753Thorax 1990;45:753-758

Assisted ventilation 1 Series editor: John Moxham

Artificial ventilation: history, equipment andtechniques

J D Young, M K Sykes

A brief history of artificial ventilators

Nuffield Departmentof Anaesthetics, JohnRadcliffe Hospital,OxfordJ D YoungM K SykesAddress for reprint requests:Dr J D Young,Nuffield Department ofAnaesthetics, John RadcliffeHospital, Oxford OX3 9DU.

An artificial ventilator is essentially a devicethat replaces or augments the function of theinspiratory muscles, providing the necessary

energy to ensure a flow of gas into the alveoliduring inspiration. When this support isremoved gas is expelled as the lung and chestwall recoil to their original volume; exhalationis a passive process. In the earliest reports ofartificial ventilation this energy was providedby the respiratory muscles of another person,

as expired air resuscitation. Baker' has tracedreferences to expired air resuscitation in thenewborn as far back as 1472, and in adultsthere is a report of an asphyxiated miner beingrevived with mouth to mouth resuscitation in1744. In the eighteenth century artificial ven-

tilation became the accepted first line treat-ment for drowning victims, though the use ofbellows replaced mouth to mouth resus-

citation.2 Automatic artificial ventilators thatdid not require a human as a power source

took another 150 years to appear, being firstsuggested by Fell3 and then made availablecommercially by Draeger4 in 1907. Thesewere still devices for resuscitation, for theDraeger company at that time was noted forits mine rescue apparatus.The introduction of artificial ventilators

into anaesthesia proceeded slowly until sur-

geons ventured into the chest. Duringthoracic surgery on spontaneously breathingpatients the inevitable pneumothoraces andmediastinal shift ("mediastinal flap") thatoccurred as the pleural cavity was openedcaused a high mortality, which was

substantially reduced when positive pressure

ventilation was used. A further boost to thedevelopment of automatic artificial ventilatorsoccurred in 1952, when a catastrophicpoliomyelitis epidemic struck Denmark.There was a very high incidence of bulbarlesions and 316 out of 866 patients withparalysis admitted over a period of 19 weeksrequired postural drainage, tracheostomy, or

respiratory support. By using tracheostomyand manual positive pressure ventilation theDanish physicians reduced the mortality frompoliomyelitis from 80% at the beginning ofthe epidemic to 23% at the end. The artificialventilation was entirely by hand, a total of1400 university students working in shifts tokeep the patients ventilated. The fear thatanother epidemic might afflict Europe

expedited research into powered mechanicalventilators.

Classification of artificial ventilatorsThe lungs can be artificially ventilated eitherby reducing the ambient pressure around thethorax (negative pressure ventilation) or byincreasing the pressure within the airways(positive pressure ventilation). Negative pres-sure ventilators use a rigid chamber thatencloses either the thorax (cuirass) or thewhole body below the neck (tank respirator or"iron lung"). The pressure in the chamber isreduced cyclically by means of a large volumedisplacement pump, thus causing the lungs toexpand and contract. Negative pressure ven-tilation is fully discussed in article 5 of thisseries. These ventilators were used extensivelyfor poliomyelitis victims, and are still in usefor long term respiratory support or overnightsupport for patients with respiratory muscleweakness, Tank ventilators occupy muchspace, access to the patient is poor, and theneck seal can create problems. They are notsuitable for use in general intensive care units.There has been a recent revival in interest innegative pressure ventilators in paediatricintensive care units to avoid the need forendotracheal intubation.5The basic classification of positive pressure

artificial ventilators was first proposed byMapleson.6 Artificial ventilators are devicesthat control inspiration; expiration is usuallypassive, and so the classification is based onthe mechanism of gas delivery during inspira-tion. There are two types of machine. A flowgenerator produces a known pattern of gasflow during inspiration, and the lungs fill at arate entirely controlled by the ventilator andindependent of any effect of lung mechanics.A pressure generator produces a preset pres-sure in the airway and the rate of lung infla-tion depends not only on the pressure gen-erated by the ventilator but also on the res-piratory resistance and compliance, whichdetermine the time constant of the lungs.With a square wave pattern of pressure thelungs fill in an exponential fashion. Theeffects of the two types of ventilator on ins-piratory flow and lung volume are shown infigure 1.

