Clinical Ventilator Adjustments That Improve Speech *

Jeannette D. Hoit, Robert B. Banzett, Heather L. Lohmeier, Thomas J. Hixon and Robert Brown Chest 2003;124;1512-1521 DOI 10.1378/chest.124.4.1512 The online version of this article, along with updated information and services can be found online on the World Wide Web at: http://chestjournal.chestpubs.org/content/124/4/1512.full.html

Chest is the official journal of the American College of Chest Physicians. It has been published monthly since 1935. Copyright2003by the American College of Chest Physicians, 3300 Dundee Road, Northbrook, IL 60062. All rights reserved. No part of this article or PDF may be reproduced or distributed without the prior written permission of the copyright holder. (http://chestjournal.chestpubs.org/site/misc/reprints.xhtml) ISSN:0012-3692

Downloaded from chestjournal.chestpubs.org by guest on December 29, 2010 2003 American College of Chest Physicians

Clinical Ventilator Adjustments That Improve 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 positivepressure ventilation. Such individuals often speak with short phrases, long pauses, and have problems with loudness and voice quality. Subjects: We studied 15 adults with spinal cord injuries or neuromuscular diseases receiving long-term ventilation. Interventions: The ventilator was adjusted using lengthened inspiratory time (TI), positive end-expiratory pressure (PEEP), and combinations thereof. Results: When TI was lengthened (by 8 to 35% of the ventilator cycle), speaking time increased by 19% and pause time decreased by 12%. When PEEP was added (5 to 10 cm H2O), speaking time 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, increasing speaking time by 55% and decreasing pause time by 36%. The combined intervention improved speech timing, loudness, voice quality, and articulation. Individual differences in subject response to the interventions were substantial in some cases. We also tested high PEEP (15 cm H2O) in three subjects and found speech to be essentially identical to that produced with a one-way valve. Conclusions: These simple interventions markedly improve ventilator-supported speech and are safe, at least when used on a short-term basis. High PEEP is a safer alternative than a one-way valve. (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 of arterial oxygen saturation; Te expiratory time; Ti inspiratory time; Vt tidal volume

ability to speak has a profound influence on T he the quality of life for individuals who depend on mechanical ventilation.1,2 Speech makes it possible for such individuals to express needs efficiently, develop and maintain personal relationships, com*From the Department of Speech and Hearing Sciences and National 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 Pulmonary and Critical Care Unit, Massachusetts General Hospital, Boston, MA. Support was provided by National Institute on Deafness and Other Communication Disorders grants DC-03425 and DC01409. Manuscript received September 30, 2002; revision accepted March 12, 2003. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (e-mail: permissions@chestnet.org). Correspondence to: Jeannette D. Hoit, PhD, Department of Speech and Hearing Sciences, PO Box 210071, University of Arizona, Tucson, AZ 85721; e-mail: hoit@email.arizona.edu
1512

municate with caregivers, use the telephone, and operate systems controlled by speech recognition software. Those who have tracheostomies and receive positive-pressure ventilation are usually able to speak if the cuff on the tracheostomy tube is deflated or if a fenestrated or cuffless tracheostomy tube is used.3,4 Nevertheless, speech produced in this way is often far from satisfactory. The settings on positive-pressure ventilators are adjusted to accommodate cardiopulmonary requirements, but they are not optimal for speech. Speech produced with typical ventilator adjustments is often characterized by short phrases, long pauses between phrases, abnormal loudness, and poor voice quality.5,6 These speech problems can be attributed, in large part, to the nature of the tracheal pressure (Pt) waveform. To produce speech, Pt must be at least 2 cm H2O to vibrate the vocal folds,7,8 and it should be relatively constant to maintain steady loudness and normal voice quality.9,10 When speech is produced in able-bodied individuals, Pt is nearly constant at 5 to 10 cm H2O throughout expiration.11,12 In sharp
Clinical Investigations in Critical Care

Downloaded from chestjournal.chestpubs.org by guest on December 29, 2010 2003 American College of Chest Physicians

