Performance ofTransport Ventilators

The paper by Chipman et al in the June2007 issue of RESPIRATORY CARE representeda monumental undertaking to evaluate 15transport ventilators.1 We reviewed the pa-per on several occasions to reconcile theirfindings with our work on transport venti-lators.2-9 We write to point out several in-accuracies in the study and ask Chipmanet al to more clearly explain their findingsand recommendations.

In their Table 1, Chipman et al list thepositive end-expiratory pressure (PEEP) ofthe LTV1000 as 0–30 cm H2O. The actualrange is 0–20 cm H2O. The table also liststhe PEEP of the Crossvent 3 as 0 cm H2O,but the Crossvent 3 is capable of PEEP upto 35 cm H2O. Were Chipman et al refer-ring to an earlier model of the Crossvent 3that couldn’t supply PEEP? Also, the Vor-tran is listed as being capable of providingcontinuous mandatory ventilation (CMV)and intermittent mandatory ventilation(IMV), but in fact the Vortran can provideonly pressure-cycled CMV. Similarly, thePercussionaire TXP provides only IMV, notCMV. With the Percussionaire TXP, if thepatient breathes spontaneously, air is en-trained from the inspiratory port of the Pha-sitron.

We are unclear as to why Chipman et allisted both assist-control and CMV for someventilators, as those 2 modes are the same.Did they mean to distinguish between ven-tilators that can be patient-triggered andthose that can deliver only mandatorybreaths? If so, then errors still remain. ThePneupac ParaPac ventilators have a modetermed SMMV (synchronized minimummandatory volume), not MMV (mandatoryminute volume) or IMV, which are quitedifferent, in that if the patient breathes spon-taneously at a respiratory rate equal to orgreater than the set rate, the ventilator doesnot provide any mandatory breaths.

Also the Crossvent 3 has a selection forpressure support but does not have the abil-ity to flow cycle. This mode results in pa-tient-triggered, pressure-limited, time-cycled support. The Crossvent 3 is listed ashaving both volume and pressure control,but in fact it is capable of only pressure-

limited ventilation via a mechanical pres-sure-relief valve. In that instance the venti-lator continues to deliver the set flow andvolume according to the volume control set-tings, but vents gas to the atmosphere if thepressure threshold is reached. These errorsmay have resulted from Chipman et al de-pending on the manufacturer’s literature fortheir data. However, since their intentionwas to compare available devices, we be-lieve these errors should be pointed out soreaders can determine the actual perfor-mance of the devices, not simply mimic themanufacturer’s specifications.

Chipman et al stated that in their benchprotocol the Impact/Uni-Vent Eagle 754 op-erated on an E-size cylinder for 35 min,which is substantially less than we have ob-served and is not consistent with that ven-tilator’s function.2 Was this test repeatedmore than once? This finding suggests ei-ther a leak in the system, a malfunctioningventilator, or unfamiliarity by the operator.Was the tidal volume (VT) continuouslymeasured to assure that the correct minutevolume was being delivered? As Chipmanet al are well aware from their experience,repetition of laboratory experiments is in-valuable in the detection of one-time errors.

In Table 2, Chipman et al concluded thatneither the Impact/Uni-Vent Eagle 754 northePulmoneticLTV1000arecapableofven-tilating injured lungs. As in our previouscomment, that conclusion is inconsistentwith the bench data and the worldwide ex-perience with those 2 devices. For over adecade the Impact/Uni-Vent Eagle 754 hasbeen successfully used to transport criticallyill soldiers (including many with acute re-spiratory distress syndrome) between mili-tary care centers. Similarly, the LTV1000 iscapable of volume and pressure control atup to 80 breaths/min, and is used by manycenters for transport of critically ill patients.We request that Chipman et al explain thenature of the failures with the Impact/Uni-Vent Eagle 754 and the PulmoneticLTV1000, and reconcile their findings withthe common experience.

