closed-loop control of mechanical ventilation: description...

Closed-Loop Control of Mechanical Ventilation:Description and Classification of Targeting Schemes

Robert L Chatburn MHHS RRT-NPS FAARC and Eduardo Mireles-Cabodevila MD

IntroductionA Taxonomy of Closed-Loop Control

Control TheoryTargeting SchemesManual Targeting Schemes

Set-Point TargetingDual Targeting

Servo Targeting SchemesAutomatic Tube CompensationNeurally Adjusted Ventilatory AssistProportional Assist Ventilation

Automatic Targeting SchemesAdaptive TargetingOptimal Targeting

Intelligent Targeting SchemesSmartCare/PS

CaveatsSummary

There has been a dramatic increase in the number and complexity of new ventilation modes overthe last 30 years. The impetus for this has been the desire to improve the safety, efficiency, andsynchrony of ventilator-patient interaction. Unfortunately, the proliferation of names for ventila-tion modes has made understanding mode capabilities problematic. New modes are generally basedon increasingly sophisticated closed-loop control systems or targeting schemes. We describe the 6basic targeting schemes used in commercially available ventilators today: set-point, dual, servo,adaptive, optimal, and intelligent. These control systems are designed to serve the 3 primary goalsof mechanical ventilation: safety, comfort, and liberation. The basic operations of these schemesmay be understood by clinicians without any engineering background, and they provide the basisfor understanding the wide variety of ventilation modes and their relative advantages for improvingpatient-ventilator synchrony. Conversely, their descriptions may provide engineers with a means tobetter communicate to end users. Key words: taxonomy; targeting schemes; ventilation modes. [RespirCare 2011;56(1):85–98. © 2011 Daedalus Enterprises]

Robert L Chatburn MHHS RRT-NPS FAARC is affiliated with the Re-spiratory Institute, The Cleveland Clinic, and with Lerner College ofMedicine, Case Western Reserve University, Cleveland, Ohio. EduardoMireles-Cabodevila MD is affiliated with the Division of Pulmonary andCritical Care Medicine, Department of Medicine, University of Arkansasfor Medical Sciences, Little Rock, Arkansas.

Mr Chatburn has disclosed relationships with IngMar, Hamilton, Covi-dien, Drager, CareFusion, Newport, IngMar, Radiometer America, Breathe

Technologies, and the Alpha-1 Antitrypsin Foundation. Dr Mireles-Cabodevila has disclosed no conflicts of interest.

Mr Chatburn presented a version of this paper at the 46th RESPIRATORY

CARE Journal Conference, “Patient-Ventilator Interaction,” held March 19-21, 2010, in Cancun, Quintana Roo, Mexico.

Correspondence: Robert L Chatburn MHHS RRT-NPS FAARC, Respi-ratory Therapy, M-56, The Cleveland Clinic, 9500 Euclid Avenue, Cleve-land OH 44195. E-mail: [email protected].

RESPIRATORY CARE • JANUARY 2011 VOL 56 NO 1 85

Introduction

Over the last 10 years there have been more than adozen good reviews of closed-loop control of mechanicalventilators and the various modes they have generated.1-17

Without exception, those authors took the approach of“name and describe,” using the mode names coined bymanufacturers and the descriptions from the operator’smanuals. That approach is appropriate for the reader with acasual, general interest in modes or the need to know thedetails of a single mode. It is, perhaps, a less useful approachfor those who need a broader understanding: the clinicianwho needs to master several ventilator platforms, the ven-dor’s representative who must be conversant with both herown products and those of competitors, and, perhaps mostimportant of all, the educator who must be able to compareand contrast all modes in a systematic, practical fashion. Toserve these needs, we will attempt to take the discussion ofthis topic to the next level: that is, to classify closed-loopcontrol schemes so that we may better understand their sim-ilarities, differences, and relative advantages. In so doing wehope to bridge the knowledge-gap between engineers andclinicians and promote communication between the 2 groups.In the process we will establish some fundamental conceptsfor a larger taxonomy of modes, and we present and discusssome representative control-system diagrams. A complete tax-onomy of modes (though beyond the scope of this paper) willserve major stakeholders, including clinicians (who wish toimprove patient care), educators (who need a basis for teach-ing systems), and vendors (to facilitate communicationsamong clinicians, engineers, and marketers).

The fundamental problem facing these stakeholders isthat there are apparently too many modes and too littleevidence that may be used to compare them. We say “ap-parently” because, despite the fact that the 8th edition ofMosby’s Respiratory Care Equipment18 lists 56 uniquemode names, only about 2 dozen of them are actuallyunique modes when described at a level of detail thatincludes the targeting scheme (ie, the feedback controlsystem the ventilator uses to deliver a specific ventilatorypattern, see Table 1). Furthermore, these modes can beidentified using only 6 basic targeting schemes. Whenviewed from this perspective, the large number of venti-lation mode names does not seem so intimidating.

A Taxonomy of Closed-Loop Control

Control Theory

The term “closed-loop control” refers to the use of afeedback signal to adjust the output of a system. Ventila-

tors use closed-loop control to maintain consistent pres-sure and flow waveforms in the face of changing patient/system conditions.19,20 This is accomplished by using the

Table 1. Ventilator Manufacturers’ Ventilation Mode Names*

ManufacturerExampleVentilator

Manufacturer’s Mode Name

CareFusion Avea Synchronized IntermittentMandatory Ventilation WithMachine Volume

CareFusion Avea Volume Control Assist ControlWith Machine Volume

Covidien Puritan Bennett840

Proportional Assist VentilationPlus

Covidien Puritan Bennett840

Volume Control VentilationAssist/Control

Dräger Evita XL Automatic Tube CompensationDräger Evita XL Intermittent Positive Pressure

Ventilation With PressureLimited Ventilation

Dräger Evita XL Mandatory Minute VolumeVentilation

Dräger Evita XL Proportional Pressure SupportDräger Evita XL SmartCare/PSDräger Evita XL Synchronized Intermittent

Mandatory Ventilation WithPressure Limited Ventilation

Hamilton G5 Adaptive Support VentilationMaquet Servo-i AutoMode (PRVC - VS)Maquet Servo-i Neurally Adjusted Ventilatory

AssistMaquet Servo-i Pressure ControlMaquet Servo-i Pressure Regulated Volume

ControlMaquet Servo-i Volume ControlMaquet Servo-i Volume SupportNewport e500 Synchronized Intermittent

Mandatory VentilationPhillips V200 Pressure Support VentilationPhillips V200 Spontaneous/Timed VentilationTaema Horus Mandatory Rate VentilationVersaMed iVent201 Assist/Control (Adaptive Flow

and I-Time)VersaMed iVent201 Pressure Controlled

Synchronized MandatoryVentilation

VersaMed iVent201 Volume ControlledSynchronized MandatoryVentilation (Adaptive Flowand I-Time)

* The numerous manufacturers’ closed-loop ventilation mode names can be narrowed down tothis list of representatives of the major control schemes. Each ventilator has more modes thanare listed, but, for simplicity, only the unique modes are shown.

CLOSED-LOOP CONTROL OF MECHANICAL VENTILATION

86 RESPIRATORY CARE • JANUARY 2011 VOL 56 NO 1

output as a feedback signal that is compared to the oper-ator-set input. The difference between the two is used todrive the system toward the desired output. For example,pressure-control modes use airway pressure as the feed-back signal to control gas flow from the ventilator. Fig-ure 1 shows a schematic drawing of a general controlsystem. The input is a reference value (eg, operator presetinspiratory pressure) that is compared to the actual outputvalue (eg, instantaneous value of airway pressure). Thedifference between those 2 values is the error signal. Theerror signal is passed to the controller (eg, the softwarecontrol algorithm). The controller converts the error signalinto a signal that can drive the effector (eg, the hardware)to cause a change in the manipulated variable (eg, inspira-tory flow). The relationship between the input and theoutput of the controller is called the transfer function incontrol theory.21

Engineers need to understand the transfer function interms of complex mathematical equations. Clinicians, how-ever, need only understand the general operation of thefunction in terms of how the mode affects the patient’sventilatory pattern, and we will use that frame of referencein defining targeting schemes. The “plant” in Figure 1refers to the process under control. In our case, the plant isthe patient and the delivery circuit connecting the patientto the ventilator. The plant is the source of the “noise” thatcauses problems with patient-ventilator synchrony. At oneextreme, a paralyzed patient and an intact delivery circuitpose little challenge for a modern ventilator to deliver apredetermined ventilatory pattern, and thus synchrony isnot an issue. On the other extreme is a patient with anintense, erratic respiratory drive and a delivery circuit withleaks (eg, around an uncuffed endotracheal tube) makingpatient-ventilator synchrony virtually impossible. The chal-lenge for both clinicians and engineers is to develop tech-nology and procedures for dealing with this wide range ofcircumstances.

The plant alters the manipulated variable to generate thefeedback signal of interest as the control (output) variable.Continuing with the example above, the manipulated vari-able is flow, but the feedback control variable is pressure(ie, ventilator flow times plant impedance equals airwaypressure), as in pressure-control modes. Closed-loop con-trol can also refer to the use of feedback signals to controlthe overall pattern of ventilation, beyond a single breath, such

as the use of end-tidal carbon dioxide tension as a feedbacksignal to control minute ventilation.22

Targeting Schemes

The process of “setting” or adjusting a ventilation modecan be thought of as presetting various target values, suchas tidal volume, inspiratory flow, inspiratory pressure, in-spiratory time, frequency, PEEP, oxygen concentration,etc. We use the term “target” for 2 reasons. First, just likein archery, a target is aimed at but not necessarily hit,depending on the precision of the control system. An ex-ample would be setting a target value for tidal volume andallowing the ventilator to adjust the inspiratory pressureover several breaths to finally deliver the desired value.Indeed, in this case we could more accurately talk aboutdelivering an average target tidal volume over time.