contrast, the Pt during volume-controlled, positivepressure ventilation rises rapidly to 20 cm H2O during inspiration, then falls rapidly to zero during expiration and remains there until the next inspiration. Thus, Pt is too low to vibrate the vocal folds for much of the ventilator cycle, and when it is high enough, 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 as to optimize Pt for speech production while simultaneously accommodating cardiopulmonary requirements and comfort. This approach involves straightforward adjustments that are within the capabilities of most clinical positive-pressure ventilators. Such adjustments are designed to improve speech by modifying the Pt waveform to: (1) maintain Pt above the minimum required for voicing for a longer period so that more speech can be produced per breath and less time need be spent in mandatory silence, and (2) reduce the rate of change of Pt to allow loudness and voice quality to be more even. We and others have demonstrated the feasibility of this general approach.6,13 The present study extends previous work by examining two single-adjustment interventions and a combined-adjustment intervention, and by including an increased range of ventilator adjustments. Materials and Methods
We studied 15 subjects with spinal cord injuries or neuromuscular diseases (Table 1), who lived in extended-care facilities or at home. Thirteen subjects received ventilation with volumecontrolled positive pressure and produced speech with the tracheostomy 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) routinely used one-way inspiratory valves for speaking (Passy-Muir Tracheostomy Speaking Valve, Passy-Muir, Irvine, CA; and Hudson RC1, Hudson, Temecula, CA). Five subjects (subjects 1, 5, 6, 10, and 15) actively triggered the ventilator to increase breathing frequency when speaking with their usual ventilator settings. The remaining two subjects (subjects 8 and 9) had used volumecontrolled, positive-pressure ventilators in past years, but at the time of the study were routinely using phrenic nerve pacers (combined with one-way valves) for ventilation and speech production. The study protocol was approved by all appropriate human subjects committees, and informed consent was obtained from all subjects. Subjects were studied using a standard ventilator (Servo Ventilator 900C; Siemens-Elema; Solna, Sweden) with settings matched 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 removed), 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 the inspiratory phase of the ventilator cycle (Fig 1, top left, a). With lengthened Ti, air flows through the larynx longer so that Pt remains 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 the lungs flows toward the ventilator because the ventilator pathway offers much lower impedance (ie, primarily resistance) to flow than does the laryngeal pathway (Fig 1, top right, b). PEEP impedes expiratory flow and adds a threshold occlusion pressure to the ventilator expiratory line so that more air flows through the larynx than toward the ventilator (Fig 1, bottom left, c). Thus, Pt stays 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). A one-way valve shunts all expired air through the larynx by occluding the ventilator line (Fig 1, bottom right, d). The specific levels of Ti and PEEP were determined individ-

Table 1Subject Information

Subject No.* 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Sex Male Male Male Female Male Male Male Male Male Female Female Male Male Male Male Age, yr 34 29 69 45 20 35 52 35 51 51 41 33 41 72 40 Primary Etiology Duchenne muscular dystrophy Spinal cord injury (C1) Guillain-Barre syndrome Amytrophic lateral sclerosis Dejerine-Sottas disease Spinal cord injury (C1) Spinal cord injury (C2) Spinal cord injury (C3/4) Spinal cord injury (C2) Limb-girdle muscular dystrophy Arteriovenous malformation Duchenne muscular dystrophy Spinal cord injury (C2) Postpoliomyelitis syndrome Duchenne muscular dystrophy Duration of Ventilation, yr 11 3 4 3 8 9 5 13 13 13 18 4 5 1 8 Articulatory Impairment Mild None Mild Profound None None None None None None Mild Moderate None None Mild Hearing Impairment None None Mild None Moderate to severe None Mild None None None None None None Moderate 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).5,6,14 17
www.chestjournal.org CHEST / 124 / 4 / OCTOBER, 2003

1513

Downloaded from chestjournal.chestpubs.org by guest on December 29, 2010 2003 American College of Chest Physicians

Figure 1. Air flow during ventilator-supported speech production. The black circles represent occlusions, 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), almost all air flows toward the ventilator. This is because the impedance of the ventilator pathway is much lower 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 flows through the larynx. During expiration with a one-way valve (bottom right, d), all air flows through the larynx. Adapted from Hoit et al.18

ually for each subject by briefly testing a range of levels. The level that was most comfortable for the subject was used for the study. It was not possible to test all subjects on all interventions due to time limitations imposed by nursing staff schedules, subject fatigue, subject preferences (eg, rejection of lengthened Ti), and the fact that several subjects also participated in additional interventions not reported here. With each condition, we recorded several breaths with nose and mouth closed (except in the four subjects whose usual condition included use of a one-way valve). The subject read a paragraph19 aloud, and then was asked How does your speech sound? and How does your breathing feel? The subject responded by using a rating scale ( 2 much worse than usual; 1 slightly worse than usual; 0 usual; 1 slightly better than usual; 2 much better than usual). The speech signal was sensed by a head-mounted microphone (C451EB; AKG Acoustics; Wein, Austria). Pt was sampled using a polyethylene catheter inserted through the sealed port of a swivel adapter (Portex 0803; Concord/Portex; Keene, NH) and advanced through the tracheostomy tube to a point just within its proximal end. The catheter was connected to a transducer (Validyne MP45 with 50 cm H2O diaphragm; Validyne Engineering; Northridge, CA) and amplifier (Validyne MC 13). Tidal Pco2 was sampled from the common ventilator line and measured with a clinical monitor (Cardiocap II; Datex Engstrom;
1514

