clinical ventilator adjustments that improve speech1

7/28/2019 Clinical Ventilator Adjustments That Improve Speech1

1/12

DOI 10.1378/chest.124.4.15122003;124;1512-1521Chest

Hixon and Robert Brown

Jeannette D. Hoit, Robert B. Banzett, Heather L. Lohmeier, Thomas J.

*SpeechClinical Ventilator Adjustments That Improve

http://chestjournal.chestpubs.org/content/124/4/1512.full.html

services can be found online on the World Wide Web at:The online version of this article, along with updated information and

ISSN:0012-3692)http://chestjournal.chestpubs.org/site/misc/reprints.xhtml(

written permission of the copyright holder.this article or PDF may be reproduced or distributed without the priorDundee Road, Northbrook, IL 60062. All rights reserved. No part ofCopyright2003by the American College of Chest Physicians, 3300

Physicians. It has been published monthly since 1935.is the official journal of the American College of ChestChest

2003 American College of Chest Physiciansby guest on December 29, 2010chestjournal.chestpubs.orgDownloaded from
http://chestjournal.chestpubs.org/content/124/4/1512.full.htmlhttp://chestjournal.chestpubs.org/content/124/4/1512.full.htmlhttp://chestjournal.chestpubs.org/content/124/4/1512.full.htmlhttp://chestjournal.chestpubs.org/site/misc/reprints.xhtmlhttp://chestjournal.chestpubs.org/site/misc/reprints.xhtmlhttp://chestjournal.chestpubs.org/http://chestjournal.chestpubs.org/http://chestjournal.chestpubs.org/http://chestjournal.chestpubs.org/http://chestjournal.chestpubs.org/site/misc/reprints.xhtmlhttp://chestjournal.chestpubs.org/content/124/4/1512.full.html


2/12

Clinical Ventilator Adjustments ThatImprove Speech*

Jeannette D. Hoit, PhD; Robert B. Banzett, PhD; Heather L. Lohmeier, MS;Thomas J. Hixon, PhD; and Robert Brown, MD

Study objectives: We sought to improve speech in tracheostomized individuals receiving positive-pressure ventilation. Such individuals often speak with short phrases, long pauses, and haveproblems with loudness and voice quality.Subjects: We studied 15 adults with spinal cord injuries or neuromuscular diseases receivinglong-term ventilation.Interventions: The ventilator was adjusted using lengthened inspiratory time (TI), positiveend-expiratory pressure (PEEP), and combinations thereof.Results: When TI was lengthened (by 8 to 35% of the ventilator cycle), speaking time increasedby 19% and pause time decreased by 12%. When PEEP was added (5 to 10 cm H2O), speakingtime was 25% longer and obligatory pauses were 21% shorter. When lengthened TI and PEEP

were combined (with or without reduced tidal volume), their effects were additive, increasingspeaking time by 55% and decreasing pause time by 36%. The combined intervention improvedspeech timing, loudness, voice quality, and articulation. Individual differences in subjectresponse to the interventions were substantial in some cases. We also tested high PEEP (15 cmH2O) in three subjects and found speech to be essentially identical to that produced with aone-way valve.Conclusions: These simple interventions markedly improve ventilator-supported speech and aresafe, at least when used on a short-term basis. High PEEP is a safer alternative than a one-wayvalve. (CHEST 2003; 124:15121521)

Key words: neurogenic communication disorders; quadriplegia; respiration; tracheostomy; ventilation, mechanical

Abbreviations: PEEP positive end-expiratory pressure; Pt tracheal pressure; Spo2 noninvasive measure ofarterial oxygen saturation; Te expiratory time; Ti inspiratory time; Vt tidal volume

The ability to speak has a profound influence onthe quality of life for individuals who depend on

mechanical ventilation.1,2 Speech makes it possiblefor such individuals to express needs efficiently,develop and maintain personal relationships, com-

municate with caregivers, use the telephone, andoperate systems controlled by speech recognitionsoftware. Those who have tracheostomies and re-ceive positive-pressure ventilation are usually able tospeak if the cuff on the tracheostomy tube is deflatedor if a fenestrated or cuffless tracheostomy tube isused.3,4 Nevertheless, speech produced in this way isoften far from satisfactory.

The settings on positive-pressure ventilators are

adjusted to accommodate cardiopulmonary require-ments, but they are not optimal for speech. Speechproduced with typical ventilator adjustments is oftencharacterized by short phrases, long pauses betweenphrases, abnormal loudness, and poor voice quali-ty.5,6 These speech problems can be attributed, inlarge part, to the nature of the tracheal pressure (Pt)waveform. To produce speech, Pt must be at least 2cm H2O to vibrate the vocal folds,7,8 and it should berelatively constant to maintain steady loudness andnormal voice quality.9,10When speech is produced inable-bodied individuals, Pt is nearly constant at 5 to

1 0 c m H2O throughout expiration.11,12

In sharp

*From the Department of Speech and Hearing Sciences andNational Center for Neurogenic Communication Disorders (Drs.Hoit, Dr. Hixon, and Ms. Lohmeier), University of Arizona,Tucson, AZ; Physiology Program (Dr. Banzett), Harvard School

of Public Health, Boston, MA; and Pulmonary Section (Dr.Brown), Veterans Administration Boston Healthcare System,Boston, MA.Drs. Banzett and Brown are currently affiliated with the Pulmo-nary and Critical Care Unit, Massachusetts General Hospital,Boston, MA.Support was provided by National Institute on Deafness andOther Communication Disorders grants DC-03425 and DC-01409.Manuscript received September 30, 2002; revision acceptedMarch 12, 2003.Reproduction of this article is prohibited without written permis-sion from the American College of Chest Physicians (e-mail:[email protected]).Correspondence to: Jeannette D. Hoit, PhD, Department ofSpeech and Hearing Sciences, PO Box 210071, University of