The second reason for using “target” is because the term“control” is over-used and we need it to preserve someconventional notions about modes such as “volume con-trol” versus “pressure control.” From this use of the termtarget, we can logically refer to the control system transferfunction (relationship between the input and the output ofthe controller) as a targeting scheme. Tehrani has provideda fascinating history of ventilator targeting schemes, bothcommercial and experimental.23,24 The history of thesesystems clearly shows an evolutionary trend toward in-creasing levels of automation. In fact, as a first pass atsorting out the commercially available modes listed inTable 1, we can identify 3 groups of targeting schemesbased on increasing levels of autonomy: manual, servo,and automatic (Table 2). Manual targeting schemes re-quire the operator to adjust all the target values. Servotargeting schemes are unique in that there are no statictarget values; rather, the operator sets the parameters of amathematical model that drives the ventilator’s output tofollow a dynamic signal (like power steering on an auto-mobile). Automatic targeting schemes allow the ventilatorto set some or all of the ventilatory targets, using eithermathematical models of physiologic processes or artifi-cial-intelligence algorithms.

Manual Targeting Schemes

Manual targeting schemes are those in which the oper-ator directly adjusts the controller output. The adjustmentsmay determine the entire output as a function of time (eg,presetting the inspiratory flow waveform [eg, rectangular,sinusoidal, or descending ramp]), peak flow, and inspira-tory time. Alternatively, one might preset only a portion ofthe output, such as peak inspiratory pressure and inspira-tory pressure rise time, with inspiratory time being un-specified. Or perhaps only one parameter of the output (eg,peak inspiratory pressure) might be adjustable, and the rest

Fig. 1. Generalized control circuit (see text for explanation). The“plant” in a control circuit for mechanical ventilation is the patient.



of the pressure waveform is variable. There are 2 forms ofmanual targeting scheme: set-point and dual.

Set-Point Targeting

In set-point targeting the operator sets specific targetvalues and the ventilator attempts to deliver them. Thesimplest examples for volume-control modes would betidal volume and inspiratory flow. The simplest equationexpressing the controller function in volume control is:

V�t� � K (1)

where V(t) represents inspiratory flow as a function oftime (t), and K is the preset constant inspiratory flow.

For pressure-control modes, the operator may set in-spiratory pressure and inspiratory time or cycle threshold.The simplest equation expressing the controller function inthis case is:

P�t� � K (2)

where P(t) is inspiratory pressure as a function of time (t),and K is the preset constant inspiratory pressure.

Figure 2 shows a control circuit for the set-point controltargeting scheme, for either volume control with a constantflow or pressure control with a constant pressure. Depend-ing on the ventilator, various flow or pressure waveformsmay be preset, with correspondingly more complex equa-tions for the controller functions.

Dual Targeting

As it relates to mechanical ventilation, volume controlmeans that inspired volume, as a function of time, is pre-determined by the operator before the breath begins. Incontrast, pressure control means that inspiratory pressureas a function of time is predetermined. “Predetermined” inthis sense means that either pressure or volume is con-strained to a specific mathematical form. In the simplecase where either pressure or flow are preset constant val-ues (eg, set-point targeting, as explained above), we cansay that they are the independent variables in the equationof motion. The equation of motion for the respiratory sys-tem is a general mathematical model of patient-ventilatorinteraction:

P�t� � EV�t� � RV�t� (3)

where P(t) is inspiratory pressure as a function of time (t),E is respiratory-system elastance, V(t) is volume as a func-tion of time, R is respiratory-system resistance, and V(t) isflow as a function of time. Thus, for example, if pressureis the independent variable, then both volume and flow aredependent variables, indicating pressure control. If volumeis the independent variable, then pressure is the dependentvariable, indicating volume control. Because volume is theintegral of flow, if V(t) is predetermined, then so is V(t).

Table 2. Modes From Table 1 Sorted by Level of Autonomy

Manufacturer’s Mode Name Autonomy

Volume Control Ventilation Assist/Control ManualPressure Control ManualSynchronized Intermittent Mandatory

VentilationManual

Pressure Controlled Synchronized MandatoryVentilation

Manual

Spontaneous/Timed Ventilation ManualPressure Support Ventilation ManualIntermittent Positive Pressure Ventilation With

Pressure Limited VentilationManual

Volume Control ManualVolume Control Assist Control With Machine

VolumeManual

Synchronized Intermittent MandatoryVentilation With Pressure Limited Ventilation

Manual

Synchronized Intermittent MandatoryVentilation With Machine Volume

Manual

Automatic Tube Compensation ServoProportional Assist Ventilation Plus ServoProportional Pressure Support ServoNeurally Adjusted Ventilatory Assist ServoAssist/Control (Adaptive Flow and I-Time) AutomaticPressure Regulated Volume Control AutomaticVolume Controlled Synchronized Mandatory

Ventilation (Adaptive Flow and I-Time)Automatic

Mandatory Minute Volume Ventilation AutomaticAutoMode (PRVC - VS) AutomaticVolume Support AutomaticMandatory Rate Ventilation AutomaticAdaptive Support Ventilation AutomaticSmartCare/PS Automatic

Fig. 2. Control circuit for set-point or dual targeting schemes. Inthis example, the operator inputs a desired value for inspiratorypressure or inspiratory flow (both for dual targeting). The controlleris designed so that either pressure as a function of time (P(t)) orflow as a function of time (V(t)), is constant (K), the set value ofinspiratory pressure or flow.



Therefore, for simplicity, we include the case of flow be-ing the independent variable as a form of volume control.

Only one variable (ie, pressure or volume) can be inde-pendent at any moment, but a ventilator controller canswitch between the two during a single inspiration (seeFig. 2). When this happens the targeting scheme is calleddual set-point control or dual targeting.1 There are 2 basicways that ventilators have implemented dual targeting.One way is to start inspiration in volume control and thenswitch to pressure control if one or more preset thresholdsare met (eg, a desired peak airway pressure limit). Anexample of such a threshold is the operator-set Pmax inVolume Control on the Drager Evita XL ventilator. Theother form of dual targeting is to start inspiration in pres-sure control and then switch to volume control (eg, if apreset tidal volume has not been met when flow decays toa preset value).25 This was originally described as VolumeAssured Pressure Support Ventilation, but is currently avail-able only as a mode called Volume Control Assist Controlwith Machine Volume, in the CareFusion Avea ventilator.

Manual targeting schemes have limited capability tomaintain synchrony with patient inspiratory efforts. Thiscan be seen in the equation of motion if a term represent-ing the patient inspiratory force (muscle pressure or Pmus)is added:

P�t� � EV�t� � RV�t� � Pmus(t) (4)

For example, with set-point volume control, volume andflow are preset. Therefore, if the patient makes an inspira-tory effort (ie, Pmus(t) � 0), then the equation dictates thattransrespiratory-system pressure, P(t), must fall. Becausework is the result of both pressure and volume delivery (ie,work � �Pdv), if pressure decreases, the work the venti-lator does on the patient decreases, and hence we haveasynchrony of work demand on the part of the patientversus work output on the part of the ventilator.

With set-point pressure control, transrespiratory pres-sure is preset. Therefore, if the patient makes an inspira-tory effort, both volume and flow increase. With constantpressure and increased volume, work per liter for the breathstays constant. While this gives better work synchronythan does volume control, it is still not ideal.

Servo Targeting Schemes

The term “servo” was coined by Joseph Farcot in 1873to describe steam-powered steering systems. Later, hy-draulic “servos” were used to position anti-aircraft guns onwarships.26 Servo control specifically refers to a controlsystem that converts a small mechanical motion into onerequiring much greater power, using a feedback mecha-nism. As such, it offers a substantial advantage in terms of

creating ventilation modes capable of a high degree ofsynchrony with patient breathing efforts. That is, ventila-tor work output can be made to match patient work de-mand with a high degree of fidelity. Younes, at the sug-gestion of Warren Sanborn, stated that Proportional AssistVentilation (PAV) was analogous to power steering inmotor vehicles.27 Therefore, we apply the name servo con-trol to targeting schemes in which the ventilator’s outputautomatically follows a varying input. This includes PAV,Automatic Tube Compensation (ATC), and Neurally Ad-justed Ventilatory Assist (NAVA), in which the airwaypressure signal not only follows but amplifies signals thatare surrogates for patient effort (ie, volume, flow, anddiaphragmatic electrical signals). Note that the term “servocontrol” has been loosely used since it was coined to referto any type of general feedback control mechanism, but weare using it in a very specific manner, as it applies toventilator targeting schemes.