Our review of the animal-model experi-ments in the Chipman et al study led us tounderstand that a single animal was used totest 5 separate ventilators, in random se-quence. Is it possible that the small number

of animals resulted in a ventilator being dis-advantaged in the trial because of the se-quence? As an example, Figure 2 indicatesthat each ventilator was used in 2 animalexperiments. What was the progression ofthe lung injury? If the first 4 ventilators failedto provide adequate oxygenation or venti-lation, would that worsen the lung injuryand thus make the fifth ventilator fail toprovide adequate gas exchange? Does theirsurfactant-wash-out model improve withtime, or does gas exchange steadily worsen?Also, the performance of every ventilator isaffected by the device characteristics, thecondition of the animal, and the understand-ing of the operator. It seems incongruentthat a given ventilator is capable of main-taining gas exchange in one animal, but notthe other, unless the ventilator model is sub-stantially different or the operator fails tomake the appropriate changes.

Finally, though Chipman et al concludedthat the Impact/Uni-Vent Eagle 754 is un-able to ventilate injured lungs, it is one ofthe ventilators they recommend in their con-clusions. Though we agree, and experienceclearly demonstrates that this recommenda-tion makes sense, how did they come to thisconclusion? At the end of their exhaustivestudy, how did Chipman et al determinewhich ventilators to recommend? Werecharacteristics ranked or weighted? Waseachventilatorgivenascore forperformancein each of the evaluations? Though we agreethat both the Impact/Uni-Vent Eagle 754and Newport HT50 meet the requirementsfor a “front-line ventilator for rescue situa-tions,” we are not sure why the LTV1000and VersaMed iVent do not.10 Did Chip-man et al eliminate the LTV1000 becauseof its high gas consumption? Was the Ver-saMed eliminated because of short batterylife and excessive weight? We request thatChipman et al describe the system they usedto come to their conclusions. If gas con-sumption, size, and battery life were the onlyfactors considered, the animal experimentsseem unnecessary.

We appreciate the substantial effort byChipman et al in this project and their com-mitment to provide the respiratory care com-munity with much-needed data. However,we believe they should correct some errorsand explain the findings that seem to run

1372 RESPIRATORY CARE • OCTOBER 2008 VOL 53 NO 10

Letters

Letters addressing topics of current interest or material in RESPIRATORY CARE will be considered for publication. The Editors may accept or

decline a letter or edit without changing the author’s views. The content of letters as published may simply reflect the author’s opinion or

interpretation of information—not standard practice or the Journal’s recommendation. Authors of criticized material will have the opportunity

to reply in print. No anonymous letters can be published. Letters should be submitted electronically via Manuscript Central. Log onto

RESPIRATORY CARE’s web site at http://www.RCJournal.com. Instructions on how to submit a manuscript are on the site and also printed in every

issue.

counter to our experience and commonsense.

Richard D Branson MSc RRT FAARCDivision of Trauma and Critical Care

Department of SurgeryUniversity of Cincinnati Medical Center

Cincinnati, Ohio

SMSgt Dario Rodriquez RRTCenter for Sustaintment of Trauma

Readiness SkillsUnited States Air Force

Cincinnati, Ohio

Mr Branson has received honoraria and/or re-search support from Cardinal, Drager, and Co-vidien. The authors report no other conflicts ofinterest related to the content of this letter.

REFERENCES

1. Chipman DW, Caramez MP, Miyoshi E,Kratohvil JP, Kacmarek RM. Performancecomparison of 15 transport ventilators. Re-spir Care 2007;52(6):740-751.

2. Lyons E, Johannigman J, Davis K, TsueiB, Branson R. Portable ventilators: a com-parison of duration of operation from anE-cylinder (abstract). Respir Care 2006;51(11):1292.

3. Austin PA, Campbell RS, Johannigman JA,Branson RD. Work of breathing character-istics of seven portable ventilators. Resus-citation 2001;49(2):163-172.

4. Austin PN, Johannigman JA, Simmons LW,Camden MP, Campbell RS, Branson RD.Surface temperature of two portable venti-lators during simulated use under clinicalconditions. Mil Med 2001;166(10):843-847.

5. Barnes SL, Branson R, Gallo LA, Beck G,Johannigman JA. En-route care in the air:snapshot of mechanical ventilation at37,000 feet. J Trauma 2008;64(2 Suppl):S129-S135.

6. Salas N, Wisor B, Agazio J, Branson RD,Austin PA. Comparison of ventilation andcardiac compressions using the Impact 730automatic transport ventilator compared toa conventional bag valve with a facemaskin a model of adult cardiopulmonary arrest.Resuscitation 2007;74(1):94-101.