Arizona, Tucson, AZ 85721; e-mail: [email protected]

1512 Clinical Investigations in Critical Care

http://chestjournal.chestpubs.org/http://chestjournal.chestpubs.org/http://chestjournal.chestpubs.org/http://chestjournal.chestpubs.org/


3/12

contrast, the Pt during volume-controlled, positive-pressure ventilation rises rapidly to 20 cm H2Oduring inspiration, then falls rapidly to zero duringexpiration and remains there until the next inspira-tion. Thus, Pt is too low to vibrate the vocal folds formuch of the ventilator cycle, and when it is highenough, it changes so rapidly that it is impossible to

maintain constant loudness and normal voice quality.One remedy would be to cycle the ventilator so asto optimize Pt for speech production while simulta-neously accommodating cardiopulmonary require-ments and comfort. This approach involves straight-forward adjustments that are within the capabilitiesof most clinical positive-pressure ventilators. Suchadjustments are designed to improve speech bymodifying the Pt waveform to: (1) maintain Pt abovethe minimum required for voicing for a longerperiod so that more speech can be produced perbreath and less time need be spent in mandatory

silence, and (2) reduce the rate of change of Pt toallow loudness and voice quality to be more even.We and others have demonstrated the feasibility ofthis general approach.6,13 The present study extendsprevious work by examining two single-adjustmentinterventions and a combined-adjustment interven-tion, and by including an increased range of ventila-tor adjustments.

Materials and Methods

We studied 15 subjects with spinal cord injuries or neuromus-

cular diseases (Table 1), who lived in extended-care facilities orat home. Thirteen subjects received ventilation with volume-controlled positive pressure and produced speech with thetracheostomy tube cuff deflated or with a cuffless tracheostomy

tube (either fenestrated or unfenestrated). All but one subject(subject 2) routinely maintained a deflated cuff throughout the

waking hours. Four subjects (subjects 2, 5, 7, and 13) routinelyused one-way inspiratory valves for speaking (Passy-Muir Trache-ostomy Speaking Valve, Passy-Muir, Irvine, CA; and HudsonRC1, Hudson, Temecula, CA). Five subjects (subjects 1, 5, 6, 10,and 15) actively triggered the ventilator to increase breathingfrequency when speaking with their usual ventilator settings. Theremaining two subjects (subjects 8 and 9) had used volume-controlled, positive-pressure ventilators in past years, but at thetime of the study were routinely using phrenic nerve pacers(combined with one-way valves) for ventilation and speechproduction. The study protocol was approved by all appropriatehuman subjects committees, and informed consent was obtainedfrom all subjects.

Subjects were studied using a standard ventilator (ServoVentilator 900C; Siemens-Elema; Solna, Sweden) with settingsmatched as closely as possible with those on the subjects own

ventilator. These settings are termed the usual condition. Next,ventilator adjustments were made (with one-way valves re-moved), including the following: (1) lengthening inspiratory time(Ti), (2) applying positive end-expiratory pressure (PEEP), and(3) combining lengthened Ti and PEEP.

Lengthened Ti can improve speech produced during theinspiratory phase of the ventilator cycle (Fig 1, top left, a). Withlengthened Ti, air flows through the larynx longer so that Ptremains above the voicing threshold longer during inspiration.Also, the flow is lower so that the rate of rise of Pt is reduced.

When speaking during usual expiration, nearly all the air in thelungs flows toward the ventilator because the ventilator pathwayoffers much lower impedance (ie, primarily resistance) to flowthan does the laryngeal pathway (Fig 1, top right, b). PEEPimpedes expiratory flow and adds a threshold occlusion pressureto the ventilator expiratory line so that more air flows through thelarynx than toward the ventilator (Fig 1, bottom left, c). Thus, Ptstays above the voicing threshold longer during expiration than

without PEEP (as long as the impedance offered by the larynx is

adequately high, the usual case during speech production). Aone-way valve shunts all expired air through the larynx byoccluding the ventilator line (Fig 1, bottom right, d).

The specific levels of Ti and PEEP were determined individ-

Table 1Subject Information

SubjectNo.* Sex

Age,yr Primary Etiology

Duration ofVentilation, yr

ArticulatoryImpairment

HearingImpairment

1 Male 34 Duchenne muscular dystrophy 11 Mild None2 Male 29 Spinal cord injury (C1) 3 None None

3 Male 69 Guillain-Barre syndrome 4 Mild Mild4 Female 45 Amytrophic lateral sclerosis 3 Profound None5 Male 20 Dejerine-Sottas disease 8 None Moderate to severe6 Male 35 Spinal cord injury (C1) 9 None None7 Male 52 Spinal cord injury (C2) 5 None Mild8 Male 35 Spinal cord injury (C3/4) 13 None None9 Male 51 Spinal cord injury (C2) 13 None None

10 Female 51 Limb-girdle muscular dystrophy 13 None None11 Female 41 Arteriovenous malformation 18 Mild None12 Male 33 Duchenne muscular dystrophy 4 Moderate None13 Male 41 Spinal cord injury (C2) 5 None None14 Male 72 Postpoliomyelitis syndrome 1 None Moderate15 Male 40 Duchenne muscular dystrophy 8 Mild None

*Subjects 8, 9, 10, 11, and 12 participated in previous studies of respiratory function (coded previously as subjects 6, 7, 10, 11, and 14,

respectively).5,6,1417

www.chestjournal.org CHEST / 124 / 4 / OCTOBER, 2003 1513



4/12

ually for each subject by briefly testing a range of levels. The levelthat was most comfortable for the subject was used for the study.It was not possible to test all subjects on all interventions due totime limitations imposed by nursing staff schedules, subjectfatigue, subject preferences (eg, rejection of lengthened Ti), andthe fact that several subjects also participated in additionalinterventions not reported here.