Automatic Tube Compensation

ATC is perhaps the simplest example of a servo target-ing scheme on a ventilator. The goal of ATC, in terms ofpatient-ventilator synchrony, is to support the resistive workof breathing through the artificial airway. Figure 3 showsa control circuit for Automatic Tube Compensation. ATCis a form of pressure control, and the mathematical modelof the controller relates airway pressure to the inspiratoryflow of the patient’s spontaneous breath. A spontaneousbreath is one in which inspiration is both triggered (started)and cycled (stopped) by the patient. A breath that is eithertriggered or cycled by the ventilator is a mandatory breath.28

The operator inputs the size and type of the artificial air-way, and the ventilator uses a look-up table to determinethe airway resistance. Then the controller manipulates flowto control pressure according to a mathematical model thatsets pressure equal to the product of resistance and thesquare of (inspiratory or expiratory) flow29,30:

P�t� � RV�t�2 (5)

Fig. 3. Control circuit for a servo targeting scheme (eg, AutomaticTube Compensation). The controller is designed so that inspira-tory pressure as a function of time (P(t)) is proportional to thesquare of flow as a function of time (V(t)2), with the constant ofproportionality being artificial airway resistance (R).



where P(t) is inspiratory pressure relative to end-expira-tory pressure as a function of time (t), R is the resistanceof the artificial airway, and V is flow as a function of time.Because airway pressure with ATC is a function of thepatient’s otherwise unassisted spontaneous inspiratory flow(eg, during CPAP), and because flow is generated by thepatient’s muscle pressure (Pmus), ATC can be thought of asa simple amplifier of patient effort.31

Neurally Adjusted Ventilatory Assist

NAVA is a servo targeting scheme similar to ATC butwith more complex requirements for implementation. Interms of patient-ventilator synchrony, NAVA supports bothresistive and elastic work of breathing in proportion to thepatient’s inspiratory effort. With NAVA the controller setsairway pressure to be proportional to Pmus, based on thevoltage recorded from diaphragmatic activity measured bysensors embedded in an orogastric tube:

P�t� � KEdi�t� (6)

where P(t) is inspiratory pressure as a function of time (t),K is the NAVA support level (an amplification factor), andEdi(t) is the electrical signal from the diaphragm as afunction of time. The operator inputs the constant of pro-portionality between voltage and pressure (ie, gain factor).Then the controller sets airway pressure to equal the prod-uct of the gain and the Edi. Figure 4 shows a control circuitfor NAVA.32

Proportional Assist Ventilation

PAV is a servo targeting scheme that uses a more com-plex model than ATC or NAVA. In terms of patient-ven-tilator synchrony, PAV, like NAVA, supports both resis-tive and elastic work of breathing in proportion to thepatient’s inspiratory effort, but it uses different feedbacksignals. PAV is a form of pressure control based on theequation of motion for the respiratory system24 in the form:

P(t)�K1V(t)�K2V(t) (7a)

where inspiratory pressure relative to end-expiratory pres-sure as a function of time [P(t)] is the sum of 2 compo-nents. The first is the “volume assist” or the amount ofelastic load supported (ie, K1 times volume as a functionof time [V(t)]). The second component is the “flow assist”or the amount of resistive load supported (ie, K2 timesflow as a function of time [V(t)]). Both volume and floware used as feedback signals and represent the combinedresult of patient effort and ventilator support. The valuesof K1 and K2 are preset by the operator and represent thesupported elastance and resistance, respectively (Fig. 5).Because volume and flow are initiated by the patient’smuscle pressure (Pmus), the pressure generated by PAVcan be thought of as an amplifier of Pmus.

The form of PAV implemented on the Drager Evita XLventilator (called Proportional Pressure Support) requiresthe operator to input desired values for elastance (K1) andresistance (K2) to be supported. ATC can be added to thismode. PAV implemented on the Puritan Bennett 840 ven-tilator (called PAV Plus) uses a slightly different equation,which automatically calculates the resistance of the artifi-cial airway and combines K1 and K2, such that the oper-ator enters only a single value representing the percentagework of breathing to be supported. PAV Plus also contin-uously monitors lung mechanics to keep the controllerinformed of changing values of respiratory-system elas-tance and resistance. The design differences between Pro-portional Pressure Support and PAV Plus lead to impor-tant performance differences. Figure 6 shows results fromour laboratory, indicating that the addition of ATC to Pro-portional Pressure Support yields a huge variation in therelationship between the work demanded by the patientand the work output by the ventilator.33 At an ATC levelof 100%, a 30% increase in patient work resulted in a200% increase in ventilator work output. Of equal con-

Fig. 4. Control circuit for a servo targeting scheme (eg, NeurallyAdjusted Ventilatory Assist). The controller is designed so thatinspiratory pressure as a function of time (P(t)) is proportional tothe electrical signal from the diaphragm (Edi(t)), which is also afunction of time. The constant of proportionality (K) is the outputgain, expressed in cm H2O/�V.

Fig. 5. Control circuit for a servo targeting scheme (eg, Propor-tional Assist Ventilation). The controller is designed so that inspira-tory pressure as a function of time (P(t)) is proportional to bothvolume as a function of time (V(t)) and flow as a function of time(V(t)). The constant of proportionality K1 represents the amount ofelastance to be supported. The constant of proportionality K2 rep-resents the amount of resistance to be supported.



cern, the inability of Proportional Pressure Support to mon-itor and correct for changes in respiratory-system mechan-ics can lead to unexpected results. Figure 6 shows that forrelatively small increases in respiratory-system compli-ance (the inverse of elastance) the delivered tidal volumewas increased with Proportional Pressure Support but wasconstant with PAV Plus.

Before concluding this section, we note that the simpledidactic construct of “independent versus dependent vari-able,” mentioned above in the section on dual targetingschemes, breaks down in the case of servo targetingschemes. For example, with PAV, pressure depends onvolume and flow. At fist thought this might suggest thatPAV is a form of volume control. But volume and flow arenot predetermined static targets, as in the case of volumecontrol: they are dynamically determined by the patient.Nevertheless, we need to classify PAV and the other servotargeting schemes as either pressure control or volumecontrol. This is accomplished by referring to the generaldefinition of the “control variable” as being the variablethat is predetermined by the operator (in this case theventilator). In the case of the servo targeting schemes,pressure as a function of time, P(t), is predetermined byspecific mathematical models; that is, P(t) � RV(t)2 for ATC,P(t) � KEdi(t) for NAVA, and P(t) � K1V(t) � K2V(t) forPAV. Therefore we conclude that these servo targetingschemes are forms of pressure control.

Automatic Targeting Schemes

Automatic targeting schemes allow the ventilator to ad-just some set-points according to either a mathematical

model derived from known physiologic processes34 or someform of artificial intelligence, such as a rule-based expertsystem, fuzzy-logic, or an artificial neural network.35 Ofcourse, it is possible to combine approaches into hybridsystems.36-38

Artificial-intelligence systems automate the clinician’sheuristics (rules derived from experimentation or experi-ence) and may enforce protocols when humans would fal-ter for various reasons. However, a disadvantage of arti-ficial-intelligence systems is that they have limitedtransparency, meaning that the way set-points are adjustedis not easily understood because of the complex interac-tion of many rules. Furthermore, such rules may be sub-jective (ie, based on expert opinion, leading to increasedcomplexity). Another option is to use mathematical mod-els, which have the advantage of being transparent andobjective.22 Models may also help the clinician understandthe patient’s physiologic state, which are represented bythe model’s parameters. If the parameters are estimated fora particular patient, the model might be used to answer“what if” questions and predict the outcome of differentventilation strategies.39 On the other hand, such modelscan be difficult to derive and can suffer from having toomany parameters.40

There are 3 general categories of automatic targetingscheme in current use: adaptive, optimal, and intelligenttargeting.

Adaptive Targeting

An adaptive targeting scheme involves modifying thefunction the controller uses to cope with the fact that the

Fig. 6. Comparison of 2 implementations of Proportional Assist Ventilation: Proportional Pressure Support (PPS) (Drager) and PAV Plus(Covidien). A: Both modes increase ventilator work output as patient work demand increases, but the relationship in PPS is highly variable,depending on the level of Automatic Tube Compensation (ATC) chosen. PAV Plus includes tube compensation as part of the controllerfunction and it is not user adjustable. B: PAV Plus delivers the same tidal volume as compliance changes, because it monitors complianceand changes the controller function. PPS does not monitor compliance and thus overcompensates as compliance increases. This effectis amplified with the use of ATC.



system parameters being controlled are time-varying.41 Asit applies to mechanical ventilation, adaptive targetingschemes allow the ventilator to set some (or conceivablyall) of the targets in response to varying patient conditions.There are currently 4 distinct approaches to basic adaptivetargeting, represented by the mode names Pressure Regu-lated Volume Control, Mandatory Rate Ventilation, Adap-tive Flow/Adaptive I-Time, and Mandatory Minute Ven-tilation. Two advanced forms of adaptive targeting (optimaland intelligent) will be treated separately.

Pressure Regulated Volume Control. Pressure Regu-lated Volume Control is one example of adaptive targetingof inspiratory pressure, with the goal of providing the flowsynchrony of set-point pressure control with the tidal-vol-ume assurance of set-point volume control. One of theconcerns with set-point pressure-control ventilation is thatit cannot guarantee a minimum tidal volume in the face ofchanging lung mechanics or patient effort. To solve thisproblem, Siemens introduced adaptive pressure targetingin a mode they called Pressure Regulated Volume Control.This targeting scheme has also been given the generalname adaptive pressure control.42,43 The operator inputs atarget tidal volume and the controller manipulates the flow,within a breath, to control the inspiratory pressure to aconstant value (pressure control). Between breaths the in-spiratory pressure is automatically adjusted by the venti-lator to achieve the target volume (Fig. 7). Adaptive pres-sure targeting has been used for both mandatory andspontaneous breaths, and the various modes that use thisscheme have been given many different proprietary namesby ventilator manufacturers (eg, Covidien uses VC� andVV�, Drager uses AutoFlow, and Hamilton uses APV)

To understand what the targeting scheme does, we firstnote that the target tidal volume has 2 components, onedue to the ventilator and the other due to the patient:

Vtarget�Vvent�Vmus (7b)

where Vtarget is the tidal volume target set by the operator,Vvent is the volume delivered by the ventilator-generatedpressure, P(t), and Vmus is the volume generated by thepatient’s muscle pressure as a function of time, Pmus(t),during the inspiratory time (assuming the presence of apatient inspiratory effort). In the simplest case of

P�t� � constant � �P (8)

then

Vvent � �C�P��1 � e�TI/RC) (9)

where R is respiratory-system resistance, C is respiratory-system compliance, �P is inspiratory pressure above PEEP,e is the base of natural logarithms (approximately 2.72),and TI is inspiratory time.