In general, the flow generator ventilator isused for adults and the pressure generatorventilator for children, or adults when controlof peak airway pressure is important. The

on 6 July 2018 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.45.10.753 on 1 October 1990. D

ownloaded from

Young, Sykes

C)

0

._co

La

U)

0

Figure 1 Lung volume and gasflow during one respiratory cycle with pressure andflowgenerator ventilators.

pressure generator is particularly useful inchildren where uncuffed endotracheal tubesare used and there is a leak of gas around theendotracheal tube during inspiration. A pres-sure generator tends to compensate for thisleak by increasing the flow into the airway,whereas with a flow generator a proportion ofthe tidal volume is lost.A ventilator must also have a mechanism to

cause it to cycle between inspiration andexpiration. Ventilators are usually subclas-sified as time cycled, pressure cycled, orvolume cycled machines. Time cycled ven-tilators switch between inspiration and expira-tion after a preset time interval, pressurecycled ventilators switch when a preset airwaypressure threshold has been reached, andvolume cycled ventilators cycle when a presettidal volume has been delivered. Cycling fromexpiration to inspiration is usually effected bya timing mechanism or by a patient triggeringdevice that senses the subatmospheric pres-sure or the flow generated in the inspiratorytube by the patient's inspiratory effort.

This traditional classification was devisedduring a period when ventilators were totallymechanical, and driven either by compressedgas or by an electrically powered piston orbellows. By 1954, however, Donald had usedan electronic trigger that initiated inspirationand in 1958 an electronic timing device wasincorporated in the Barnet ventilator.7 In 1971the Siemens-Elema company introduced theServo 900 ventilator, which combined a sim-ple pneumatic system with a sophisticatedelectronic measuring and control unit.8 Gasflow to and from the patient's lungs was

controlled by a pair of scissor valves andmonitored with pressure and flow sensors.The control unit adjusted the scissor valves toensure that the flow patterns measured by thesensors corresponded to those selected by theoperator. This method of control is termed aservo or feedback system and is very flexible.Potentially one machine could mimic all thepreviously described categories of ventilator.

Modern intensive care unit ventilatorsMost modern intensive care unit ventilatorsuse similar technology. A simplified diagram ofsuch a ventilator is shown in figure 2. Therespiratory gas is held in a pressurised reservoirand delivered to the patient via an inspiratoryvalve. The inspiratory valve and hence theinspiratory flow is controlled by the electroniccontrol unit. The airway pressure and flow ofgas into the patient are monitored by thepressure and inspiratory flow sensors. Theexpiratory flow can also be monitored to checkfor leaks and disconnection of the patient fromthe ventilator. This design enables the ven-tilator to be used as either a flow or a pressuregenerator. For example, if the operator selects aconstant flow during inspiration the ventilatorwill open the inspiratory valve until the flowsensor measures the required flow. If theinspiratory flow decreases the inspiratory valvewill be opened further to compensate, and viceversa. If the operator wishes the ventilator toact as a constant pressure generator, the ven-tilator will open the inspiratory valve until thepressure sensor indicates that the desired pres-sure has been reached. The ventilator will thenmaintain the airway pressure at the desired

754 on 6 July 2018 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.45.10.753 on 1 O

ctober 1990. Dow

nloaded from

Artificial ventilation: history, equipment and techniques

Topatient

Figure 2 A block diagram of a modern ventilator. The shaded area indicates the gaspath during inspiration.

level by opening or closing the inspiratoryvalve. The bulk of a modern ventilator is nowthe electronic unit and the pneumatic units areoften very compact. Figure 3 is a photograph ofa Servo 900D ventilator. The pneumatic sys-tems are housed in the top section and thebottom section contains only electronics.The use of electronic feedback systems in

ventilators has many advantages. The movingparts within the ventilator are kept to a min-imum, sophisticated alarm systems are possi-ble, the ventilators can be made physicallysmall, and maintenance and repair are simple.If a new mode of ventilation appears to beuseful clinically it can often be added to anexisting machine by altering the software in theelectronic unit-indeed, our ability to producenew modes of artificial ventilation far exceedsour ability to test them clinically.