With each condition, we recorded several breaths with noseand mouth closed (except in the four subjects whose usualcondition included use of a one-way valve). The subject read a

paragraph19 aloud, and then was asked How does your speechsound? and How does your breathing feel? The subjectresponded by using a rating scale ( 2much worse than usual; 1 slightly worse than usual; 0 usual; 1 slightly betterthan usual; 2much better than usual).

The speech signal was sensed by a head-mounted microphone(C451EB; AKG Acoustics; Wein, Austria). Pt was sampled usinga polyethylene catheter inserted through the sealed port of aswivel adapter (Portex 0803; Concord/Portex; Keene, NH) andadvanced through the tracheostomy tube to a point just within itsproximal end. The catheter was connected to a transducer(Validyne MP45 with 50 cm H2O diaphragm; Validyne Engi-neering; Northridge, CA) and amplifier (Validyne MC 13). TidalPco2 was sampled from the common ventilator line and mea-

sured with a clinical monitor (Cardiocap II; Datex Engstrom;

Helsinki, Finland). Noninvasive measure of arterial oxygen satu-ration (Spo2), heart rate, and noninvasive BP were monitoredcontinually. A digital audiotape recorder (PC208A; Sony; Tokyo,Japan) was used to record the speech signal, Pt, end-tidal Pco2,and ventilator flow.

Ten objective measures were computed (Macintosh Quadra950, LabVIEW Software; National Instruments; Austin, TX)[Table 2]. Auditory perceptual analysis was conducted by fivelisteners, all of whom were certified speech-language patholo-gists. Each listener was presented pairs of speech samples, one

with the subjects usual ventilator settings and one with anintervention. Listeners were asked to mark on a visual analogscale whether the second sample was better, worse, or the sameas the first (usual) sample, and provide descriptors. The scale wascentered around zero (same as usual), and ranged from 2(much worse than usual) to 2 (much better than usual).

An intervention was deemed successful if it improved objectivemeasures and/or listener ratings of speech and if Spo2 remained 90% (and did not decrease by 3%), heart rate and BP were

within the expected range, and subjects reported that breathingwas comfortable. In the case of subjects who used one-way valves,an intervention was deemed successful if the speech was as goodas with the one-way valve. Statistical inferences were made withpaired t tests (objective measures) or one-sample t tests (listener

ratings). An

level of 0.05 was used.

Figure 1. Air flow during ventilator-supported speech production. The black circles representocclusions, and the gray circle represents higher-than-usual impedance. During inspiration (top left, a),air flows both toward the lungs and through the larynx. During usual expiration (top right, b), almostall air flows toward the ventilator. This is because the impedance of the ventilator pathway is muchlower than that of the laryngeal pathway during speech production. During expiration with PEEP(bottom left, c), the impedance of the ventilator pathway is higher than usual so that more air flowsthrough the larynx. During expiration with a one-way valve (bottom right, d), all air flows through the

larynx. Adapted from Hoit et al.18




5/12

Results

Speech improved in 12 of our 15 subjects with one ormore of the interventions. Results of successful inter-ventions are presented below according to type. Un-successful cases are considered separately thereafter.

Lengthened TI

Six subjects (subjects 1, 3, 5, 7, 14, and 15) weretested with lengthened Ti (lengthened by 8 to 35%of the cycle) compared to their usual Ti (Table 3).

Expiratory time (Te) decreased by the same amountin subjects who did not actively trigger the ventilator(subjects 3, 7, and 14). For those who did trigger, Te

did not necessarily decrease because they triggeredless often than usual.

Lengthened Ti increased actual and potentialspeaking time (mean, 19% for both) and numberof syllables produced per breath (mean, 21%).This increase was entirely attributable to an increasein syllables produced during the inspiratory phase ofthe cycle (mean, 58%). No significant changes

Table 2Objective Measures

Measures Definition Significance Goal

Pause time Time (in s) per cycle during which speech is notproduced (ie, silent time)

Silent pauses between speech phrasesinterrupt the flow of speech.

Decrease

Speaking time Time (in s) per cycle during which speech isproduced

Short speech phrases make itnecessary to pause at nonlinguisticjunctures.

Increase

Syllables/breath No. of syllables produced per breathing cycle Syllables are units of speech.Syllables/breath may or may notcovary with speaking time.

Increase

Syllables/inspiration No. of syllables produced per inspiratory phase ofthe cycle

Syllables/inspiration are expected toincrease with lengthened Ti.

Increase

Syllables/expiration No. of syllables produced per expiratory phase ofthe cycle

Syllables/expiration are expected toincrease with PEEP.

Increase

Articulation rate Speed at which speech is produced (in syllables/s)calculated for speaking time only (ie, pausetime deleted)

Changes in articulation rate canchange syllables/breathindependent of changes inspeaking time.

No change

Sound pressure level Average amplitude (in decibels) of the (acoustic)speech signal

Sound pressure level is the acousticcorrelate of loudness (an auditorypercept).

No change

Peak Pt Highest Pt (in cm H2O) within a cycle (occurs at

end-inspiration)

Peak Pt is usually high and should

not be made higher withinterventions.