Vmus is unpredictable because Pmus(t) is unpredictable.Therefore, we can express the target tidal volume as:

Vtarget � �C�P��1 � e�TI/RC) � Vmus (10)

Solving this equation for �P gives a simple model for howadaptive targeting automatically adjusts inspiratory pres-sure:

�P � �Vtarget � Vmus�/C�1 � e�TI/RC)] (11)

This equation predicts that, if Vmus stays constant, theventilator will increase �P if either R increases or C de-creases (or both). This is intuitively obvious, because if thelungs are stiffer or flow is restricted by narrower airways,then it will take more inspiratory pressure to hit the targettidal volume. Likewise, the ventilator will decrease �P ifeither R decreases or C increases (or both). On the otherhand, the equation shows that the ventilator will decrease�P if patient inspiratory effort increases (ie, Vmus increases).This latter effect is what has led some authors to concludethat adaptive pressure targeting is a type of automatic wean-ing. The problem is that the controller has no way ofknowing whether the increase in Vmus and resultant de-crease in inspiratory pressure are appropriate (eg, the pa-tient is recovering from sedation) or inappropriate (eg,Vmus increasing due to pain or anxiety). Furthermore, thecontroller cannot distinguish an increase in patient effort(increased Vmus) from improved lung mechanics (increasein C or decrease in R). The controller always follows thepatient’s lead (by decreasing support in response to anincrease in effort), whereas the clinician would lead thepatient by decreasing support and then observing the pa-tient’s response. Therefore, this form of adaptive pressuretargeting should not be considered a form of automatic

Fig. 7. Control circuit for an adaptive pressure targeting scheme(eg, Pressure Regulated Volume Control). The controller is de-signed so that inspiratory pressure relative to PEEP (IP) is adjustedto meet the target tidal volume, given changes in resistance (R),compliance (C), and volume (Vmus) generated by patient effort(Pmus). IP will increase if R increases or C decreases. IP will de-crease if Vmus increases. TI � inspiratory time.



weaning per se, although it may have that effect in certainsituations.

Mandatory Rate Ventilation. An interesting variationof adaptive targeting is Mandatory Rate Ventilation, on theTaema-Horus ventilator (Air Liquide, Antony, France),44

which is a form of pressure support in which the inspira-tory pressure is automatically adjusted based on a targetvalue for the patient’s spontaneous breath frequency(Fig. 8). The operator sets a target frequency and the con-troller adjusts the inspiratory pressure (ie, the pressure-support level) in proportion to the difference between theactual and target breath frequencies:

��P� factual � ftarget (12)

where �P is inspiratory pressure relative to PEEP, and�(�P) is the change in inspiratory pressure made by theventilator in response to changes in the patient’s sponta-neous breathing frequency. The operator-set target fre-quency, ftarget is the spontaneous respiratory frequency thepatient is expected to perform when comfortable. The ven-tilator compares the average actual frequency from the last4 breaths (factual) to the target frequency. If factual is higherthan the ftarget, the inspiratory pressure is automaticallyincreased. If the factual is lower than the ftarget, the inspira-tory pressure is automatically decreased. The basic idea isthat when the inspiratory pressure is correctly adjusted, thepatient will have an acceptable ventilatory frequency (eg,15–25 breaths/min) and will be breathing comfortably.When the pressure is too low, the patient’s frequency willincrease and vice versa. Hence, the controller attempts tokeep the patient in a comfortable range of ventilatory as-sistance as assessed by frequency, assuming the operatorhas selected the appropriate target frequency.

Adaptive Flow and Adaptive I-Time. The General Elec-tric Versamed iVent ventilator has 2 unique adaptive tar-

geting schemes available in volume-control modes calledAdaptive Flow and Adaptive I-Time. Working togetherwith a preset tidal volume, these algorithms automaticallyadjust the inspiratory time and inspiratory flow to maintaina target inspiratory-expiratory ratio of 1:2 (considered anormal value) in the face of a variable breathing frequencyon the part of the patient (Fig. 9). This action can bedescribed by the equation:

�VT/VI�:�1/f � TI� � 1:2 (13)

where VT is the preset tidal volume, VI is the ventilatoradjustable inspiratory flow (constant during a given breath),f is the patient’s spontaneous breathing frequency and TI isthe ventilator adjustable inspiratory time. For example, asthe frequency increases, the ventilatory period (inspiratorytime plus expiratory time, or 1/f) decreases. However, be-cause VT and VI are preset before the breath, the TI ispreset, hence the inspiratory-expiratory ratio increases. Inan effort to maintain the target inspiratory-expiratory ratioof 1:2, the ventilator will respond by automatically de-creasing the TI of the next breath. However, to maintainthe target VT, it must also increase VI (ie, VT � VI � TI).In summary, VT is preset by the operator, f is determinedby the patient, VI and TI are automatically adjusted by theventilator between breaths to maintain a target inspiratory-expiratory ratio of 1:2.

Mandatory Minute Ventilation. The idea of automat-ically adjusting the ventilation pattern to maintain a con-stant minute volume was first described in 1977.45 Asimplemented, for example, on the Drager Evita XL ven-tilator, Mandatory Minute Ventilation is a form of volume-control intermittent mandatory ventilation. The operatorpresets the target minute ventilation by setting tidal vol-ume and frequency. The ventilator then monitors the totalminute ventilation as the sum of the minute ventilations

Fig. 9. Control circuit for an adaptive pressure targeting scheme(eg, Adaptive Flow and Adaptive I-Time). The controller is de-signed so that for a preset tidal volume, inspiratory time and floware adjusted automatically, in the face of a changing breathingfrequency, to maintain the inspiratory-expiratory ratio at 1:2.VT � tidal volume. V � inspiratory flow (constant during a givenbreath). f � patient’s spontaneous breathing frequency. TI � in-spiratory time. inspiratory-expiratory ratio is the ratio of inspiratorytime to expiratory time.

Fig. 8. Control circuit for an adaptive pressure targeting scheme(eg, Mandatory Rate Ventilation). The controller is designed so thatinspiratory pressure as a function of time (P(t)) is proportional tothe difference between the target and the actual spontaneousbreathing frequencies (ftarget and factual). If the difference is posi-tive, the inspiratory pressure is increased. If the difference is neg-ative, the inspiratory pressure is reduced. Thus, the system con-verges on a Pressure Support level that should keep the patientcomfortable, assuming the target frequency is selected correctly.



generated by mandatory and spontaneous breaths (Fig. 10).The controller function can be expressed as:

MVtotal � MVmandatory � MVspontaneous (14)

where MV is minute ventilation. If the total minute ven-tilation is below the target value, the mandatory breathfrequency will increase. However, as long as the sponta-neous minute ventilation is at least equal to the targetvalue, mandatory breaths will be suppressed. In this way,the proportion of the total minute ventilation generated byspontaneous breaths can range from 0% to 100%. As withadaptive pressure targeting schemes, Mandatory MinuteVentilation has been called a mode of automatic weaning.However, Mandatory Minute Ventilation has the same lim-itations mentioned above, in that the suppression of man-datory breaths may be due to an inappropriate increase inspontaneous breaths, perhaps even with a rapid, shallowpattern that actually decreases minute alveolar ventilation.The Hamilton Veolar ventilator used a variation of thistargeting scheme by making Mandatory Minute Ventila-tion a form of adaptive pressure support. The target minuteventilation was maintained by automatic adjustment ofinspiratory pressure. That mode was replaced by AdaptiveSupport Ventilation (ASV) (see optimal targeting below)on newer Hamilton ventilators.

Optimal Targeting

Optimal targeting is an advanced form of adaptive tar-geting.46 Optimal targeting in this context means that theventilator controller automatically adjusts the targets ofthe ventilatory pattern to either minimize or maximizesome overall performance characteristic.

Adaptive Support Ventilation. ASV, on the Hamiltonventilators, is the only commercially available mode todate that uses optimal targeting. This targeting scheme wasfirst described by Tehrani in 1991,47,48 and was designedto minimize the work rate of breathing, mimic natural

breathing, stimulate spontaneous breathing, and reduceweaning time.24 The operator inputs the patient’s weight,from which the ventilator estimates the required minutealveolar ventilation, assuming a normal dead-space frac-tion. Next, an optimum frequency is calculated based onwork by Otis et al,49 that predicts a frequency resulting inthe least mechanical power (work rate)24:

f �

�1 � �1 � 4�2RCE�MV � fVD

VD�

1 � 2�2RCE(15)

where MV is predicted minute ventilation (in L/min) basedon patient weight and the setting for percent-of-predictedMV to support, VD is predicted dead space (in L) based onpatient weight, RCE is the expiratory time constant calcu-lated as the slope of the expiratory flow-volume curve, andf is the computed optimal frequency (in breaths/min). Thetarget tidal volume is calculated as MV/f. The controlcircuit for ASV shows that the controller uses the Otisequation to set the tidal volume (Fig. 11). As with simpleadaptive pressure targeting, the inspiratory pressure withina breath is controlled to achieve a constant value, andbetween breaths the inspiratory pressure is adjusted toachieve a target tidal volume. However, unlike simple adap-tive pressure targeting, the target is not set by the operator,but rather it is estimated by the ventilator in response tochanges in respiratory-system mechanics and patient ef-fort. Individual pressure-targeted breaths may be manda-tory (time-triggered and time-cycled) or spontaneous (flow-triggered and flow-cycled).

ASV adds some expert rules that put safety limits onfrequency and tidal-volume delivery and reduce the risk ofauto-PEEP. In that sense, this mode may be considered anintelligent targeting scheme or, more appropriately, a hy-brid system (ie, using a mathematical model and artificialintelligence).

Fig. 10. Control circuit for an adaptive pressure targeting scheme(eg, Mandatory Minute Ventilation). The controller is designed sothat the sum of the mandatory minute ventilation (MV) and thespontaneous MV is equal to the preset value (K) determined by thepreset tidal volume and frequency.