Helsinki, Finland). Noninvasive measure of arterial oxygen saturation (Spo2), heart rate, and noninvasive BP were monitored continually. 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 Quadra 950, LabVIEW Software; National Instruments; Austin, TX) [Table 2]. Auditory perceptual analysis was conducted by five listeners, all of whom were certified speech-language pathologists. Each listener was presented pairs of speech samples, one with the subjects usual ventilator settings and one with an intervention. Listeners were asked to mark on a visual analog scale whether the second sample was better, worse, or the same as the first (usual) sample, and provide descriptors. The scale was centered 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 objective measures 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 breathing was comfortable. In the case of subjects who used one-way valves, an intervention was deemed successful if the speech was as good as with the one-way valve. Statistical inferences were made with paired t tests (objective measures) or one-sample t tests (listener ratings). An level of 0.05 was used.
Clinical Investigations in Critical Care

Downloaded from chestjournal.chestpubs.org by guest on December 29, 2010 2003 American College of Chest Physicians

Table 2Objective Measures

Measures Pause time Speaking time Definition Time (in s) per cycle during which speech is not produced (ie, silent time) Time (in s) per cycle during which speech is produced No. of syllables produced per breathing cycle Significance Silent pauses between speech phrases interrupt the flow of speech. Short speech phrases make it necessary to pause at nonlinguistic junctures. Syllables are units of speech. Syllables/breath may or may not covary with speaking time. Syllables/inspiration are expected to increase with lengthened Ti. Syllables/expiration are expected to increase with PEEP. Changes in articulation rate can change syllables/breath independent of changes in speaking time. Sound pressure level is the acoustic correlate of loudness (an auditory percept). Peak Pt is usually high and should not be made higher with interventions. Pt of at least 2 cm H2O is required to maintain vocal fold vibration. Breathing frequency can change if the active triggering frequency changes. Goal Decrease Increase

Syllables/breath

Increase

Syllables/inspiration Syllables/expiration Articulation rate

No. of syllables produced per inspiratory phase of the cycle No. of syllables produced per expiratory phase of the cycle Speed at which speech is produced (in syllables/s) calculated for speaking time only (ie, pause time deleted) Average amplitude (in decibels) of the (acoustic) speech signal Highest Pt (in cm H2O) within a cycle (occurs at end-inspiration) Time (in s) per cycle during which Pt is at least 2 cm H2O Breathing frequency (in breaths/min)

Increase Increase No change

Sound pressure level

No change

Peak Pt

No change or decrease Increase No change

Potential speaking time Breathing frequency

Results Speech improved in 12 of our 15 subjects with one or more of the interventions. Results of successful interventions are presented below according to type. Unsuccessful cases are considered separately thereafter. Lengthened TI Six subjects (subjects 1, 3, 5, 7, 14, and 15) were tested 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 amount in 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 triggered less often than usual. Lengthened Ti increased actual and potential speaking time (mean, 19% for both) and number of syllables produced per breath (mean, 21%). This increase was entirely attributable to an increase in syllables produced during the inspiratory phase of the cycle (mean, 58%). No significant changes

Table 3Group Means of Measures Associated With Usual Ventilator Settings and With the Intervention of Lengthened TI, Mean Differences Between Usual and Intervention Measures, and Results of t Tests
Measure Pause time, s Speaking time, s Syllables/breath Syllables/inspiration Syllables/expiration Articulation rate, syllables/s Sound pressure level, decibels Peak Pt, cm H2O Potential speaking time, s Breathing frequency, breaths/min Listener ratings *p 0.05.
www.chestjournal.org CHEST / 124 / 4 / OCTOBER, 2003