No change or

decrease

Potential speaking time Time (in s) per cycle during which Pt is at least 2cm H2O

Pt of at least 2 cm H2O is requiredto maintain vocal fold vibration.

Increase

Breathing frequency Breathing frequency (in breaths/min) Breathing frequency can change if the active triggering frequencychanges.

No change

Table 3Group Means of Measures Associated With Usual Ventilator Settings and With the Intervention ofLengthened TI, Mean Differences Between Usual and Intervention Measures, and Results of t Tests

Measure Mean Usual Mean Intervention Mean t p Value

Pause time, s 1.63 1.43 0.20 1.39 0.223Speaking time, s 2.62 3.10 0.48 10.13 0.001*Syllables/breath 10.23 12.33 2.10 5.64 0.002*Syllables/inspiration 4.48 7.07 2.59 5.81 0.002*Syllables/expiration 5.73 5.28 0.45 1.19 0.287Articulation rate, syllables/s 3.88 4.00 0.12 1.23 0.272Sound pressure level, decibels 71.18 70.13 1.05 1.09 0.324Peak Pt, cm H2O 20.87 19.42 1.45 2.03 0.098Potential speaking time, s 3.23 3.80 0.57 5.94 0.002*Breathing frequency, breaths/min 14.75 13.57 1.18 1.79 0.133Listener ratings 0 0.15 0.15 0.70 0.515

*

p

0.05.




6/12

were found in pause time, syllables/expiration, artic-ulation rate, sound pressure level, breathing fre-quency, or peak Pt. Listeners rated subjects speechwith lengthened Ti as generally better than usual,but this was not statistically significant. Nearly allsubjects rated their own speech as sounding better(four subjects) or the same (one subject) as usual,

and breathing as feeling better (three subjects) orthe same (two subjects) as usual.Thus, objective measures of speech improved with

lengthened Ti. The primary improvement was anincrease in speech produced during inspiration.

PEEP

Eight subjects (subjects 1, 3, 6, 8, 9, 11, 12, and14) who did not use one-way valves were tested withlow levels of PEEP (5 to 10 cm H2O) compared tono PEEP (Table 4). Minimum Pt equaled the PEEPsetting when the nose and mouth were occluded,

and it usually fell near (or to) zero during speakingtrials.

With PEEP, pause time decreased in all subjects(mean, 21%), and speaking time increased in mostsubjects (mean, 25%). All subjects produced moresyllables per breath (mean, 25%). This was dueentirely to more syllables being produced during theexpiratory phase of the ventilator cycle (mean, 67%). Pt remained above the voicing thresholdlonger (ie, potential speaking time increased; mean, 32%), which explains why subjects were able toproduce more speech. PEEP did not significantly

alter syllables/inspiration, articulation rate, soundpressure level, peak Pt, or breathing frequency.Listeners rated subjects speech with PEEP as gen-erally better than usual, but this was not statisticallysignificant. Nearly all subjects rated their ownspeech as sounding better (five subjects) or the same(two subjects) as usual, and their breathing as feelingbetter (five subjects) or the same (two subjects) asusual.

Thus, objective measures revealed that speechimproved with PEEP. Improvement was primarily inthe form of shorter pauses and more speech pro-duced during expiration.

Lengthened TI Plus PEEP (Plus Reduced TidalVolume)

Six subjects (subjects 3, 6, 11, 12, 14, and 15) weretested with the combination of lengthened Ti(lengthened by 17 to 25% of the cycle) and lowPEEP (5 to 10 cm H2O) [Table 5]. For subjects 11and 15, tidal volume (Vt) was also reduced slightly(by 0.1 to 0.2 L) because they commented that thepressure felt too high.

These combined adjustments resulted in shorterpause time (mean, 36%), longer speaking time(mean, 55%), and more syllables per breath(mean, 61%). The increase in syllable productionoccurred during both the inspiratory phase (mean,

45%) and expiratory phase (mean, 96%) of theventilator cycle in all subjects. Potential speakingtime increased by about the same amount as actualspeaking time (mean, 52%). Articulation rate,sound pressure level, and peak Pt did not change.Breathing frequency did not change significantly;however, the two subjects who actively triggered theventilator (subjects 6 and 15) reduced their breath-ing frequency with the combined adjustments. Lis-teners rated speech as better than usual for allsubjects (this was statistically significant). Their de-scriptors indicated improvements in timing (pause

time, speaking time), loudness (overall loudness,loudness variation), voice quality (overall voice qual-ity, voice quality variation), and articulation (articu-lation precision). Subjects rated their speech andbreathing as better (four subjects) or the same (twosubjects) as usual.

Speech improvements were in the form of bothobjective and subjective measures. Specific improve-ments were shorter pauses, more speech produced

Table 4 Group Means of Measures Associated With Usual Ventilator Settings and With the Intervention of AddingPEEP, Mean Differences Between Usual and Intervention Measures, and Results of t Tests

Measures Mean Usual Mean Intervention Mean t p Value

Pause time, s 2.82 2.24 0.58 4.60 0.002*Speaking time, s 2.03 2.50 0.47 3.68 0.008*Syllables/breath 8.11 10.06 1.95 2.48 0.042*Syllables/inspiration 5.10 5.09 0.01 0.04 0.968Syllables/expiration 3.01 5.00 1.99 2.91 0.023*Articulation rate, syllables/s 4.11 4.08 0.03 0.27 0.797Sound pressure level, decibels 77.05 76.36 0.69 0.73 0.488Peak Pt, cm H2O 18.56 18.48 0.08 0.16 0.881Potential speaking time, s 2.79 3.69 0.90 6.80 0.001*Breathing frequency, breaths/min 13.40 13.09 0.31 0.45 0.668Listener ratings 0 0.33 0.33 1.55 0.166

*

p

0.05.