Fig. 11. Control circuit for an optimum targeting scheme (eg, Adap-tive Support Ventilation). The controller is designed to select afrequency and tidal volume that minimizes the work-rate of breath-ing. The tidal volume is delivered using the adaptive pressure con-trol targeting scheme in Figure 7. MV � minute ventilation. IP �inspiratory pressure.



Intelligent Targeting Schemes

Intelligent targeting is another form of adaptive target-ing, which uses artificial-intelligence techniques.50

SmartCare/PS

SmartCare/PS on the Drager XL ventilator is the onlyventilation mode commercially available to date in theUnited States with a targeting scheme that relies entirelyon a rule-based expert system. SmartCare/PS is a special-ized form of pressure support that is designed for true (ie,ventilator-led) automatic weaning. The SmartCare/PS con-troller uses predefined acceptable ranges for spontaneousbreathing frequency, tidal volume, and end-tidal carbondioxide tension to automatically adjust the inspiratory pres-sure to maintain the patient in a “respiratory zone of com-fort”51 (Fig. 12).

The SmartCare/PS system divides the control processinto 3 steps. The first step is to stabilize the patient withinthe “zone of respiratory comfort,” defined as combinationsof tidal volume, respiratory frequency and end-tidal CO2

defined as acceptable by the artificial-intelligence programfor the operator-set patient diagnosis (ie, COPD or neuro-muscular disorder). The second step is to progressivelydecrease the inspiratory pressure while making sure thepatient remains in the “zone.” The third step tests readi-ness for extubation by maintaining the patient at the lowestlevel of inspiratory pressure, which depends on the type ofartificial airway (endotracheal tube vs tracheostomy tube),the type of humidifier (heat-and-moisture exchanger vsheated humidifier), and the use of Automatic Tube Com-pensation. Once the lowest level of inspiratory pressure isreached, a one-hour observation period is started, duringwhich the patient’s breathing frequency, tidal volume, andend-tidal CO2 are monitored. Upon successful completionof this step (essentially a spontaneous breathing trial), amessage on the screen suggests that the clinician “considerseparation” of the patient from the ventilator. This methodof automatic weaning reduced the duration of mechanical

ventilation and intensive care unit stay in a multicenterrandomized controlled trial.52 However, the advantage ofartificial intelligence may be less noticeable in environ-ments where natural intelligence is plentiful. Rose et alrecently concluded that “substantial reductions in weaningduration previously demonstrated were not confirmed whenthe SmartCare/PS system was compared to weaning man-aged by experienced critical care specialty nurses, using a1:1 nurse-to-patient ratio. The effect of SmartCare/PS maybe influenced by the local clinical organizational context.”53

Caveats

All closed-loop control systems are only as good astheir feedback signals. For example, if adaptive pressuretargeting schemes have inaccurate feedback informationabout target volume, due to leaks, then the system will notwork properly and a more primitive system (eg, set-point)must be used. If a servo system such as NAVA cannotobtain an Edi signal, it reverts back to a set-point mode.For systems that rely on physiologic models, the ventilatormust be able to obtain the required model parameters. Forexample, if a patient has a high respiratory drive, PAV Plusmay not be able to calculate respiratory-system compli-ance (using a brief inspiratory hold) and thus will not beable to properly control airway pressure. In the case ofmathematical models, the model assumptions must reason-ably fit the actual patient conditions. For example, ASVassumes a normal ratio of dead space to tidal volume anda normal minute ventilation requirement. If these assump-tions do not apply to the patient at hand, the mode will notwork as expected. Another example is SmartCare/PS. Ifthe patient is not ready to wean, SmartCare/PS is not theright mode, because its rule-based system is designed forpatients who can tolerate repeated reductions in support.SmartCare/PS also assumes a “normal” serum bicarbonatelevel, which may not be a correct assumption in peopleliving at high altitudes.54 If anything, the more advancedthe mode, the more the need for vigilance and understand-ing on the part of clinicians: hence this paper.

Summary

A major problem facing clinicians, educators, and evenvendors, related to ventilation modes, is the vast prolifer-ation of names and the lack of a standardized approach tounderstanding them. A key component of a ventilation-mode taxonomy is a set of concepts that succinctly de-scribe the operation of closed-loop control schemes in ven-tilators. Fortunately, the multitude of modes commerciallyavailable today can be described with only 6 targetingschemes (set-point, dual, servo, adaptive, optimal, and in-telligent, Table 3), which serve the 3 primary goals ofmechanical ventilation: safety, comfort, and liberation (Ta-

Fig. 12. Control circuit for an intelligent targeting scheme (eg, Smart-Care/PS). The controller is designed to identify automatically ad-just the inspiratory pressure relative to PEEP (IP) to keep the pa-tient in a “zone of comfort,” defined in terms of the instantaneousvalues of tidal volume, frequency and end-tidal CO2.



ble 4). The basic operation of the 6 schemes may be un-derstood by clinicians without any engineering background,yet their descriptions may provide clinicians and engineerswith a better means to communicate. The way that target-ing schemes contribute to a taxonomy of modes has beendescribed elsewhere,28 and is illustrated in Table 5.

Clearly, closed-loop control of mechanical ventilators isevolving at a rapid pace. The United States Food and DrugAdministration’s recent acceptance of intelligent targetingschemes such as ASV and SmartCare/PS has paved theway for even more sophisticated systems. Indeed, in thenot too distant future the concept of a discrete “mode” ortargeting system may become obsolete. As ventilators be-come more intelligent and able to adapt to a broader rangeof feedback signals indicating changing patient require-ments, the ventilatory patterns will become fluid and in-terchangeable. We would go so far as to predict that theparadigm will change from the current one, “set the mode and monitor the outcome,” to “set the outcome and mon-

itor the mode.” Modes such as “Assist/Control” and“SIMV” will be transitory states instead of preset patterns,and hence the way we think about setting ventilators, modes,and the operator-interface paradigms, must change.

Nevertheless, there have been very few randomized con-trolled trials exploring the use of advanced targetingschemes, and in those trials (eg, of PAV, ATC, and Smart-Care) superiority has not been clearly (nor robustly) dem-onstrated. Further such studies would be expensive andtime-consuming, and might never be performed. That said,the absence of such studies should imbue the reader witha healthy dose of skepticism. Greater complexity of ven-tilation modes may be better, but we don’t know for sure.This is also relevant because we know that poor patient-ventilator interaction is associated with poor outcomes, butwhether patient-ventilator interaction is the cause of thosepoor outcomes remains uncertain.

The technical developments described in this paper willdemand even higher levels of understanding and skill fromclinicians, just as automated flight controls have placedhigher training demands on pilots. We trust that such fu-ture developments do not instill a fear of human reason

Table 4. Goals and Objectives of Mechanical Ventilation

Promote SafetyOptimize ventilation-perfusion of the lung

Maximize alveolar ventilationMinimize shunt

Optimize pressure/volume curveMinimize tidal volumeMaximize compliance

Promote ComfortOptimize patient-ventilator synchrony

Maximize trigger/cycle synchronyMinimize auto-PEEPMaximize flow synchronyCoordinate mandatory and spontaneous breaths

Optimize work demand versus work deliveredMinimize inappropriate shifting of work from ventilator to patient

Promote LiberationOptimize the weaning experience

Minimize adverse eventsMinimize duration of ventilation

Table 3. Targeting Schemes

ManualSet-pointDual

ServoSimple servoAdvanced servo

AutomaticAdaptive

Simple adaptiveOptimalIntelligent

Table 5. Outline of a Ventilation Mode Taxonomy, Showing theContribution of Targeting Schemes*

Primary Control Variable for Inspiration (refers to mandatory breathsin continuous mandatory ventilation (CMV) and intermittentmandatory ventilation (IMV), or spontaneous breaths incontinuous spontaneous ventilation (CSV)

VolumePressure

Breath SequenceCMVIMVCSV

Primary Targeting Scheme (refers to mandatory breaths in CMV andIMV, or spontaneous breaths in CSV)

Set-pointServoDualAdaptiveOptimalIntelligent

Secondary Targeting Scheme (refers to spontaneous breaths in IMV)Set-pointServoAdaptiveOptimalIntelligent

* All ventilation modes can be placed in one of 5 mutually exclusive groups, using the first 2levels of the outline: volume-controlled continuous mandatory ventilation (VC-CMV),volume-controlled intermittent mandatory ventilation (VC-IMV), pressure-controlledcontinuous mandatory ventilation (PC-CMV), pressure-controlled intermittent mandatoryventilation (PC-IMV), and pressure-controlled continuous spontaneous ventilation (PC-CSV).



being replaced by machine intelligence, nor a false senseof security such that we “fall asleep at the wheel.”

Glossary

Adaptive targeting scheme. Feedback control systemin which the ventilator determines some (or conceivablyall) of the targets in response to changes in patient-deriveddata.

Automatic targeting scheme. Feedback control systemthat allows the ventilator to set some or all of the venti-latory targets using either mathematical models of physi-ologic processes or artificial-intelligence algorithms.

Closed-loop control. The use of a feedback signal toadjust the output of a system.

Dual targeting scheme. Feedback control in which theventilator switches between volume control and pressurecontrol during a single inspiration.

Equation of motion for the respiratory system. Amathematical model of patient-ventilator interaction. Thesimplest version of this equation, assuming passive inspi-ration and no auto-PEEP is

P�t� � EV�t� � RV�t�

where P(t) is inspiratory pressure as a function of time, Eis respiratory-system elastance, V(t) is volume as a func-tion of time, R is respiratory-system resistance, and V(t) isflow as a function of time.

Intelligent targeting scheme. Targeting schemes thatuse artificial-intelligence techniques such as rule-based ex-pert systems, fuzzy logic and artificial neural networks.

Mandatory breath. A breath in which inspiration istriggered (started) or cycled (stopped) by the ventilator (orboth).

Manual targeting scheme. Feedback control systemthat requires the operator to adjust all the target values.