Interactions between patients andventilatorsAs early as 1929 it was observed that a patientwho "fought" the ventilator was difficult to

manage and suffered complications.9 To min-imise these problems many ventilated patientsare sedated, and in some cases paralysed withneuromuscular blocking agents (for example,pancuronium). In the early 1980s paralysis ofventilated patients was very common in Britishintensive care units, 1011 and sedation was usedto dissociate the patients from their surround-ings. Over the last decade the amount ofsedative and paralysing drugs given to inten-sive care patients has been reduced consider-ably.

Paralysis of ventilated patients is not withoutrisks.'2 The paralysed patient is unable to makeany spontaneous respiratory effort, so a discon-nection from the ventilator is rapidly fatal. Thepatient may be aware ofhis or her surroundingsbut is unable to communicate, and the risk ofpulmonary embolus is probably increased.Even sedative drugs are not without risk,'3 asshown by the substantial increase in mortalityin trauma patients given etomidate infusions.This was subsequently shown to be caused bysuppression of adrenal activity by the drug.Only a few clinicians now favour deep seda-

tion and paralysis,'4 the preferred approachbeing a comfortable patient who can be easilyaroused and can communicate with the staff.To achieve this goal ventilators have to be moreflexible and adapted to the needs of the patient.In its simplest form this is achieved by insert-ing a valve into the ventilator tubing, whichallows the patient to breathe spontaneouslyfrom another gas source between the breathsdelivered by the ventilator. This mode ofventilation is termed intermittent mandatoryventilation (IMV) and was initially used as atechnique to wean patients from mechanicalventilation to spontaneous respiration. By1987, however, over 70% of American inten-sive care units were using intermittent man-datory ventilation as their primary mode ofventilatory support."' It is not a truly interac-tive mode ofventilation-the ventilator contin-ues to deliver the preset pattern of ventilationregardless of the patient's respiratory effort.The advent of microprocessor controlled ven-tilators has produced several "smart" modes ofventilation, where the ventilator adjusts theventilatory support it supplies according to thepatient's respiratory efforts. The commonestmodes used are synchronised intermittentmandatory ventilation, mandatory minute ven-tilation, and inspiratory pressure support.

a

Figure 3 A Siemens Servo 900D ventilator.

Ventilatory strategies in common use

INTERMITTENT POSITIVE PRESSURE VENTILATIONEach year millions of patients are ventilatedduring anaesthesia for surgical procedures thatrequire muscular relaxation. Nearly all of thesepatients are ventilated with basic intermittentpositive pressure ventilation by means of asimple mechanical ventilator. For these shortperiods of respiratory support these ventilatorsare perfectly adequate, for as the patient is bothparalysed and anaesthetised there is no need forinteractive ventilators. Weaning is accompli-

755 on 6 July 2018 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.45.10.753 on 1 O

ctober 1990. Dow

nloaded from

Young, Sykes

Figure 4 "Stacking" ofbreaths during intermittentmandatory ventilation.The arrow indicates wherea mandatory breath startsbefore the patient hasfullyexpired.

LA.

0

0

co0E0

-j

Time

shed by reversing the muscle relaxation andlightening the anaesthetic. In intensive care

units this mode of ventilation is used when thepatient has to be heavily sedated and paralysedto treat the primary condition (for example,tetanus) or when the patient is unable to makeany respiratory movement (for example, severe

Guillain-Barre syndrome). Intermittentpositive pressure ventilation can be achievedwith almost any artificial ventilator as no

special refinements are needed.

INTERMITTENT MANDATORY VENTILATIONAs outlined above, intermittent mandatoryventilation describes a mode of mechanicalventilation that allows the patient to breathespontaneously through the ventilator circuit.At predetermined intervals a positive pressurebreath is provided by the ventilator totallyindependently of the patient's spontaneousventilatory pattern. The system is a simplemechanical device that can be fitted to almostany ventilator, though care has to be taken toensure that the oxygen concentration is thesame in the spontaneous breaths as during thebreaths delivered by the ventilator and that theinspired gas is appropriately humidified.'6Though simple it is not without disadvantages.There is no synchronisation between thepatient's respiratory efforts and the ventilator,and this can lead to "stacking" of breaths andhigh airway pressures. Figure 4 shows a plot oflung volume against time for a patient on

intermittent mandatory ventilation. Thepatient can take a spontaneous breath, andbefore he has fully exhaled the ventilatordelivers a breath. A very large tidal volumewith high peak airway pressure results. Toovercome this problem ventilators were

developed that were able to detect spontaneousbreaths taken by the patient.