7/12

during inspiration and expiration, and listener ratingsreflecting improvements in several perceptual di-mensions.

High PEEP vs One-Way Valve

High PEEP (15 cm H2O) was tested in threesubjects (subjects 5, 7, and 13) who used one-wayvalves with positive-pressure ventilation (Table 6).Objective measures obtained with high PEEP werenearly identical to those for the one-way valve in allsubjects. Listeners tended to rate speech with highPEEP as better than with the one-way valve, thoughthis was not statistically significant. The subjects

rated their speech and breathing as the same as usualwith high PEEP. Pt waveforms generated with highPEEP were highly similar to those generated withone-way valves (Fig 2).

Unsuccessful Cases

The interventions were unsuccessful in three sub-jects. We attribute lack of success to significant

laryngeal dysfunction (two subjects) or to a signifi-cant stoma leak (one subject).

Subject 4 had advanced amyotrophic lateral scle-rosis and concomitant laryngeal neuromotor andupper airway control problems. Because she couldnot speak, we tested her while she sustained a vowel.Interventions were attempted with the hope ofincreasing loudness and duration of voicing to im-prove her ability to alert caregivers when she wantedto communicate (she used a foot-operated computerto type messages); however, the interventions werenot effective.

Subject 2, unlike other subjects in this study, spentmost of his waking time with his cuff inflated. Hiscuff was deflated only occasionally to allow him tospeak (with a one-way valve). When we deflated hiscuff and adjusted his ventilator, his larynx remainedopen during inspiration so that only a small portionof the ventilator-delivered volume reached his lungs.His Spo2 dropped below 90%, and we immediatelystopped the test and reinflated his cuff.

Subject 10 had a significant air leak because her

Table 5Group Means of Measures Associated With Usual Ventilator Settings and With the Intervention ofCombining Lengthened TI and PEEP (and, in Two Cases, Reduced VT), Mean Differences Between Usual and

Intervention Measures, and Results of t Tests


Pause time, s 2.53 1.55 0.98 3.94 0.011*Speaking time, s 2.17 3.42 1.25 5.14 0.004*Syllables/breath 7.92 12.72 4.80 4.21 0.008*Syllables/inspiration 5.53 7.98 2.45 4.01 0.010*Syllables/expiration 2.40 4.73 2.33 2.74 0.041*Articulation rate, syllables/s 3.60 3.67 0.07 0.26 0.809Sound pressure level, decibels 74.48 73.77 0.71 1.34 0.237Peak Pt, cm H2O 19.83 17.78 2.05 1.53 0.186Potential speaking time, s 2.85 4.38 1.53 4.75 0.005*Breathing frequency, breaths/min 13.12 12.17 0.95 1.46 0.206Listener ratings 0 0.77 0.77 5.04 0.004*

*p 0.05.

Table 6 Group Means of Measures Associated With Use of a One-Way Valve (Usual) and High PEEP

(Intervention), Mean Differences Between Usual and Intervention Measures, and Results of t Tests


Pause time, s 1.00 1.00 0.00 0.00 1.000Speaking time, s 3.73 3.70 0.03 0.19 0.868Syllables/breath 15.07 15.40 0.33 0.53 0.651Syllables/inspiration 3.90 3.87 0.03 0.07 0.953Syllables/expiration 11.17 11.47 0.30 0.37 0.749Articulation rate, syllables/s 4.03 4.23 0.20 2.00 0.184Sound pressure level, decibels 75.97 75.17 0.80 1.14 0.373Peak Pt, cm H2O 14.83 15.43 0.60 1.09 0.390Potential speaking time, s 4.27 4.23 0.04 0.20 0.860Breathing frequency, breaths/min 12.97 12.93 0.04 0.07 0.951Listener ratings 0 0.37 0.37 2.08 0.173




8/12

tracheostoma was larger than the largest tube hertrachea could accommodate. Thus, expiratory Pt fellto zero and speech improvement was minimal. Theleak was reduced substantially by securing a donut-shaped foam seal around the shaft of the tube. Asa result, her potential speaking time increasedmarkedly (by 3.5 s); however, her actual speakingtime increased only slightly (by 0.5 s). Also,pause time increased (by 0.7 s) because she ceasedtriggering extra breaths, thereby causing cycle peri-ods to increase (by 31%). Thus, her speech becameworse because of her behavioral response to theintervention.

Discussion

These interventions improved speech, especiallywhen used in combination. Also, the application of

high PEEP was found to be as effective as a one-wayvalve.

What Improvements Result From TheseInterventions?

All three interventionslengthened Ti, PEEP,and lengthened Ti plus PEEP (plus reduced Vt)reduced the duration of the obligatory pause andincreased the amount of speech produced perbreath. This was because they maintained Pt abovethe voicing threshold for a greater portion of the

ventilator cycle (not because subjects altered articu-

lation rate or breathing frequency). The combined-adjustment intervention was the most successful ofthe three. It elicited especially high ratings fromlisteners who noted improvements not only in tem-poral features of the speech, but also in loudness,voice quality, and articulation precision. These latter

features reflect changes in laryngeal and upperairway behavior, which appear to have been facili-tated by modifications to the Pt waveform.

Speech improved immediately after the ventilatorwas adjusted. Speech might have improved evenmore with practice and behavioral therapy. Forexample, subjects used only a portion of their poten-tial speaking time (range, 49 to 93%). With practiceand therapy, subjects might have learned to takegreater advantage of the entire time available tospeak.