Mean Usual 1.63 2.62 10.23 4.48 5.73 3.88 71.18 20.87 3.23 14.75 0

Mean Intervention 1.43 3.10 12.33 7.07 5.28 4.00 70.13 19.42 3.80 13.57 0.15

Mean 0.20 0.48 2.10 2.59 0.45 0.12 1.05 1.45 0.57 1.18 0.15

t 1.39 10.13 5.64 5.81 1.19 1.23 1.09 2.03 5.94 1.79 0.70

p Value 0.223 0.001* 0.002* 0.002* 0.287 0.272 0.324 0.098 0.002* 0.133 0.515

1515

Downloaded from chestjournal.chestpubs.org by guest on December 29, 2010 2003 American College of Chest Physicians

were found in pause time, syllables/expiration, articulation rate, sound pressure level, breathing frequency, or peak Pt. Listeners rated subjects speech with lengthened Ti as generally better than usual, but this was not statistically significant. Nearly all subjects rated their own speech as sounding better (four subjects) or the same (one subject) as usual, and breathing as feeling better (three subjects) or the same (two subjects) as usual. Thus, objective measures of speech improved with lengthened Ti. The primary improvement was an increase in speech produced during inspiration. PEEP Eight subjects (subjects 1, 3, 6, 8, 9, 11, 12, and 14) who did not use one-way valves were tested with low levels of PEEP (5 to 10 cm H2O) compared to no PEEP (Table 4). Minimum Pt equaled the PEEP setting when the nose and mouth were occluded, and it usually fell near (or to) zero during speaking trials. With PEEP, pause time decreased in all subjects (mean, 21%), and speaking time increased in most subjects (mean, 25%). All subjects produced more syllables per breath (mean, 25%). This was due entirely to more syllables being produced during the expiratory phase of the ventilator cycle (mean, 67%). Pt remained above the voicing threshold longer (ie, potential speaking time increased; mean, 32%), which explains why subjects were able to produce more speech. PEEP did not significantly alter syllables/inspiration, articulation rate, sound pressure level, peak Pt, or breathing frequency. Listeners rated subjects speech with PEEP as generally better than usual, but this was not statistically significant. Nearly all subjects rated their own speech as sounding better (five subjects) or the same (two subjects) as usual, and their breathing as feeling better (five subjects) or the same (two subjects) as usual.

Thus, objective measures revealed that speech improved with PEEP. Improvement was primarily in the form of shorter pauses and more speech produced during expiration. Lengthened TI Plus PEEP (Plus Reduced Tidal Volume) Six subjects (subjects 3, 6, 11, 12, 14, and 15) were tested with the combination of lengthened Ti (lengthened by 17 to 25% of the cycle) and low PEEP (5 to 10 cm H2O) [Table 5]. For subjects 11 and 15, tidal volume (Vt) was also reduced slightly (by 0.1 to 0.2 L) because they commented that the pressure felt too high. These combined adjustments resulted in shorter pause time (mean, 36%), longer speaking time (mean, 55%), and more syllables per breath (mean, 61%). The increase in syllable production occurred during both the inspiratory phase (mean, 45%) and expiratory phase (mean, 96%) of the ventilator cycle in all subjects. Potential speaking time increased by about the same amount as actual speaking 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 the ventilator (subjects 6 and 15) reduced their breathing frequency with the combined adjustments. Listeners rated speech as better than usual for all subjects (this was statistically significant). Their descriptors indicated improvements in timing (pause time, speaking time), loudness (overall loudness, loudness variation), voice quality (overall voice quality, voice quality variation), and articulation (articulation precision). Subjects rated their speech and breathing as better (four subjects) or the same (two subjects) as usual. Speech improvements were in the form of both objective and subjective measures. Specific improvements were shorter pauses, more speech produced

Table 4 Group Means of Measures Associated With Usual Ventilator Settings and With the Intervention of Adding PEEP, Mean Differences Between Usual and Intervention Measures, and Results of t Tests
Measures Pause time, s Speaking time, s Syllables/breath Syllables/inspiration Syllables/expiration Articulation rate, syllables/s Sound pressure level, decibels Peak Pt, cm H2O Potential speaking time, s Breathing frequency, breaths/min Listener ratings *p 0.05. 1516
Clinical Investigations in Critical Care

Mean Usual 2.82 2.03 8.11 5.10 3.01 4.11 77.05 18.56 2.79 13.40 0

Mean Intervention 2.24 2.50 10.06 5.09 5.00 4.08 76.36 18.48 3.69 13.09 0.33

Mean 0.58 0.47 1.95 0.01 1.99 0.03 0.69 0.08 0.90 0.31 0.33

t 4.60 3.68 2.48 0.04 2.91 0.27 0.73 0.16 6.80 0.45 1.55

p Value 0.002* 0.008* 0.042* 0.968 0.023* 0.797 0.488 0.881 0.001* 0.668 0.166

Downloaded from chestjournal.chestpubs.org by guest on December 29, 2010 2003 American College of Chest Physicians