SYNCHRONISED INTERMITTENT MANDATORY

VENTILATIONTo avoid "stacking" the ventilator must be ableto sense that the patient has taken a breath, andthen avoid delivering a mandatory breath dur-ing the period of the spontaneous breath. Thislevel of sophistication was not easily achievedbefore electronic control was used in artificialventilators. One of the first machines to have a

synchronised intermittent mandatory ventila-tion facility was the Servo 900B, a refinementofthe original Servo 900. This machine dispen-

sed with external intermittent mandatory ven-tilation valves and used its internal valves andsensors to provide synchronised intermittentmandatory ventilation. Any spontaneous in-spiratory activity by the patient was sensed bythe pressure sensor and then the expiratoryvalve was closed and the inspiratory valveopened. The flow of gas through the in-spiratory valve was matched to the patient'sinspiratory flow rate by opening and closing theinspiratory valve. This allowed the patient tobreathe spontaneously through the ventilator.When a mandatory breath was due the ven-tilator would wait until the patient began toinspire and then deliver the mandatory breath,synchronising the mandatory breath with thepatient's own inspiratory effort.

All ventilators with synchronised intermit-tent mandatory ventilation systems use someform of synchronisation between patient andmachine, the exact details varying betweenmachines. All synchronised intermittent man-datory ventilation machines are not equallyeffective, for the time taken for the inspiratoryvalve to open after the pressure sensor has beentriggered varies between machines. During theperiod between the beginning of the patient'sinspiration and the opening of the inspiratoryvalve the patient inspires against a closed valve,increasing the work of breathing.'7 Machineswith a short lag between the beginning of aninspiratory effort and the opening of theinspiratory valve are preferred. Althoughsubjectively patients usually appear morecomfortable with synchronised intermittentmandatory ventilation than intermittentpositive pressure ventilation or intermittentmandatory ventilation, there is no evidencethat it reduces morbidity or mortality inintensive care units.

MANDATORY MINUTE VENTILATIONThe synchronised intermittent mandatoryventilation systems described add a presetnumber of breaths of a preset volume to thepatient's own respiratory efforts. If thepatient's spontaneous breaths are providingmost of the required minute ventilation, themandatory breaths may be too large or toofrequent. Alternatively, if the patient's ownrespiratory efforts diminish the mandatorybreaths may not be enough to provide anadequate minute ventilation. Ideally the ven-tilator should monitor exhaled volumes and"top up" the patient's respiratory efforts ifneeded. This is mandatory minute ventilation.In this mode of ventilation a preset minutevolume is selected. If the patient's own res-piratory efforts produce a minute ventilationequal to or greater than the preset value theventilator is not activated. If the preset minutevolume is not achieved the ventilator givessome assistance to the patient, to bring theminute ventilation back to the target value.This system works well if the patient's res-piratory rate is relatively normal. If, however,the patient has rapid, shallow breaths theefficiency of breathing will be reduced and themandatory volume may become inadequate.The method used to provide ventilatory

756 on 6 July 2018 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.45.10.753 on 1 O

ctober 1990. Dow

nloaded from

Artificial ventilation: history, equipment and techniques

assistance depends on the design of the ven-tilator. The Ohmeda CPU1 and EngstromErica will provide synchronised mandatorybreaths of a preset tidal volume at an increasingfrequency as the patient's own respiratoryefforts diminish. The Hamilton Veolar uses adifferent approach and provides increasinginspiratory pressure support (see below) as thepatient's own efforts decrease. Whichevermethod is used, the machines ensure that thepatient always receives a predetermined minutevolume.