We observed speech improvement during a read-

ing task. Improvement may have been even moreimpressive if speech had been assessed while sub-jects engaged in conversation. Conversation requirescontinual cognitive-linguistic formulation and carriesmore rigorous timing demands than reading (eg, tofollow turn-taking conventions). Because these inter-ventions allow greater flexibility in speech timing,they are likely to be especially potent in enhancingconversational interchange.

A side benefit of these interventions was thatbreathing felt better. Why this occurred is not clear,but we can speculate. First, tidal stretch of pulmo-

nary afferents is known to relieve air hunger,20

and it

Figure 2. Pt waveforms generated during speech production with a one-way valve and with a PEEPvalve set to 15 cm H2O for subject 13.




9/12

may be that the increase in mean lung volumecaused by PEEP has a similar effect. Second, certainsubjects may have experienced a reduction in per-ceived breathing effort because they did not need totrigger as many breaths as they usually did whenspeaking.

Which Intervention Is Most Effective forImproving Speech?

Lengthened Ti and PEEP each improved speech.When combined, their effects on speaking time wereadditive because they worked on opposite phases ofthe ventilator cycle. Specifically, lengthened Ticaused Pt to stay above the voicing threshold longerduring the inspiratory phase, and PEEP caused Pt tostay above the threshold longer during the expiratoryphase.

The relative benefits of the interventions can be

appreciated by examining the amount of speech (eg,number of syllables) produced over an extendedperiod (eg, a minute). Syllables per minute is a globalmeasure that reflects interactions among severalspeech variables (speaking time, articulation rate,pause time, and breathing frequency) across a seriesof consecutive breaths. Using this measure to com-pare interventions, we determined that subjects pro-duced an average of nearly 20 syllables per minutemore than usual with lengthened Ti, an average of 23syllables per minute more than usual with PEEP,and an average improvement of 50 syllables per

minute (50.3% increase) with combined lengthenedTi and PEEP (with or without reduced Vt) [Fig 3,top]. A similar additive effect was evident in a singlesubject who received all interventions (Fig 3, bottom;the combined intervention appears more than addi-tive in this subject because PEEP was slightly higherfor the combined intervention than for the PEEP-alone intervention).

Although the group results were clear, the magni-tude of speech change was strikingly different forsome subjects. For example, speech improved withPEEP in all subjects (as reflected in changes in one

or more measures), but the magnitude of improve-ment was slight for subject 14 and marked forsubject 1. Despite the fact that these two subjectswere exposed to nearly the same level of PEEP (7cm H2O and 8 cm H2O, respectively), subject 14showed only a 7% increase in syllables per breath(from 10.7 to 11.4 syllables per breath), whereassubject 1 showed a 78% increase in syllables perbreath (from 8.5 to 15.1 syllables per breath). Whythese subjects responded so differently to the samePEEP intervention is not known.

Responses to interventions were more predictable

within a given subject than across different subjects.

For example, data from one subject (subject 6) show

a nearly linear relation of syllables per minute to themagnitude of PEEP (Fig 4).

Another example in the same subject demon-strates the effect of reduced Vt. When Vt wasreduced by 0.25 L (in combination with 12 cm H2OPEEP and lengthened Ti), slightly less speech was

Figure 3. Changes (from usual) in speaking rate (syllables perminute) for each of the three interventions: lengthened Ti,PEEP, and lengthened Ti plus PEEP (plus reduced Vt) for thesubject group (top) and for subject 14, who had all threeinterventions (bottom).

Figure 4. Changes (from usual) in speaking rate with 5, 8, and12 cm H2O PEEP (combined with lengthened Ti from 33 to

50%) for subject 6.




10/12

produced (by approximately five syllables perminute) than with the usual Vt, but listener ratingswere higher (average rating of 1.8 compared to 1.1with the usual Vt). Listeners commented that loud-ness was more even and that voice quality improved.From this, we can infer that the more gradual Pt riseand the lower Pt peak resulted in vocal fold vibration

with less amplitude variation and more normal wave-form characteristics.During this study, we noted that sometimes the

acceptance of an intervention was influenced bythe sequence with which the components wereintroduced. For example, an individual may haverejected lengthened Ti alone. However, if PEEP wasadded first, the same individual may have been morelikely to accept, or even prefer, the lengthened Ti.The same was true for reduced Vt.

Who Are Candidates for These Interventions?

The conclusions drawn from our work apply toindividuals with spinal cord injuries and neuromus-cular diseases (eg, muscular dystrophy, Guillain-Barre syndrome, postpoliomyelitis syndrome, arte-riovenous malformation) who are without significantairway disease and who are receiving long-termventilation. Clinical experience of one of the authors(R. Brown) indicates that the interventions describedhere can also be applied successfully to those whohave received ventilation for only a short time or whomay be in an acute stage of recovery.

Although our findings suggest that most individu-

als with neuromotor disorders can benefit from theseventilator-adjustment interventions, those with sig-nificant laryngeal dysfunction or significant stomaleaks may not. When the problem is maladaptivebehavioral control of the larynx (eg, subject 2),behavioral therapy from a speech-language patholo-gist may be beneficial. When the problem is poorneuromotor control of the larynx (eg, subject 4),there may be little that can be done to improvespeech. When there is a significant stoma leak thatcannot be repaired, adding PEEP may not improvespeech (although improvement may still be possible

with lengthened Ti).

Are These Interventions Safe?