Ventilation mode. A predetermined pattern of interac-tion between a mechanical ventilator and the system beingventilated (eg, a patient); a specific combination of controlvariable, breath sequence, and targeting scheme.

Optimal targeting scheme. Feedback control system inwhich the ventilator automatically adjusts the targets toeither minimize or maximize some overall performancecharacteristic.

Pressure control. A general category of ventilationmodes in which inspired pressure as a function of time ispredetermined before the breath begins. For simple pres-sure-control modes (eg, pressure support), pressure is theindependent variable and volume and flow are the depen-dent variables in the equation of motion for the respiratorysystem.

Servo targeting scheme. Feedback control system inwhich the operator sets the parameters of a mathematicalmodel that drives the ventilator output to follow a dynamicsignal.

Set-point targeting scheme. Feedback control systemin which the operator sets specific target values.

Spontaneous breath. A breath in which inspiration isboth triggered (started) and cycled (stopped) by the pa-tient.

Target. A predetermined goal during inspiration or ex-piration (eg, inspiratory flow or average tidal volume).

Targeting scheme. Feedback control system used by amechanical ventilator to deliver a specific ventilatory pat-tern; a key component of a ventilation mode.

Tidal pressure. The change in transalveolar pressure(the pressure in the alveolar region minus the pressure inthe pleural space) associated with the delivery of a tidalvolume. In pressure-control modes the tidal pressure isestimated as peak airway pressure minus PEEP. In vol-ume-control modes it is estimated as plateau pressure (ie,pressure during an inspiratory hold) minus PEEP. Tidalpressure is a useful parameter when switching betweenvolume control and pressure control and attempting toadjust ventilator settings to maintain the same alveolarventilation. For example, when switching from volumecontrol to pressure control, suppose the plateau pressure involume control is 20 cm H2O and PEEP is 5 cm H2O. Thetidal pressure is 20 – 5 � 15 cm H2O. Thus, to achieveapproximately the same tidal volume in pressure control,inspiratory pressure should be set at 15 cm H2O abovePEEP, and inspiratory time should be at least 3 time-constants long to allow equilibration of airway pressurewith alveolar pressure.

Volume control. A general category of ventilation modesin which inspired volume (or flow), as a function of time,is predetermined before the breath begins. For simplevolume-control modes (eg, Assist/Control or SIMV) vol-ume is the independent variable and pressure is the de-pendent variable in the equation of motion for the respi-ratory system.

REFERENCES

1. Branson RD, Davis K Jr. Dual control modes: combining volumeand pressure breaths. Respir Care Clin N Am 2001;7(3):397-408.

2. Branson RD, Johannigman JA, Campbell RS, Davis K Jr. Closed-loop mechanical ventilation. Respir Care 2002;47(4):427-451.

3. Navalesi P, Costa R. New modes of mechanical ventilation: propor-tional assist ventilation, neurally adjusted ventilatory assist, and frac-tal ventilation. Curr Opin Crit Care 2003;9(1):51-58.

4. Branson R. Understanding and implementing advances in ventilatorcapabilities. Curr Opin Crit Care 2004;10(1):23-32.

5. Branson RD, Johannigman JA. What is the evidence base for thenewer ventilation modes? Respir Care 2004;49(7):742-60.

6. Snow TM, Brandon DH. A nurse’s guide to common mechanicalventilation techniques and modes used in infants: nursing implica-tions. Adv Neonatal Care 2007;7(1):8-21.



7. Wysocki M, Brunner JX. Closed-loop ventilation: an emerging stan-dard of care? Crit Care Clin 2007;23(2):223-240.

8. Branson RD, Davis K Jr. Does closed loop control of assist controlventilation reduce ventilator-induced lung injury? Clin Chest Med2008;29(2):343-350.

9. Burns KE, Lellouche F, Lessard MR. Automating the weaning pro-cess with advanced closed-loop systems. Intensive Care Med 2008;34(10):1757-1765.

10. Santanilla JI, Daniel B, Yeow ME. Mechanical ventilation. EmergMed Clin North Am 2008;26(3):849-862.

11. Lellouche F, Brochard L. Advanced closed loops during mechanicalventilation (PAV, NAVA, ASV, SmartCare). Best Pract Res ClinAnaesthesiol 2009;23(1):81-93.

12. Johannigman JA, Muskat P, Barnes S, Davis K, Branson RD. Auton-omous control of ventilation. J Trauma 2008;64(4 Suppl):S302-S320.

13. Puri N, Puri V, Dellinger RP. History of technology in the intensivecare unit. Crit Care Clin 2009;25(1):185-200.

14. Donn SM. Neonatal ventilators: how do they differ? J Perinato.2009;29(Suppl 2):S73-S78.

15. Hummler H, Schulze A. New and alternative modes of mechanicalventilation in neonates. Semin Fetal Neonatal Med 2009;14(1):42-48.

16. Branson RD, Johannigman JA. Innovations in mechanical ventila-tion. Respir Care 2009;54(7):933-947.

17. Keszler M. State of the art in conventional mechanical ventilation. JPerinatol 2009;29(4):262-275.

18. Cairo JM, Pilbeam SP. Mosby’s respiratory care equipment, 8thedition. St Louis: Mosby Elsivier; 2009.

19. Chatburn RL. Computer control of mechanical ventilation. RespirCare 2004;49(5):507-515.

20. Pauldine R, Beck G, Salinas J, Kaczka DW. Closed-loop strategiesfor patient care systems. J Trauma 2008;64(4 Suppl):S289-S294.

21. Zaleski JR. Integrating device data into the electronic medical record.Erlangen, Germany: Publicis; 2010:209-234.

22. Martinoni EP, Pfister ChA, Stadler KS, Schumacher PM, Leibundgut D,Bouillon T, et al. Model-based control of mechanical ventilation: designand clinical validation. Br J Anaesth 2004;92(6):800-807.

23. Tehrani FT. Automatic control of mechanical ventilation. Part 1:Theory and history of the technology. J Clin Monit Comput 2008;22(6):409-415.

24. Tehrani FT. Automatic control of mechanical ventilation. Part 2: Theexisting techniques and future trends J Clin Monit Comput 2008;22(6):417-424.

25. Amato MB, Barbas CS, Bonassa J, Saldiva PH, Zin WA, de Car-valho CR. Volume-assured pressure support ventilation (VAPSV): anew approach for reducing muscle workload during acute respiratoryfailure. Chest 1992;102(4):1225-1234.

26. Wikipedia contributors. Feedback. http://en.wikipedia.org/wiki/feedback. Accessed November 3, 2010.

27. Younes M. Porportional assist ventilation, a new approach to venti-latory support. Theory. Am Rev Respir Dis 1992;145(1):114-120.

28. Chatburn RL. Classification of ventilator modes: update and pro-posal for implementation. Respir Care 2007;52(3):301-323.

29. Guttmann J, Eberhard L, Fabry B, Bertschmann W, Wolff G. Con-tinuous calculation of intratracheal pressure in tracheally intubatedpatients. Anesthesiology 1993;79(3):503-513.

30. Elsasser S, Guttmann J, Stocker R, Mols G, Priebe HJ, Haberthur C.Accuracy of automatic tube compensation in new-generation me-chanical ventilators. Crit Care Med 2003;31(11):2619-2626.

31. Unoki T, Serita A, Grap MJ. Automatic tube compensation duringweaning from mechanical ventilation: evidence and clinical impli-cations. Crit Care Nurse 2008;28(4):34-42.

32. Sinderby C, Navalesi P, Beck J, Skrobik Y, Comtois N, Friberg S, etal. Neural control of mechanical ventilation. Nat Med 1999;5(12):1433-1436.

33. Chatburn RL, Wear JL. Comparison of two different proportional assistventilation algorithms (abstract). Respir Care 2009;54(11):1583.

34. Tehrani FT. Critiquing treatment and setting ventilatory parametersby using physiological modeling. Conf Proc IEEE Eng Med Biol Soc2009;2009:286-288.

35. Tzavaras A, Weller PR, Spyropoulos B. A neuro-fuzzy controller forthe estimation of tidal volume and respiration frequency ventilatorsettings for COPD patients ventilated in control mode. Conf ProcIEEE Eng Med Biol Soc 2007;2007:3765-3768.

36. Kwok HF, Linkens DA, Mahfouf M, Mills GH. SIVA: a hybridknowledge-and-model-based advisory system for intensive care ven-tilators. IEEE Trans Inf Technol Biomed 2004;8(2):161-172.

37. Lozano S, Moller K, Brendle A, Gottlieb D, Schumann S, Stahl CA,Guttmann J. AUTOPILOT-BT: a system for knowledge and modelbased mechanical ventilation. Technol Health Care 2008;16(1):1-11.

38. Wang A, Mahfouf M, Mills GH, Panoutsos G, Linkens DA, GoodeK, et al. Intelligent model-based advisory system for the manage-ment of ventilated intensive care patients: hybrid blood gas patientmodel. Comput Methods Programs Biomed 2010;99(2):195-207.

39. Rees SE, Allerød C, Murley D, Zhao Y, Smith BW, Kjaergaard S, et al.Using physiological models and decision theory for selecting appropri-ate ventilator settings. J Clin Monit Comput 2006;20(6):421-429.

40. Sundaresan A, Yuta T, Hann CE, Chase JG, Shaw GM. A minimalmodel of lung mechanics and model-based markers for optimizingventilator treatment in ARDS patients. Comput Methods ProgramsBiomed 2009;95(2):166-180.

41. Wikipedia contributors. Adaptive control http://en.wikipedia.org/wiki/adaptive_control. Accessed November 3, 2010.

42. Mireles-Cabodevila E, Diaz-Guzman E, Heresi GA, Chatburn RL.Alternative modes of mechanical ventilation: a review for the hos-pitalist. Cleve Clin J Med 2009;76(7):417-430.