7. Stevenson VW, Haas CF, Wahl WL. Intra-hospital transport of the adult mechanicallyventilated patient. Respir Care Clin N Am2002;8(1):1-35.

8. Austin PN, Campbell RS, Johannigman JA,Branson RD. Impsed work of breathing dur-ing ventilator failure. Respir Care 2002;47(6):667-674.

9. Campbell RS, Johannigman JA, BransonRD, Austin PN, Matacia G, Banks GR. Bat-tery duration of portable ventilators: effects

of control variable, positive end-expiratorypressure, and inspired oxygen concentra-tion. Respir Care 2002;47(10):1173-1183.

10. American Association for Respiratory Care.Guidelines for acquisition of ventilators tomeet demand for pandemic flu and masscasualty incidents. http://www.aarc.org/head l ines /ven t i l a to r_acqu is i t ions /vent_guidelines.pdf. Accessed July 28,2008.

The authors respond:

We thank Branson and Rodriquezfor their letter.1 We agree there weremistakes in one of our tables that wemissed during editing, and that thereis considerable disagreement about ter-minology. However, we also believethat most of the issues they raise werecovered in the paper.

Branson and Rodriquez are correctthat the positive end-expiratory pres-sure (PEEP) ranges for the Pulmon-etic LTV1000 and Crossvent 3 are0–20 cm H2O and 0–35 cm H2O, re-spectively, and we thank them forpointing this out. We do, however, takeexception to what they describe as in-accuracies with respect to various ven-tilation modes.

It is unfortunate that no standard-ized nomenclature exists for classify-ing ventilation modes. Ventilator man-ufacturers use various names todescribe their devices’ modes. Con-sider 2 commonly used ventilators, thePuritan Bennett 840 and the DragerEvita 4. The Evita 4 offers a contin-uous-mandatory-ventilation mode thatallows timed or patient triggering ofconstant-volume breath delivery, aswell as the PCV � Assist mode thatallows timed or patient triggering ofpressure-controlled breaths.2 Both ofthose modes may be defined as a formof assist/control ventilation (ie, allbreaths are delivered by the ventilatorat a set minimum rate [mechanicalbreaths], but the patient may triggerthe ventilator and thus cause a rategreater than the set rate).3 Contrast thisto controlled mechanical ventilation,in which all breaths are delivered bythe ventilator at a set rate and patienttriggering is not allowed.

With the Puritan Bennett 840 theclinician must first select a ventilationmode, and we will again use the ex-ample of assist/control, followed bythe selection of a mandatory breathtype, such as volume control or pres-sure control.4

In the neonatal arena, the DragerBabylog has a mode in which the ven-tilator delivers a set number of non-synchronized, time-cycled, pressure-limited breaths, and the patient isallowed to breath spontaneously in be-tween the breaths from a clinician-ad-justable, continuous gas flow. Al-though we have known this ventilationmode for many years as intermittentmandatory ventilation (IMV), on theventilator the manufacturer describesit as CMV.5

We did not evaluate any of theseventilators during spontaneous venti-lation, which we clearly identified asa major limitation of our study. Con-sequently, we relied on the manufac-turers’ documentation regarding trig-gering and spontaneous breathingcapabilities.

Several years ago, Chatburn andPrimiano, in an attempt to promotestandardization of nomenclature, pub-lished an extensive description of ven-tilation modes.6 However, no consen-sus on the use of that nomenclaturehas been established. Neither the en-gineers who design ventilators, themarketing and sales people, nor clini-cians use Chatburn and Primiano’s no-menclature. In the absence of such con-sensus, it has been our practice todescribe ventilation modes as con-trolled, assist/control, synchronizedintermittent mandatory ventilation(SIMV), or spontaneous, and thebreath-delivery types as volume-con-trolled, pressure-controlled and/orpressure-supported. In our paper weintentionally disregarded the manufac-turers’ descriptions of modes and ap-plied the above classification system.Thus, since the Percussionaire deliv-ers a set number of mechanical breathsto a patient making no spontaneousrespiratory efforts, it was classified as