INSPIRATORY PRESSURE SUPPORTThe most recent mode of ventilation to bewidely used is inspiratory pressure support. Asthe patient initiates a breath the ventilatorraises the airway pressure to a preset value. Thepositive airway pressure provides some of theenergy needed to expand the lung and theefforts of the patient provide the rest. At theend of inspiration the positive airway pressureis removed to allow unimpeded expiration. Byselection of an appropriate level of airwaypressure patients can be given only the res-piratory assistance they actually require. Thepatient also determines his or her own res-piratory rate. This is a relatively new mode ofventilation and has not yet been fully assessed.This form of ventilation differs from the assistmode commonly used in the 1960s. In the assistmode the magnitude ofthe breath was preset bythe adjustment ofcontrols on the ventilator andthe breath was initiated whenever the patienttrigger was activated. In inspiratory pressuresupport a constant pressure is applied after thepatient triggers the ventilator so that thepatient can determine the flow pattern and sizeof breath. Inspiration is terminated when theinspiratory flow ceases.

POSITIVE END EXPIRATORY PRESSUREPatients who are anaesthetised or comatose orwho have recently undergone abdominal sur-gery have a reduced functional residualcapacity. The reduction may be sufficient tocause airway closure in dependent areas of thelung before the end of expiration, leading tounderventilation of these areas and ventilation-perfusion mismatch.'8 This in turn leads tohypoxaemia. The functional residual capacitycan be maintained by leaving a constant stand-ing pressure on the lungs, keeping themslightly inflated even at the end of expiration.This is positive end expiratory pressure orPEEP. The use of PEEP, and how best todetermine its optimum level, has been thesubject of debate for many years. On thepositive side, it causes an increase in arterialoxygen tension for the same inspired oxygentension. On the negative side, the constantlyraised intrathoracic pressure causes a diminu-tion in venous return to the heart, a decrease incardiac output, and an increased risk ofpneumothorax. The controversy may now beresolved, for a recent study showed that mor-tality increased with the aggressive applicationof PEEP and that the minimum level whichkept arterial oxygenation within acceptablelimits was "best PEEP."'19

UNUSUAL MODES OF VENTILATIONThere are several modes of ventilation that arestill being evaluated clinically. In highfrequency jet ventilation brief, frequentlyrepeated pulses of gas (up to 300 a minute) aredirected from a high pressure nozzle down theairway. The tidal volumes delivered are small,airway pressures are low, and the patient canbreathe spontaneously while being ventilated.20High frequency oscillation techniques oscillategas in the airways using a piston pump ordiaphragm pump. The tidal volumes used arevery small, much less than the anatomical deadspace, so gas movement occurs by mechanismsother than tidal exchange. Because of the largesize of oscillators needed for adults this tech-nique has been confined to use in infants.There are also two extracorporeal systems inclinical use. In adults with acute respiratoryfailure extracorporeal membrane oxygenationwas abandoned after a large trial showed that itdid not reduce mortality.2' In infants withneonatal respiratory distress, however, themethod seems to be very effective. A remarka-ble reduction in mortality from this conditionhas been reported in America when extracor-poreal membrane oxygenation has been ins-tigated.22 In adults a new approach has beenpioneered by Gattinoni in Italy. He startedwith the hypothesis that artificial ventilationitself may exacerbate acute respiratory failure.Tidal ventilation is required only to removecarbon dioxide, and oxygenation can be main-tained in apnoeic patients if the carbon dioxideis removed by an extracorporeal circuit. Byremoving carbon dioxide with an extracor-poreal artificial lung and reducing ventilationto the minimum he has claimed to reducemortality from acute respiratory failure.23

ConclusionThe distinction between a ventilated patientand a spontaneously breathing patient isbecoming increasingly blurred as more sophis-ticated means of respiratory support are dev-ised. In many cases "respiratory assistance"may be a more appropriate term than "artificialventilation." The efficacy of many of theseventilatory strategies has not been properlyassessed, in terms either of acceptability to thepatient or of the effect on morbidity andmortality. It is still true, however, that "thetype of ventilator used, provided that its designis satisfactory, is of less importance than theexperience of the person using it."'8

1 Baker AB. Early attempts at expired air respiration, intuba-tion and manual ventilation. In: Atkinson RS, BoultonTB, eds. The history ofanaesthesia. London: Royal Societyof Medicine, 1987:372-4.