43. Branson RD, Chatburn RL. Controversies in the critical care setting.Should adaptive pressure control modes be utilized for virtually allpatients receiving mechanical ventilation? Respir Care 2007;52(4):478-488.

44. Taniguchi C, Eid RC, Saghabi C, Souza R, Silva E, Knobel E, et al.Automatic versus manual pressure support reduction in the weaningof post-operative patients: a randomised controlled trial. Crit Care2009;13(1):R6.

45. Hewlett AM, Platt AS, Terry VG. Mandatory minute volume. An-esthesia 1977;32(2):163-169.

46. Stengel RF. Optimal control and estimation. Mineola, NY: DoverPublications; 1994.

47. Tehrani FT. Method and apparatus for controlling an artificial respi-ratory. US patent 4,986,268, issued January 22, 1991.

48. Tehrani FT. Automatic control of an artificial respirator. Proc IEEEEMBS Conf 1991;13:1738-1739.

49. Otis AB, Fenn WO, Rahn H. Mechanics of breathing in man. J ApplPhysiol 1950;2(11):592-607.

50. Wikipedia contributors. Intelligent control. http://en.wikipedia.org/wiki/intelligent_control. Accessed November 3, 2010.

51. Rose L, Presneill JJ, Cade JF. Update in computer-driven weaning frommechanical ventilation. Anesth Intensive Care 2007;35(2):213-221.

52. Lellouche F, Mancebo J, Jolliet P, Roeseler J, Schortgen F, Dojat M,et al. A multicenter randomized trial of computer-driven protocol-ized weaning from mechanical ventilation. Am J Respir Crit CareMed 2006;174(8):894-900.

53. Rose L, Presneill JL, Johnston L, Cade JF. A randomised, controlledtrial of conventional versus automated weaning from mechanicalventilation using SmartCare/PS. Intensive Care Med 2008;34(10):1788-1795.

54. Maldonado PA, Díaz UC, Unzueta AS. SmartCare closed-loopsystem and the altitude problem. Intensive Care Med 2009;35(4):753.



Discussion

Epstein: I like the idea of your rank-ing system. You seem to make theassumption that all those boxes areequivalent, that there’s no weightingof importance, so gas exchange andprotecting the lung and liberation areall the same. It may be true, but it alsomay not be true, and it also may de-pend upon the patient. Clearly, pro-tecting the lung is important in sev-eral kinds of patients, but not in allkinds of patients.

Chatburn: That’s exactly true. Theproblem is that once you get into thatkind of multivariate decision analysisand weaning factors based on patientconditions, it becomes unmanageable. Ijust wanted to show that there was ageneralized rational approach in rank-ing these things, and it would be an in-teresting academic exercise to see howfar you could carry that. In fact, youmight even come up with an artificial-intelligence program to run it based onthe situation and the things you felt wereimportant in that particular case.

MacIntyre: I want to follow up onScott’s point, because I think it’s in-credibly important. A number of yearsago at a conference, Art Slutsky said,“When you think about it, in manag-ing ventilator synchrony there are re-ally only 4 parameters you’re tryingto balance: plateau pressure, PO2

, pH,and FIO2

. How you weigh those 4 thingswill determine your optimal ventilatorstrategy.” There were about 15 of usin the room at the time, and he askedus to rank—let’s take FIO2

as the ref-erence—what is “equitoxic” to an FIO2

of 100%? Is a plateau pressure of 25,30, 35, 40, 45 cm H2O equitoxic? WhatPO2

is equitoxic: 55, 60, 70 mm Hg?And so on. He submitted this to thepeople around the room, and the re-sults were all over the chart. Therewas absolutely no consistency or con-sensus. Everybody’s got in their braina weighing system for those 4 param-eters, and the ranking system depends

on the relative weights in your valuesystem. I applaud you for coming upwith an interesting schema, but it’s go-ing to be very difficult to implement.

Chatburn: Keep in mind that theranking system describes what venti-lators do, not how they are used. It’slike saying I have a hammer, a screw-driver, and a saw, and now you haveto tell me how you want to use themaccording to the most important goalat the moment. We are starting to seeartificial intelligence used to balancethe parameters you mentioned, suchas in SmartCare, but that particularsystem uses only one tool: pressuresupport. Systems that make expert de-cisions with more tools, such as Ham-ilton’s Intellivent, are starting to gen-erate patient data.

Arnal et al reported at this year’sATS [American Thoracic Society]meeting that, compared to AdaptiveSupport Ventilation, patients venti-lated with Intellivent spent more timein optimal and suboptimal ventilationand less time in nonsecure ventilation.In addition, Intellivent delivered lowervolumes and pressures for equivalentgas-exchange results, suggesting moreefficient ventilation.1 Also at thatmeeting, Lellouche, et al reported thatIntellivent seemed safer than conven-tional ventilation after cardiac sur-gery.2

1. Arnal JM, Wysocki M, Demory D, DurishG, Laubscher T, Novotni D, et al. Prospec-tive randomized crossover controlled studycomparing Adaptive Support Ventilation(ASV) and a fully close loop control solu-tion (Intellivent) in adult ICU patients withacute respiratory failure (abstract). Am JRespir Crit Care Med 2010;181(Suppl):A3004.

2. Lellouche F, Bouchard PA, Wysocki M,Laubscher T, Novotni D, Lopez R, et al. Pro-spective randomized controlled study com-paring conventional ventilation versus a fullyclosed-loop ventilation (Intellivent) in postcardiac surgery ICU patients (abstract). Am JRespir Crit Care Med 2010;181(Suppl):A6035.

Younes:* I may appear to be pro-moting my own PAV [proportional as-sist ventilation], but if you look at thesystems that emphasize safety, theycannot provide synchrony with the pa-tient. But the systems that provide syn-chrony, such as NAVA [neurally ad-justed ventilatory assist] and PAV canbe made to be safe by adding backupfeatures and so on, and they can bemade to provide liberation by havingalgorithms that adjust, whereas theother ones cannot do that. So it seemsto me that the rational thing to do is totake the ones that can do what theothers cannot do and build up the otherfeatures that would give them thesafety and the ability to liberate thepatient.

Chatburn: Thank you. That was ex-actly what I was hoping someonewould say. You take the best featureson that grid and see which ones wecan combine to make the next gener-ation of better modes. I think that’show things should evolve; theyshouldn’t evolve haphazardly just be-cause someone thinks they can sell aventilator. They should evolve with atleast some crude sense of evolution asto what’s better and what’s not. I thinkthat’s an excellent way to look atthings.

Pierson:† So you’re not advising usto manage our patients on multiple dif-ferent ventilators concurrently?

Chatburn: On the contrary, I wouldargue for standardization. Pick one andlearn how to use it to the best of yourability.

Kacmarek: He is recommending toget rid of the [Puritan Bennett] MA1s.

Branson: Rob, you and I have spentmore time than anybody talking about

* Magdy Younes MD FRCP(C) PhD, Depart-ment of Medicine, University of Manitoba, Win-nipeg, Manitoba, Canada.



terminology and I’ve had this discus-sion with you before. The term “adap-tive” is explanatory or descriptive,whereas “intelligent” or “optimum”are, I think, value statements. I under-stand where you got the words from,but I think you’re in danger of appear-ing to promote or provide a value basedon the name, and I wonder if there isnot another way to describe those tech-niques with descriptive words thatdon’t suggest inherent value. Most ofus, when we say “intelligent” we gen-erally think that’s good, right?

Chatburn: You can argue that bothways, Rich. I didn’t make up thosewords: they’re used in artificial intel-ligence and engineering, and in thebigger world those words don’t havethe emotional connotation that you’retalking about. Should we bring our-selves down to arbitrary, “non-emo-tional” words or should we bring our-selves up to contemporary standardsof terminology? Should we go out andrename things that have already beenestablished in chemistry and biologyand pharmacology just because wedon’t like those words? I would argueno. Why do that? We already have toomuch gobbledygook, so let’s stick withsomething that at least has some con-text, some larger system of sciencebehind it. I’m not in favor of invent-ing new words to keep some peopleemotionally comfortable.

Branson: I’m not uncomfortablewith it. I just think the word “intelli-gent” may be perceived by some read-ers to be better than “adaptive.”

Chatburn: I recently did this lecturein another country. Before I gave it,my sponsors asked me to take out theranking section because they wereafraid the culture there would takesomething like that literally, as yousuggest. So you’re right that in certaincontexts you have to take into consid-eration how people will interpret thewords. I was hoping mostly, believe itor not, to speak to the engineers. Be-

cause if the engineers can describe tous what they’re doing in a systematicand rational fashion, and put that inthe operating manuals, then the rest ofus don’t have to struggle so hard tounderstand what their products do. Inthe printed version of this talk I willmake very clear what the words intel-ligent and adaptive mean, along withtheir engineering context. I trust thatreaders will be sophisticated enoughto grasp their proper meanings.

Gentile: When you showed us allthose modes on the screen, it mademe think that we have been takingthem apart. We test them on the benchor somewhere else, and we have ourown ranking. With SmartCare it’s end-tidal CO2 monitoring that we don’t doanywhere else: we have this adapterinline. We’ve already proven that themore times you break the ventilatorcircuit the higher the risk of ventila-tor-associated pneumonia. Neil[MacIntyre] and I did a bench studyon ASV [adaptive support ventilation]and found that when these resistance/compliance problems surface, thetidal volumes can increase dramati-cally. And that may be applicable formany other modes. These new modescome along and, until they’re proven,we are all very skeptical. I think we’reall a little long in the tooth, and wealso have to decide on an endpoint:ventilator-liberation, safety, or somesimilar measure.

Branson: Rob, so ASV, which Ithink works perfectly fine—and,again, if you told me tomorrow that Ihad to use ASV on all patients, I thinkI could successfully accomplish that—but ASV has been around since 1994.Why isn’t it more widely accepted andused? I feel the same way about a lotof these things: if they’re so much bet-ter, wouldn’t we have all figured itout by now?