LETTERS TO THE EDITOR

RESPIRATORY CARE • OCTOBER 2008 VOL 53 NO 10 1373

a controlled mode. And because spon-taneous efforts are met with unsup-ported air entrainment, it was also clas-sified as having an IMV mode. Weused similar logic in classifying theVortran. It functions in a controlledmode with a patient who does not havea spontaneous respiratory rate. How-ever, upon further inspection, despitethe manufacturer’s claim that “spon-taneous breathing patients may entrainroom air,” the one-way valve is pres-surized during normal operation anddoes not allow spontaneous breathing.It functions only as a fail-safe valve,allowing spontaneous breathing ofroom air if gas flow to the system islost (eg, empty cylinder).

The Parapac devices provide a modeidentified as synchronized mandatoryminute ventilation (SMMV), which isintended to allow spontaneous breath-ing but also provides mandatorybreaths at the set tidal volume (VT) inthe event the patient does not meet theminute volume requirement. To morecarefully evaluate this, we set the tidalvolume at 1,000 mL and the respira-tory rate at 10 breaths/min. If the pa-tient has a small spontaneous VT

(� 200 mL), the ventilator deliversthe set VT at the set rate (IMV). If thespontaneous VT is � 200 mL but lessthan the set VT, the ventilator deliversthe set VT at a rate lower than the setrate. And if the spontaneous breathsequal the set VT, no mandatory breathsare delivered. This is consistent withthe manufacturer’s description: “Thetidal volume required to completelyinhibit the ventilator is fixed at 450mL, but the frequency is determinedby that set on the ventilator.”7

Regarding the Crossvent 3, a VT

and pressure limit are set. If the pres-sure limit is above the pressure neededto deliver the set VT, the set volume isdelivered every breath (volume venti-lation)! However, if the pressure is setlower than the pressure needed to de-liver the set VT, the excess volume isvented to the atmosphere and the setpressure is held for the remainder of

the inspiratory time (pressure ventila-tion)! In no case did we indicate thatany ventilator provides both controlledmechanical ventilation and assist/con-trol ventilation, as Branson and Ro-driquez indicated.

Gas consumption was defined as theamount of time the ventilator func-tioned on one full E-size oxygen cyl-inder (capacity 660 L of oxygen) at aVT of 1,000 mL and a respiratory rateof 10 breaths/min, on 100% oxygen.We performed this test once with eachventilator. In theory the maximum timeof operation would be 66 min, assum-ing there were no leaks in the system,no bias flow or other diversion of gasby the ventilator, 100% of the gas wasdevoted to minute volume, the cylin-der contents were correct, and the ven-tilator functioned as expected through-out the test.

As is well known, the actual con-tent of oxygen cylinders differs con-siderably. However, the reason for theless-than-expected duration of opera-tion of the Impact/Uni-Vent Eagle 754was its inability to maintain set pa-rameters during this aspect of the eval-uation. A peculiar finding with the Im-pact/Uni-Vent Eagle 754 was thatbefore the cylinder became depleted,the ventilator alarmed “O2 LOW/FAIL” and substantially decreased theVT of alternate breaths. We identifiedthis as the point at which the ventila-tor was unable to maintain set param-eters.

To more carefully address Bransonand Rodriquez’s concern, we recentlyrepeated this portion of the test withthe Impact/Uni-Vent Eagle 754, with3 different ventilators, at the previ-ously described settings, and with andwithout PEEP set at 5 cm H2O. Allcylinders were verified to be full, withpressure range 2,000–2,300 psi. Pre-dicted duration of operation was 60–69 min. All 3 ventilators exhibited sim-ilar changes in VT delivery, but atvarious cylinder pressures. The rangeof duration of normal operation was22– 61 min. The cylinder-pressurerange when the malfunction began was

300–1,100 psi, and the delivered VT

of alternate breaths decreased to ap-proximately 50% of the set value. Con-tinued operation eventually resulted infurther variance in delivered VT in allbreaths. Performance was unaffectedby the addition of PEEP.