2 Kite C. An essay on the recovery of the apparently dead.London: Dilly, 1788.

3 Goerig M, Filos K, Ayisi KW. George Edward Fell and thedevelopment of respiratory machines. In: Atkinson RS,Boulton TB, eds. The history of anaesthesia. London:Royal Society of Medicine, 1987:386-93.

4 Rendell-Baker L, Pettis JL. The development of positivepressure ventilators. In: Atkinson RS, Boulton TB, eds.The history of anaesthesia. London: Royal Society ofMedicine, 1987:402-21.

757 on 6 July 2018 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.45.10.753 on 1 O

ctober 1990. Dow

nloaded from

Young, Sykes

5 Samuals MP, Southall DP. Negative extrathoracic pressure

in the treatment of respiratory failure in infants and young

children. Br Med J 1989;299:1253-7.6 Mapleson WW. The effect of lung characteristics on the

functioning of artifical ventilators. Anaesthesia1 962;31 :300-14.

7 Mushin WW, Rendell-Baker L, Thompson PW, MaplesonWW. Automatic ventilation of the lungs. Oxford: Black-wells, 1980:312-30.

8 Ingelstedt S, Johnson B, Nordstrom L, Olsson S-G. Aservo-controlled ventilator measuring expired minutevolume, airway flow and pressure. Acta AnaesthesiolScand 1972;supple 47:9-28.

9 Drinker P, McKhann CF. The use of a new apparatus forprolonged administration of artificial respiration. JAMA1929;92: 1658-61.

10 Miller-Jones, CMH, Williams JH. Sedation for ventilation.A retrospective study of fifty patients. Anaesthesia1980;35:1104-7.

11 Merriman HM. 'I'he techniques used to sedate ventilatedpatients. Intens Care Med 1981;7:217-24.

12 Willatts SM. Paralysis for ventilated patients? Yes or No?Intens Care Med 1985;11:2-4.

13 Watt I, Ledingham I. McA. Mortality amongst multipletrauma patients admitted to an intensive therapy unit.Anaesthesia 1984;39:973-81.

14 Bion JF, Ledingham I. McA. Sedation in intensive care-a

postal survey. Intens Care Med 1987;13:215-6.15 Venus B, Smith RA, Mathru M. National survey ofmethods

and criteria used for weaning from mechanical ventilation.Crit Care Med 1987;15:530-3.

16 Pybus DA, Kerr JH. A simple system for adminsteringintermittent mandatory ventilation (IMV) with theOxford ventilator. Br J Anaesth 1978;50:271-4.

17 Lemaire LF. Weaning from mechanical ventilation. In:Ledingham I, McA, ed. Recent advances in critical caremedicine 3. Edinburgh: Churchill Livingstone, 1988:15-30.

18 Sykes MK, McNicol MW, Campbell EJM. Respiratoryfailure. Oxford: Blackwells, 1976:320-2.

19 Carroll GC, Tuman KJ, Braverman B, et al. Minimalpositive end-expiratory pressure (PEEP) may be bestPEEP. Chest 1988;93:1020-5.

20 Smith BE, Hanning CD. Advances in respiratory support.Br J Anaesth 1986;58:138-50.

21 Zapol WM, Snider M'r, Hill JD, et al. Extracorporealmembrane oxygenation in severe acute respiratory failure.JAMA 1979;242:2193-6.

22 Toomasian JM, Snedcor SM, Cornell RG, Cilley RE,Bartlett RH. National experience with extracorporealmembrane oxygenation for newborn respiratory failure.Trans Am Soc Artif Intern Organs 1988;34:140-7.

23 Gattinoni L, Pesenti A, Mascheroni D, et al. Low-frequencypositive-pressure ventilation with extra-corporeal CO,removal in severe acute respiratory failure. JAMA1986;256:881-6.

24 Sykes MK, McNicol MW, Campbell EJM. Respiratoryfailure. Oxford: Blackwells, 1976:212.

758

on 6 July 2018 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.45.10.753 on 1 October 1990. D

ownloaded from