Certain of these interventions can reduce alveolarventilation. Lengthened Ti can reduce ventilationbecause there is more time to speak during inspira-tion. Speech produced during inspiration diverts airaway from the lungs (through the larynx) so that it isnot available for ventilation (whereas air used forspeech during expiration has already ventilated thelungs). Because individuals with respiratory paralysis

can usually perceive air hunger,15,21

this can serve as

a signal to simply close the larynx and stop speaking.This increases ventilation because the full Vt isdelivered to the lungs. By contrast, reduced Vtdecreases alveolar ventilation whether or not thesubject is speaking. Although acute increases inarterial Pco2 can produce discomfort (air hun-ger),15,22 this did not occur in our subjects, probably

because the Pco

2 changes were small (within 2 to 5mm Hg). Even if they had experienced acute airhunger, subjects receiving mechanical ventilation areknown to adapt within 3 days to an increased level ofarterial Pco2.16 In certain cases, a reduction inventilation may actually be beneficial. This is be-cause many individuals who receive long-term posi-tive-pressure ventilation because of neuromotor im-pairments are overventilated (eg, end-tidal Pco2 was 30 mm Hg in nearly all of our subjects). In suchcases, speech adjustments could have the advantageof moving alveolar ventilation toward normal.

Whether a given adjustment actually does alteralveolar ventilation depends on how the individualmanipulates the ventilator-delivered inspiration.

Decreasing Te (by increasing Ti) reduces the timeavailable for the lungs to empty. This did not poseproblems for the present subjects, as they were allable to expire completely even when Te occupied aslittle as 33% of the ventilator cycle. Nevertheless,such a short Te could pose problems for individualswith significant obstructive airway disease.

PEEP has theoretical safety risks that includebarotrauma and reduced cardiac output due to re-

duced venous return. These risks are associated withhigh intrathoracic pressure, but usually are of con-cern only at levels of PEEP higher than those used inthis study. PEEP did not increase peak Pt duringspeech production because diversion of air throughthe larynx allowed Pt to return to zero during speechbreaths. PEEP did not produce adverse effects inthe present study, in our previous work,6 or in arecent study13 of speech with pressure-support ven-tilation and PEEP. Nevertheless, caution shouldalways be used when applying PEEP to individualswith compromised cardiac function (eg, congestive

heart failure) or those with autonomic nervous sys-tem dysfunction (eg, spinal cord injury, Guillain-Barre syndrome, and syringomyelia).

Is High PEEP Better Than a One-Way Valve?

If PEEP is set at or above the peak Pt, it has thesame functional effect as occlusion of the expiratoryline (as achieved with a one-way valve in the com-mon ventilator line or a cork in the ventilatorexpiratory line). We found that speech producedwith 15 cm H2O of PEEP was as good as speech

produced with a one-way valve. The advantage of




11/12

PEEP is that it is safer and less expensive than aone-way valve.23 With a one-way valve, inflation ofthe tracheostomy tube cuff or occlusion of the upperairway could be harmful or fatal. Severe hypoventi-lation will ensue if the pressure-limit safety device ofthe ventilator works properly. Severe barotraumaand reduced cardiac output (and reduced venous

return) will ensue if the pressure-limit safety devicefails to work properly. Substitution of PEEP substan-tially reduces these risks.

Summary and Conclusions

All of our interventions improved speech, but themost successful was the combination of lengthenedTi and PEEP (and, in some cases, reduced Vt).Speech improved immediately and substantially.Further improvement could probably be achievedwith practice and with behavioral therapy providedby a speech-language pathologist.

These ventilator adjustment interventions weresafe and comfortable. Such interventions are simpleand inexpensive (or free) and can be implementedeasily with any clinical ventilator. Speech producedwith high PEEP was as good as speech producedwith a one-way valve. For individuals who use one-way valves, we recommend that PEEP be substi-tuted (using an adjustment internal to the ventilatoror by coupling an external valve to the expiratoryline) because PEEP is safer.

ACKNOWLEDGMENTS: We thank the medical staff at NewEngland Sinai Hospital and Rehabilitation Center, West Roxbury

Veterans Administration Medical Center, Posada del Sol HealthCare Center, and John C. Lincoln Hospital, especially Dr. LawrenceHotes, Dr. Leonard Ditmanson, and James Ruf, RRT. We alsothank Marie Duggan, RRT, for her assistance in data collection;Robert Chase, RRT, for helping us use the stoma seal he developed;and those who helped with data analysis and other aspects of thestudy, most notably Monica Christian, DiAnne Gagliardo, WolfgangGolser, Esther Kim, and Amanda Zimmerman. Special thanks go tothe subjects who participated in this research.

References

1 Bach JR. Amyotrophic lateral sclerosis: communication statusand survival with ventilatory support. Am J Phys Med Rehabil1993; 72:343349

2 Wang T-G, Bach JR, Avilla C, et al. Survival of individualswith spinal muscular atrophy on ventilatory support. Am JPhys Med Rehabil 1994; 73:207211

3 Bach JR, Alba AS. Tracheostomy ventilation: a study ofefficacy with deflated cuffs and cuffless tubes. Chest 1990;97:679 683

4 Tippett DC, Siebens AA. Preserving oral communication inindividuals with tracheostomy and ventilator dependency.Am J Speech Lang Pathol 1995; 4:55 61

5 Hoit JD, Shea SA, Banzett RB. Speech production duringmechanical ventilation in tracheostomized individuals. JSpeech Hear Res 1994; 37:53 63

6 Hoit JD, Banzett RB. Simple adjustments can improveventilator-supported speech. Am J Speech Lang Pathol 1997;6:8796

7 Draper MH, Ladefoged P, Whitteridge D. Expiratory pres-sures and air flow during speech. BMJ 1960; 1:18371843

8 Lieberman P, Knudson R, Mead J. Determination of the rateof change of fundamental frequency with respect to subglottalair pressure during sustained phonation. J Acoust Soc Am1969; 45:15371543