Chatburn: You might be right aboutthat. The other piece is that, as we’vetalked about before, even the most ba-

sic advances in medicine take a longtime to be accepted into practice, evenwhen the evidence is staring us rightin the face. There’s an emotional com-ponent regarding resistance to change,and then there’s the intellectual com-ponent of “where’s the evidence?”

Gentile: I think the answer is thatit’s all proprietary. We have model A,model B, and model C, and they allhave different features on them. Wecan’t do PAV on a Servo-i, and wecan’t do NAVA on a Puritan Bennett,so I think that’s going to be some-thing we’ll have to live with.

Chatburn: First of all, Mike, I wasjust in China, and I can assure youthey use the same kinds of ventilatorswe do and with just as rich an assort-ment in a single department. In fact,their range is wider than ours. I actu-ally saw a modern Japanese high-fre-quency ventilator sitting right next toan original Emerson iron lung, bothbeing commonly used with patients.

Rich, I think part of the adoptionthing of course comes down to bud-get, but it’s not only, “Do I have themoney to buy the latest thing?” It’s“Do I have the will to buy the latestthing?” With a large purchase (half amillion to a million dollars, say), peo-ple tend to buy products they’re fa-miliar with. It’s human nature. It hasnothing to do with the intrinsic valueof the technology.

Branson: I would still submit that if15 years ago a revolutionary techniquehad been introduced that made patientsget better, it would have been evidentby now. It just goes back to the factthat, depending on whether your pa-tient needs to be oxygenated or ven-tilated because of respiratory acidosisor supported because of respiratorymuscle fatigue, the mode I wouldchoose would be different in every pa-tient. I think therein lies the problemof why one thing doesn’t get adopted.



Parthasarathy: Even with propri-etary modes there are instances wheredifferent companies have been com-ing together. In sleep-disorderedbreathing, Respironics and ResMedgot together to fund a multicenter trial.It’s unfortunate that the CANPAP[Canadian Continuous Positive Air-way Pressure for Patients with Cen-tral Sleep Apnea and Heart Failure]trial1 was negative, but I think the in-vestigators learned from that, andthey’re moving on with the AD-VENT-HF [Multi-Centre, Random-ized Study to Assess the Effects ofAdaptive Servo Ventilation on Sur-vival and Frequency of Cardiovascu-lar Hospital Admissions in Patientswith Heart Failure and Sleep Apnea,http://www.controlled-trials.com/isrctn/pf/67500535] trial now.

These kinds of discussions are great,because when these get published thefield should move in that direction,and I think one of the impediments isfunding in critical care research. Whatwe’re talking about here is a very in-homogenous population, and whatworks for one group may not work foranother, or may be harmful to another.It can only answered by a network,and that’s the kind of agenda that asociety such as the AARC [AmericanAssociation for Respiratory Care] orthe ATS [American Thoracic Society]should push for.

The biggest criticism I got from thereviewers of a [recent] grant applica-tion was that critical illness is a verycomplex condition with multiple vari-ables and that sleep is also a very com-plex condition with multiple variables,and you’re trying to do research onboth of them. The concern is that crit-ical care does not have a review groupof its own in the NIH [National Insti-tutes of Health]. Mine went to a panelof psychologists, who looked at myapplication kindly.

I think the situation is such that thereshould be something done at a differ-ent level to actually propel critical careresearch forward, to make it so thateven individual proprietary companies

can come together so that we can getthese questions answered. Otherwise,in 25 years will we still be sitting inthis same room talking about the sameissues, just with a more complex clas-sification system?

1. Bradley TD, Logan AG, Kimoff RJ, SeriesF, Morrison D, Ferguson K, et al; CanPAPInvestigators. Continuous positive airwaypressure for central sleep apnea and heartfailure N Engl J Med 2005;353(19):2025-2033.

Kallet: When you showed that slideon optimizing WOB [work of breath-ing], it made me think. WOB has beenmy focus all these years, and you getinto this frame of mind, almost sub-consciously, of “you gotta minimizeit, you gotta minimize it!” But it oc-curred to me recently, would we havesurvived as a species if we weren’table to at least double our WOB forextended periods of time? Granted,critical illness is a special context, butI think it illustrates how our assump-tions might influence the design of so-phisticated ventilator modes in a waythat might not be beneficial. Wheredo you factor that in? As closed-loopventilation becomes more complex,could our preoccupation with tightlycontrolling some targets lead us downthe wrong path?

Rich Branson and I have talked overthe years about how rigidly to controlplateau pressure in trauma patientswith decreased chest-wall compliance.Do you need to keep the plateau at30 cm H2O all the time? When do youloosen that up? Same goes for PEEP/FIO2

titration. I think when we discussmaking a perfect mode, there’s somuch context to consider that it canalso befuddle the design, particularlyif we rigidly adhere to rules such askeep the plateau pressure at30 cm H2O, WOB [lt] 0.5 J/L, FIO2

below 0.7, or whatever. I don’t knowwhere we go with that.

Branson: It reminds me of normal-izing other things in critical care. Whathappens when we normalize the he-

matocrit? Patients don’t do better, theydo worse. Think of all the things thatwe make normal and maybe it’s not in-telligent to normalize the WOB to whatit is in a patient with normal complianceand resistance, I just don’t know.

Epstein: We need to know whetherthese things really cause a bad out-come or if instead they’re just mark-ers. If they’re markers, then a lot ofwhat we’re talking about in terms ofchanging the way we ventilate patientsmay not be relevant. We need to knowwhether this is cause or effect.

Branson: I have to believe, beingsomeone who’s been scuba diving, thatif you try to take a breath and youdon’t get any air, it’s a frightening,distressing situation.

Epstein: But we’re not talking aboutnot getting any air, we’re talking aboutmaybe only getting 85% of the air youmight want. What is the thresholdwhere it makes a clinical difference? Idon’t think we know.

Branson: But I think a missed trig-ger is more like that. With a missedtrigger they get a little bit of gas out ofthe circuit, but for the most part arenot rewarded with a breath. It seemsthat has to be distressing and uncom-fortable and lead to asynchrony. I thinkthat, but I don’t know.

Epstein: I agree with the discomfortpart. I’m looking at the different kindsof outcomes in terms of the duration ofmechanical ventilation, and survival.Some data show it’s uncomfortable.

Kacmarek: There is a difference be-tween asynchrony and WOB. The pa-tient can be synchronous with the ven-tilator and we can still titrate the levelof work or effort they do. WOB andsynchrony should not be used synony-mously. At this conference we’re talk-ing about the ventilator not working inconjunction with the patient’s desires—not necessarily saying we should un-



load the patient completely. We have awhole spectrum of levels of unloadingof the patient, but what we desire is thatthe ventilator and the patient work insynchrony as much as possible, regard-less of what level of work we considerfor that given patient.

Branson: I guess what I’m saying isthat if your mode targets a work, thenyou are making that a goal.

Epstein: You’re making the goal thatyou want to set the work at X level,but it’s not a specific level: it can be atany level you choose. It can be a dif-ferent level in different patients or ina given patient, depending on the pa-tient’s progress.

Branson: I think I agree with you,Bob, but I’m saying that if you choosea mode that has a goal of maintainingWOB at X, then no matter what thepatient does, it tries to maintain WOBat X. All I’m saying is, how do weknow that level X is good or bad?

Epstein: We don’t.

Branson: And we see patients withCOPD get extubated with a WOB of1.0 J/L who do fine.

Epstein: Sure: once you get the en-dotracheal tube out the work changespretty dramatically.

Younes: I want to emphasize that badinteractions happen only when we arenot awake. When we are awake, wedo not allow poor interactions: we tryto enforce the good interactions. If wefail to do that, we get panicky and usesedation, and I think the link betweenthis topic and outcome has to do withsedation. If you’re using a mode thatbecomes asynchronous when you’reasleep, then it would automaticallylead to sedation when you’re awake,because the patient would not tolerateit. We all know that excessive seda-tion contributes to bad outcomes, andI think we may need to look at what isassociated with the most sedation andthe least sedation.

Parthasarathy: We’ve observed thattoo. On these patients, electroencepha-lography more likely than not showsthat the non-triggering events occurduring sleep, because they won’t al-low it when they’re awake. We had ahypothesis that non-triggering is asso-ciated with the disruption from sleep,but we couldn’t find it, because whenthe patients were non-triggering, theywere usually asleep. There were veryfew non-triggering breaths when theywere awake. It seems that if they re-ally need to get that breath, they willfind the effort to get that breath. Theywon’t be in the circumstance wherethey’re trying to breathe and they don’tget a breath while they’re awake.

Pierson: It’s like I’ve been at a de-bate among the mechanical ventila-tion gods on respiratory care’s MountOlympus, and it has been a headyexperience. This conference has sev-eral goals, but I hope one thing thatwill come out of it is something thatcan spill down the slopes of themountain to the readership in termsof practical advice on what we cando about patient-ventilator interac-tion, right now, with the ventilatorswe have. I’m hoping we don’t comeaway with the conclusion that youhave to have a particular brand orkind of machine to really do me-chanical ventilation right.

And I’m also hoping that onedoesn’t have to have clinicians ofthe expertise that we have in thisroom to take good care of patientson ventilators. I think we might allagree that mechanical ventilation aspracticed on an everyday basis fallsshort in terms of both knowledge ofthe current evidence and the practi-cal application of that evidence: evenwith what we know for sure, we knowthat it doesn’t get used. In additionto this splendid discussion of the is-sues we’re dealing with, some ofthem admittedly more theoreticalthan practical, I hope we will alsocome out with as much as we can interms of practical guidance on whatto do about these things.



closed-loop control of mechanical ventilation: description...

Documents