Regarding the animal studies, all theventilators that failed to ventilate theinjured lung model failed because ofthe development of auto-PEEP, whichprevented further increase in the rate.We should have explained this in moredetail. The lung model we used is thesame that had been used by us in nu-merous other animal studies and is sta-ble for longer than 4 hours, which ex-ceeded the time needed to evaluatethe group of 5 ventilators.8-10 Couldthe sequence of ventilators evaluatedhave affected their ability to ventilateand the development of auto-PEEP?Absolutely. This is in fact one of thereasons we did not use this failure asa “strike” against a ventilator in thefinal analysis

In the introduction of our paper welisted the 6 criteria we used to selectthe most suitable ventilator for use inthe out-of-hospital setting; we focusedon forward military positions. As westated above, because of the size ofsome of the injured animals, we dis-regarded the failure to ventilate injuredlungs in our final evaluation. The spe-cific criteria that separated the New-port HT50 and Impact/Uni-Vent Ea-gle 754 from the VersaMed iVent andPulmonetic LTV1000 was the dura-tion of the internal battery. We did notconsider the attachment of additionalbattery capabilities, and we stated thatboth the VersaMed iVent and Pulmo-netic LTV1000 would also be consid-ered at the same level as the NewportHT50 and Impact/Uni-Vent Eagle 754if their internal battery life was longer.

Daniel W Chipman RRTRobert M Kacmarek PhD RRT

Respiratory Care ServicesMassachusetts General Hospital

Boston, Massachusetts

LETTERS TO THE EDITOR

1374 RESPIRATORY CARE • OCTOBER 2008 VOL 53 NO 10

Dr Kacmarek has been a consultant for SpaceLaboratories, and has received research grantsand/or honoraria from Puritan Bennett, Maquet,Cardinal Health, Newport Medical, HamiltonMedical, Respironics, and Drager. Mr Chipmanhas been a consultant for and has received lec-ture honoraria from Maquet. The authors reportno other conflicts of interest related to the con-tent of this paper.

REFERENCES

1. Chipman D, Caramez MP, Miyoshi E, Kra-tohvil J, Kacmarek RM. Performance com-parison of fifteen transport ventilators. Re-spir Care 2007;52(6):740-751.

2. Operating instructions. Drager Evita 4 in-tensive care ventilator, 7th edition. Telford,Pennsylvania: Drager; 2004;70-78.

3. Martinez-Ruis R, Bigatello RM, Hess D.Mechanical ventilation. In: Critical CareHandbook of the Massachusetts General

Hospital, 3rd edition. Hurford WE, editor.Philadelphia: Lippincott Williams andWilkins; 2000:87-88.

4. 840 ventilator system operator’s and tech-nical reference manual, part 4-075609-00revision C. Carlsbad, California: NellcorPuritan Bennett; 1999:4.3.

5. Babylog 8000 infant care ventilator opera-tor’s manual, software release 3. Chantilly,Virginia: Drager; 1993.

6. Chatburn RL, Primiano FP Jr. A new sys-tem for understanding modes of mechani-cal ventilation. Respir Care 2001;46(6):604-621.

7. Parapac Transport 200D Ventilator user’smanual. PN 504-1117A_NUS, Issue 2, 07/2003, Draft B, Pneupac Ltd. Bedfordshire,England, page 31.

8. Sedeek KA, Takeuchi M, Suchodolski K,Vargas SO, Shimaoka M, Schnitzer JJ,

Kac-marek RM. Open-lung protectiveventilation with pressure control ventila-tion, high-frequency oscillation, and in-tratracheal pulmonary ventilation resultsin similar gas exchange, hemodynamics,and lung mechanics. Anesthesiology 2003;99(5):1102-1111.

9. Caramez MP, Schettino G, Suchodolski K,Nishida T, Harris RS, Malhotra A, Kac-marek RM. The impact of endotracheal suc-tioning on gas exchange and hemodynam-ics during lung-protective ventilation inacute respiratory distress syndrome. RespirCare 2006;51(5):497-502.

10. Caramez MP, Miyoshi E, Harris RS, Kac-marek RM, Malhotra A. Gas exchange im-pairment induced by open suctioning inacute respiratory distress syndrome: impactof permissive hypercapnia. Crit Care Med2008;36(2):560-564.

LETTERS TO THE EDITOR

RESPIRATORY CARE • OCTOBER 2008 VOL 53 NO 10 1375