Table 5Group Means of Measures Associated With Usual Ventilator Settings and With the Intervention of Combining Lengthened TI and PEEP (and, in Two Cases, Reduced VT), Mean Differences Between Usual and Intervention Measures, and Results of t Tests
Measures Pause time, s Speaking time, s Syllables/breath Syllables/inspiration Syllables/expiration Articulation rate, syllables/s Sound pressure level, decibels Peak Pt, cm H2O Potential speaking time, s Breathing frequency, breaths/min Listener ratings *p 0.05. Mean Usual 2.53 2.17 7.92 5.53 2.40 3.60 74.48 19.83 2.85 13.12 0 Mean Intervention 1.55 3.42 12.72 7.98 4.73 3.67 73.77 17.78 4.38 12.17 0.77 Mean 0.98 1.25 4.80 2.45 2.33 0.07 0.71 2.05 1.53 0.95 0.77 t 3.94 5.14 4.21 4.01 2.74 0.26 1.34 1.53 4.75 1.46 5.04 p Value 0.011* 0.004* 0.008* 0.010* 0.041* 0.809 0.237 0.186 0.005* 0.206 0.004*

during inspiration and expiration, and listener ratings reflecting improvements in several perceptual dimensions. High PEEP vs One-Way Valve High PEEP (15 cm H2O) was tested in three subjects (subjects 5, 7, and 13) who used one-way valves with positive-pressure ventilation (Table 6). Objective measures obtained with high PEEP were nearly identical to those for the one-way valve in all subjects. Listeners tended to rate speech with high PEEP as better than with the one-way valve, though this was not statistically significant. The subjects rated their speech and breathing as the same as usual with high PEEP. Pt waveforms generated with high PEEP were highly similar to those generated with one-way valves (Fig 2). Unsuccessful Cases The interventions were unsuccessful in three subjects. We attribute lack of success to significant

laryngeal dysfunction (two subjects) or to a significant stoma leak (one subject). Subject 4 had advanced amyotrophic lateral sclerosis and concomitant laryngeal neuromotor and upper airway control problems. Because she could not speak, we tested her while she sustained a vowel. Interventions were attempted with the hope of increasing loudness and duration of voicing to improve her ability to alert caregivers when she wanted to communicate (she used a foot-operated computer to type messages); however, the interventions were not effective. Subject 2, unlike other subjects in this study, spent most of his waking time with his cuff inflated. His cuff was deflated only occasionally to allow him to speak (with a one-way valve). When we deflated his cuff and adjusted his ventilator, his larynx remained open during inspiration so that only a small portion of the ventilator-delivered volume reached his lungs. His Spo2 dropped below 90%, and we immediately stopped the test and reinflated his cuff. Subject 10 had a significant air leak because her

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

www.chestjournal.org

CHEST / 124 / 4 / OCTOBER, 2003

1517

Downloaded from chestjournal.chestpubs.org by guest on December 29, 2010 2003 American College of Chest Physicians

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

tracheostoma was larger than the largest tube her trachea could accommodate. Thus, expiratory Pt fell to zero and speech improvement was minimal. The leak was reduced substantially by securing a donutshaped foam seal around the shaft of the tube. As a result, her potential speaking time increased markedly (by 3.5 s); however, her actual speaking time increased only slightly (by 0.5 s). Also, pause time increased (by 0.7 s) because she ceased triggering extra breaths, thereby causing cycle periods to increase (by 31%). Thus, her speech became worse because of her behavioral response to the intervention. Discussion These interventions improved speech, especially when used in combination. Also, the application of high PEEP was found to be as effective as a one-way valve. What Improvements Result From These Interventions? All three interventionslengthened Ti, PEEP, and lengthened Ti plus PEEP (plus reduced Vt) reduced the duration of the obligatory pause and increased the amount of speech produced per breath. This was because they maintained Pt above the voicing threshold for a greater portion of the ventilator cycle (not because subjects altered articu1518

lation rate or breathing frequency). The combinedadjustment intervention was the most successful of the three. It elicited especially high ratings from listeners who noted improvements not only in temporal features of the speech, but also in loudness, voice quality, and articulation precision. These latter features reflect changes in laryngeal and upper airway behavior, which appear to have been facilitated by modifications to the Pt waveform. Speech improved immediately after the ventilator was adjusted. Speech might have improved even more with practice and behavioral therapy. For example, subjects used only a portion of their potential speaking time (range, 49 to 93%). With practice and therapy, subjects might have learned to take greater advantage of the entire time available to speak. We observed speech improvement during a reading task. Improvement may have been even more impressive if speech had been assessed while subjects engaged in conversation. Conversation requires continual cognitive-linguistic formulation and carries more rigorous timing demands than reading (eg, to follow turn-taking conventions). Because these interventions allow greater flexibility in speech timing, they are likely to be especially potent in enhancing conversational interchange. A side benefit of these interventions was that breathing felt better. Why this occurred is not clear, but we can speculate. First, tidal stretch of pulmonary afferents is known to relieve air hunger,20 and it
Clinical Investigations in Critical Care