9 Bouhuys A, Proctor DF, Mead J. Kinetic aspects of singing.J Appl Physiol 1966; 21:483 496

10 Hixon TJ, Mead J, Goldman MD. Dynamics of the chest wallduring speech production: function of the thorax, rib cage,diaphragm, and abdomen. J Speech Hear Res 1976; 19:297356

11 Netsell RW. Subglottal and intraoral air pressures during theintervocalic contrast of /t/ and /d/. Phonetica 1969; 20:68 73

12 Murry T, Brown WS. Subglottal air pressure during two typesof vocal activity: vocal fry and modal phonation. Folia Phoni-atr (Basel) 1971; 23:440 449

13 Prigent H, Samuel C, Louis B, et al. Comparative effects oftwo ventilatory modes on speech in tracheostomized neuro-muscular patients. Am J Respir Crit Care Med 2003; 167:114 119

14 Banzett RB, Lansing RW, Brown R. High-level quadriplegicsperceive lung volume change. J Appl Physiol 1987; 62:567573

15 Banzett RB, Lansing RW, Reid MB, et al. Air hunger arisingfrom increased Pco2 in mechanically ventilated quadriple-gics. Respir Physiol 1989; 76:53 68

16 Bloch-Salisbury E, Shea SA, Brown R, et al. Air hunger inducedby acute increase in Pco

2

adapts to chronic elevation of Pco2

inventilated humans. J Appl Physiol 1996; 82:949956

17 Hoit JD, Shea SA. Speech production and speech with aphrenic nerve pacer. Am J Speech Lang Pathol 1996; 5:53 60

18 Hoit JH, Banzett RB, Brown R. Improving ventilator-sup-ported speech [TELEROUNDS #39 live satellite transmis-sion]. National Center for Neurogenic Communication Dis-orders, Tucson, AZ. December 10, 1997

19 Hoit JD, Hixon TJ. Age and speech breathing. J Speech HearRes 1987; 30:351366

20 Manning HL, Shea SA, Schwartzstein RM, et al. Reducedtidal volume increases air hunger at fixed Pco2 in ventilatedquadriplegics. Respir Physiol 1992; 90:19 30

21 Banzett RB, Lansing RW, Brown R, et al. Air hunger fromincreased Pco2 persists after complete neuromuscular block

in humans. Respir Physiol 1990; 81:11822 Banzett RB, Lansing RW, Evans K, et al. Stimulus-response

characteristics of CO2-induced air hunger in normal subjects.Respir Physiol 1996; 103:19 31

23 Hoit JD, Banzett RB. Je peux parler! Am J Respir Crit CareMed 2003; 167:101102




12/12

DOI 10.1378/chest.124.4.15122003;124; 1512-1521Chest

and Robert BrownJeannette D. Hoit, Robert B. Banzett, Heather L. Lohmeier, Thomas J. Hixon

*Clinical Ventilator Adjustments That Improve Speech

December 29, 2010This information is current as of

http://chestjournal.chestpubs.org/content/124/4/1512.full.htmlUpdated Information and services can be found at:

Updated Information & Services

http://chestjournal.chestpubs.org/content/124/4/1512.full.html#ref-list-1This article cites 19 articles, 9 of which can be accessed free at:

References

http://chestjournal.chestpubs.org/content/124/4/1512.full.html#related-urlsThis article has been cited by 4 HighWire-hosted articles:

Cited Bys

http://www.chestpubs.org/site/misc/reprints.xhtmlfound online at:Information about reproducing this article in parts (figures, tables) or in its entirety can bePermissions & Licensing

http://www.chestpubs.org/site/misc/reprints.xhtmlInformation about ordering reprints can be found online:

Reprints

"Services" link to the right of the online article.Receive free e-mail alerts when new articles cite this article. To sign up, select the

Citation Alerts

PowerPoint slide format. See any online figure for directions.

articles can be downloaded for teaching purposes inCHESTFigures that appear inImages in PowerPoint format

b D b 29 2010h j l h bD l d d f
http://chestjournal.chestpubs.org/content/124/4/1512.full.htmlhttp://chestjournal.chestpubs.org/content/124/4/1512.full.htmlhttp://chestjournal.chestpubs.org/content/124/4/1512.full.htmlhttp://chestjournal.chestpubs.org/content/124/4/1512.full.html#ref-list-1http://chestjournal.chestpubs.org/content/124/4/1512.full.html#ref-list-1http://chestjournal.chestpubs.org/content/124/4/1512.full.html#related-urlshttp://chestjournal.chestpubs.org/content/124/4/1512.full.html#related-urlshttp://www.chestpubs.org/site/misc/reprints.xhtmlhttp://www.chestpubs.org/site/misc/reprints.xhtmlhttp://www.chestpubs.org/site/misc/reprints.xhtmlhttp://www.chestpubs.org/site/misc/reprints.xhtmlhttp://chestjournal.chestpubs.org/http://chestjournal.chestpubs.org/http://chestjournal.chestpubs.org/http://chestjournal.chestpubs.org/http://www.chestpubs.org/site/misc/reprints.xhtmlhttp://www.chestpubs.org/site/misc/reprints.xhtmlhttp://chestjournal.chestpubs.org/content/124/4/1512.full.html#related-urlshttp://chestjournal.chestpubs.org/content/124/4/1512.full.html#ref-list-1http://chestjournal.chestpubs.org/content/124/4/1512.full.html

clinical ventilator adjustments that improve speech1

Documents