Downloaded from chestjournal.chestpubs.org by guest on December 29, 2010 2003 American College of Chest Physicians

may be that the increase in mean lung volume caused by PEEP has a similar effect. Second, certain subjects may have experienced a reduction in perceived breathing effort because they did not need to trigger as many breaths as they usually did when speaking. Which Intervention Is Most Effective for Improving Speech? Lengthened Ti and PEEP each improved speech. When combined, their effects on speaking time were additive because they worked on opposite phases of the ventilator cycle. Specifically, lengthened Ti caused Pt to stay above the voicing threshold longer during the inspiratory phase, and PEEP caused Pt to stay above the threshold longer during the expiratory phase. The relative benefits of the interventions can be appreciated by examining the amount of speech (eg, number of syllables) produced over an extended period (eg, a minute). Syllables per minute is a global measure that reflects interactions among several speech variables (speaking time, articulation rate, pause time, and breathing frequency) across a series of consecutive breaths. Using this measure to compare interventions, we determined that subjects produced an average of nearly 20 syllables per minute more than usual with lengthened Ti, an average of 23 syllables per minute more than usual with PEEP, and an average improvement of 50 syllables per minute (50.3% increase) with combined lengthened Ti and PEEP (with or without reduced Vt) [Fig 3, top]. A similar additive effect was evident in a single subject who received all interventions (Fig 3, bottom; the combined intervention appears more than additive in this subject because PEEP was slightly higher for the combined intervention than for the PEEPalone intervention). Although the group results were clear, the magnitude of speech change was strikingly different for some subjects. For example, speech improved with PEEP in all subjects (as reflected in changes in one or more measures), but the magnitude of improvement was slight for subject 14 and marked for subject 1. Despite the fact that these two subjects were exposed to nearly the same level of PEEP (7 cm H2O and 8 cm H2O, respectively), subject 14 showed only a 7% increase in syllables per breath (from 10.7 to 11.4 syllables per breath), whereas subject 1 showed a 78% increase in syllables per breath (from 8.5 to 15.1 syllables per breath). Why these subjects responded so differently to the same PEEP intervention is not known. Responses to interventions were more predictable within a given subject than across different subjects.
www.chestjournal.org

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

For example, data from one subject (subject 6) show a nearly linear relation of syllables per minute to the magnitude of PEEP (Fig 4). Another example in the same subject demonstrates the effect of reduced Vt. When Vt was reduced by 0.25 L (in combination with 12 cm H2O PEEP and lengthened Ti), slightly less speech was

Figure 4. Changes (from usual) in speaking rate with 5, 8, and 12 cm H2O PEEP (combined with lengthened Ti from 33 to 50%) for subject 6.
CHEST / 124 / 4 / OCTOBER, 2003

1519

Downloaded from chestjournal.chestpubs.org by guest on December 29, 2010 2003 American College of Chest Physicians

produced (by approximately five syllables per minute) than with the usual Vt, but listener ratings were higher (average rating of 1.8 compared to 1.1 with the usual Vt). Listeners commented that loudness was more even and that voice quality improved. From this, we can infer that the more gradual Pt rise and the lower Pt peak resulted in vocal fold vibration with less amplitude variation and more normal waveform characteristics. During this study, we noted that sometimes the acceptance of an intervention was influenced by the sequence with which the components were introduced. For example, an individual may have rejected lengthened Ti alone. However, if PEEP was added first, the same individual may have been more likely 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 to individuals with spinal cord injuries and neuromuscular diseases (eg, muscular dystrophy, GuillainBarre syndrome, postpoliomyelitis syndrome, arteriovenous malformation) who are without significant airway disease and who are receiving long-term ventilation. Clinical experience of one of the authors (R. Brown) indicates that the interventions described here can also be applied successfully to those who have received ventilation for only a short time or who may be in an acute stage of recovery. Although our findings suggest that most individuals with neuromotor disorders can benefit from these ventilator-adjustment interventions, those with significant laryngeal dysfunction or significant stoma leaks may not. When the problem is maladaptive behavioral control of the larynx (eg, subject 2), behavioral therapy from a speech-language pathologist may be beneficial. When the problem is poor neuromotor control of the larynx (eg, subject 4), there may be little that can be done to improve speech. When there is a significant stoma leak that cannot be repaired, adding PEEP may not improve speech (although improvement may still be possible with lengthened Ti). Are These Interventions Safe? Certain of these interventions can reduce alveolar ventilation. Lengthened Ti can reduce ventilation because there is more time to speak during inspiration. Speech produced during inspiration diverts air away from the lungs (through the larynx) so that it is not available for ventilation (whereas air used for speech during expiration has already ventilated the lungs). Because individuals with respiratory paralysis can usually perceive air hunger,15,21 this can serve as
1520

a signal to simply close the larynx and stop speaking. This increases ventilation because the full Vt is delivered to the lungs. By contrast, reduced Vt decreases alveolar ventilation whether or not the subject is speaking. Although acute increases in arterial Pco2 can produce discomfort (air hunger),15,22 this did not occur in our subjects, probably because the Pco2 changes were small (within 2 to 5 mm Hg). Even if they had experienced acute air hunger, subjects receiving mechanical ventilation are known to adapt within 3 days to an increased level of arterial Pco2.16 In certain cases, a reduction in ventilation may actually be beneficial. This is because many individuals who receive long-term positive-pressure ventilation because of neuromotor impairments are overventilated (eg, end-tidal Pco2 was 30 mm Hg in nearly all of our subjects). In such cases, speech adjustments could have the advantage of moving alveolar ventilation toward normal. Whether a given adjustment actually does alter alveolar ventilation depends on how the individual manipulates the ventilator-delivered inspiration. Decreasing Te (by increasing Ti) reduces the time available for the lungs to empty. This did not pose problems for the present subjects, as they were all able to expire completely even when Te occupied as little as 33% of the ventilator cycle. Nevertheless, such a short Te could pose problems for individuals with significant obstructive airway disease. PEEP has theoretical safety risks that include barotrauma and reduced cardiac output due to reduced venous return. These risks are associated with high intrathoracic pressure, but usually are of concern only at levels of PEEP higher than those used in this study. PEEP did not increase peak Pt during speech production because diversion of air through the larynx allowed Pt to return to zero during speech breaths. PEEP did not produce adverse effects in the present study, in our previous work,6 or in a recent study13 of speech with pressure-support ventilation and PEEP. Nevertheless, caution should always be used when applying PEEP to individuals with compromised cardiac function (eg, congestive heart failure) or those with autonomic nervous system dysfunction (eg, spinal cord injury, GuillainBarre syndrome, and syringomyelia). Is High PEEP Better Than a One-Way Valve? If PEEP is set at or above the peak Pt, it has the same functional effect as occlusion of the expiratory line (as achieved with a one-way valve in the common ventilator line or a cork in the ventilator expiratory line). We found that speech produced with 15 cm H2O of PEEP was as good as speech produced with a one-way valve. The advantage of
Clinical Investigations in Critical Care

Downloaded from chestjournal.chestpubs.org by guest on December 29, 2010 2003 American College of Chest Physicians

PEEP is that it is safer and less expensive than a one-way valve.23 With a one-way valve, inflation of the tracheostomy tube cuff or occlusion of the upper airway could be harmful or fatal. Severe hypoventilation will ensue if the pressure-limit safety device of the ventilator works properly. Severe barotrauma and reduced cardiac output (and reduced venous return) will ensue if the pressure-limit safety device fails to work properly. Substitution of PEEP substantially reduces these risks. Summary and Conclusions All of our interventions improved speech, but the most successful was the combination of lengthened Ti and PEEP (and, in some cases, reduced Vt). Speech improved immediately and substantially. Further improvement could probably be achieved with practice and with behavioral therapy provided by a speech-language pathologist. These ventilator adjustment interventions were safe and comfortable. Such interventions are simple and inexpensive (or free) and can be implemented easily with any clinical ventilator. Speech produced with high PEEP was as good as speech produced with a one-way valve. For individuals who use oneway valves, we recommend that PEEP be substituted (using an adjustment internal to the ventilator or by coupling an external valve to the expiratory line) because PEEP is safer.
ACKNOWLEDGMENTS: We thank the medical staff at New England Sinai Hospital and Rehabilitation Center, West Roxbury Veterans Administration Medical Center, Posada del Sol Health Care Center, and John C. Lincoln Hospital, especially Dr. Lawrence Hotes, Dr. Leonard Ditmanson, and James Ruf, RRT. We also thank 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 the study, most notably Monica Christian, DiAnne Gagliardo, Wolfgang Golser, Esther Kim, and Amanda Zimmerman. Special thanks go to the subjects who participated in this research.

References
1 Bach JR. Amyotrophic lateral sclerosis: communication status and survival with ventilatory support. Am J Phys Med Rehabil 1993; 72:343349 2 Wang T-G, Bach JR, Avilla C, et al. Survival of individuals with spinal muscular atrophy on ventilatory support. Am J Phys Med Rehabil 1994; 73:207211 3 Bach JR, Alba AS. Tracheostomy ventilation: a study of efficacy with deflated cuffs and cuffless tubes. Chest 1990; 97:679 683

4 Tippett DC, Siebens AA. Preserving oral communication in individuals with tracheostomy and ventilator dependency. Am J Speech Lang Pathol 1995; 4:55 61 5 Hoit JD, Shea SA, Banzett RB. Speech production during mechanical ventilation in tracheostomized individuals. J Speech Hear Res 1994; 37:53 63 6 Hoit JD, Banzett RB. Simple adjustments can improve ventilator-supported speech. Am J Speech Lang Pathol 1997; 6:8796 7 Draper MH, Ladefoged P, Whitteridge D. Expiratory pressures and air flow during speech. BMJ 1960; 1:18371843 8 Lieberman P, Knudson R, Mead J. Determination of the rate of change of fundamental frequency with respect to subglottal air pressure during sustained phonation. J Acoust Soc Am 1969; 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 wall during 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 the intervocalic contrast of /t/ and /d/. Phonetica 1969; 20:68 73 12 Murry T, Brown WS. Subglottal air pressure during two types of vocal activity: vocal fry and modal phonation. Folia Phoniatr (Basel) 1971; 23:440 449 13 Prigent H, Samuel C, Louis B, et al. Comparative effects of two ventilatory modes on speech in tracheostomized neuromuscular patients. Am J Respir Crit Care Med 2003; 167: 114 119 14 Banzett RB, Lansing RW, Brown R. High-level quadriplegics perceive lung volume change. J Appl Physiol 1987; 62:567 573 15 Banzett RB, Lansing RW, Reid MB, et al. Air hunger arising from increased Pco2 in mechanically ventilated quadriplegics. Respir Physiol 1989; 76:53 68 16 Bloch-Salisbury E, Shea SA, Brown R, et al. Air hunger induced by acute increase in Pco2 adapts to chronic elevation of Pco2 in ventilated humans. J Appl Physiol 1996; 82:949 956 17 Hoit JD, Shea SA. Speech production and speech with a phrenic nerve pacer. Am J Speech Lang Pathol 1996; 5:53 60 18 Hoit JH, Banzett RB, Brown R. Improving ventilator-supported speech [TELEROUNDS #39 live satellite transmission]. National Center for Neurogenic Communication Disorders, Tucson, AZ. December 10, 1997 19 Hoit JD, Hixon TJ. Age and speech breathing. J Speech Hear Res 1987; 30:351366 20 Manning HL, Shea SA, Schwartzstein RM, et al. Reduced tidal volume increases air hunger at fixed Pco2 in ventilated quadriplegics. Respir Physiol 1992; 90:19 30 21 Banzett RB, Lansing RW, Brown R, et al. Air hunger from increased Pco2 persists after complete neuromuscular block in humans. Respir Physiol 1990; 81:118 22 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 Care Med 2003; 167:101102

www.chestjournal.org

CHEST / 124 / 4 / OCTOBER, 2003

1521

Downloaded from chestjournal.chestpubs.org by guest on December 29, 2010 2003 American College of Chest Physicians

Clinical Ventilator Adjustments That Improve Speech* Jeannette D. Hoit, Robert B. Banzett, Heather L. Lohmeier, Thomas J. Hixon and Robert Brown Chest 2003;124; 1512-1521 DOI 10.1378/chest.124.4.1512 This information is current as of December 29, 2010
Updated Information & Services Updated Information and services can be found at: http://chestjournal.chestpubs.org/content/124/4/1512.full.html References This article cites 19 articles, 9 of which can be accessed free at: http://chestjournal.chestpubs.org/content/124/4/1512.full.html#ref-list-1 Cited Bys This article has been cited by 4 HighWire-hosted articles: http://chestjournal.chestpubs.org/content/124/4/1512.full.html#related-urls Permissions & Licensing Information about reproducing this article in parts (figures, tables) or in its entirety can be found online at: http://www.chestpubs.org/site/misc/reprints.xhtml Reprints Information about ordering reprints can be found online: http://www.chestpubs.org/site/misc/reprints.xhtml Citation Alerts Receive free e-mail alerts when new articles cite this article. To sign up, select the "Services" link to the right of the online article. Images in PowerPoint format Figures that appear in CHEST articles can be downloaded for teaching purposes in PowerPoint slide format. See any online figure for directions.

Downloaded from chestjournal.chestpubs.org by guest on December 29, 2010 2003 American College of Chest Physicians