speech restoration post-pharyngolaryngoesophagectomy using ... · speech restoration...

VeteransAdministration

Journal of Rehabilitation Researchand Development Vol . 25 No. 3Pages 33—40

Speech restoration post-pharyngolaryngoesophagectomy usingtracheogastric fistula

KRZYSZTOF IZDLBSKI, FK, MA, PhD, CCC, JOEL C . ROSS, MD, FACS,DOUGLAS HETZLER, MD, JAMES FONTANESI, MD, and PETER KRUMPE, MDPacific Presbyterian Medical Center, San Francisco, California 94120 andVeterans Administration Medical Center, Martinez, California 94553

Abstract—Restoration of voice and speech in patientswith gastric pull-up presents a formidable challenge, andmany of these patients are left at best with a poorlyfunctional electrolaryngeal speech . To improve this condi-tion, a tracheogastric puncture stented with a biflangedself-retaining Groningen voice button was accomplished,resulting in gastric mucosa vibrations during exhalatoryphase. The biomechanical characteristics of gastric vibra-tions and tracheogastric puncture candidate selectioncriteria are discussed.

Key words : biomechanics, fistula, gastric pull-up, gastricvibrations, tracheogastric voice, Groningen voice button/prosthesis, pharyngolaryngoesophagectomy, speech/voicerestoration, tracheoesophageal and tracheogastric punc-tures.

INTRODUCTION

A successful pharyngolaryngoesophagectomy withpharyngogastric anastomosis—the so-called gastric

Address all correspondence to : Dr . K . Izdebski, Pacific PresbyterianMedical Center, P .O . Box 7999, San Francisco, CA 94120.Dr . Izdebski is Chief of Speech Pathology and Audiology, Director ofVoice Physiology Laboratory, Head and Neck Function Clinic, PacificPresbyterian Medical Center, and Associate Clinical Professor ofOtorhinolaryngology, University of California, Davis.Dr . Ross is Attending Otolaryngologist, VAMC, Assistant ClinicalProfessor, Department of Medicine, University of California, Davis.Dr . Hetzler is Chief, Otorhinolaryngology, VAMC, Assistant ClinicalProfessor, Otorhinolaryngology, University of California, Davis.Dr . Fontanesi is Brachy Therapy Fellow, West Coast Cancer Founda-tion, VAMC.Dr . Krumpe is Chief, Pulmonary Medicine, VAMC, Associate Profes-sor, Department of Medicine, University of California, Davis .

pull-up surgery—involves a variety of postoperativefunctional alterations, including alimentation andspeech . Although the problems involved in solvingthe postoperative difficulties of swallowing anddigestive processes have been discussed extensively,restoration of voice and speech in these patients hasbeen largely unsuccessful (1-4, 7-9) . For example, ofthe 136 cases of gastric pull-up performed at theQueen Mary Hospital in Hong Kong, speech reha-bilitation was highly unsatisfactory, with only 9patients (6 .6 percent) able to produce audible whis-per, and 6 patients (4 .4 percent) able to use anelectrolarynx. This occurred despite the fact that 87percent of patients achieved satisfactory alimentaryfunctions (8).

Reviewing their experience with pharyngoeso-phageal reconstructive techniques, Schechter et al .,report that gastric pull-up resulted in better func-tions of swallowing, weight maintenance, andspeech than deltopectoral flap, pectoralis majorflap, or jejunal graft (7) . Nonetheless, speechrestoration for all types of pharyngoesophagealreconstruction was not considered to be satisfactory,and involved single word utterances or was achievedonly with the aid of an electronic device . Althoughbetter speech quality was obtained for the gastricpull-up group than for the other categories, sincethese patients were able to inject air transorally andgenerate neoesophageal sound, the functional scoresfor speech were still poor.

Of the 101 gastric pull-up surgeries performed inEngland by Harrison and his colleagues, acquisition

33

34

Journal of Rehabilitation Research and Development Vol . 25 No. 3 Summer 1988

of an adequate voice was possible only in a smallnumber of patients, in contrast to the number whoregained trouble-free eating ability (1) . Patients whowere able to produce voice did so by manualcompression of the cervical stomach, or whenexperiencing pharyngogastric fistulas . AlthoughHarrison et at. (2) stated that, in their view,tracheogastric shunt may be of considerable use inrestoring speech in pharyngolaryngoesophagecto-mized patients, they did not elaborate on thisapproach. A tracheojejunal fistula stented with aBlom-Singer voice prosthesis resulted in successfulrestoration of speech in 3 patients reported recentlyby Salamoun et al. (6). Five successful cases oftracheogastric fistula for speech restoration werealso reported by Holden at London's Charing CrossHospital (3).

In our attempt to restore voice function in apatient with gastric pull-up, we drew upon experi-ence involving voice rehabilitation usingtracheoesophageal (TE) puncture procedures (5).The results of this modified TE approach, which wetermed TG (tracheogastric) puncture procedure, aredescribed in this paper.

CASE REPORT

The subject of the study was a 73-year-old malewith a 2-month history of intermittent hoarseness,dysphagia, and odynophagia, who had had a 16-pound weight loss in the 6 months prior to admis-sion . His past history included T, squamous cellcarcinoma of the left vocal fold treated with externalbeam radiation therapy 10 years prior to admission.The patient had continued his smoking habit duringthose 10 years.

Endoscopy showed a large ulcerated lesion of theleft pyriform sinus extending into the cervicalesophagus, which prevented passage of the rigidesophagoscope.

Endolaryngeal structures appeared normal . Bar-ium swallow showed a 3 .5 cm diameter mass of thecervical esophagus . Nutritional support was insti-tuted in the form of nasogastric (NG) feedings . Thepatient experienced airway compromise and requireda tracheostomy . He subsequently underwent a phar-yngolaryngoesophageoctomy in conjunction with aleft radical neck dissection, total thyroidectomy,

gastric pull-up, parathyroid reimplantation in theright forearm, and a feeding jejunostomy.

Postoperatively, the patient developed a left pleu-ral effusion requiring a thoracostomy tube ; hypocal-cemia requiring supplementation of calcium andvitamin D; weakness of the right arm which resolvedspontaneously; and a pharyngocutaneous fistulawhich closed with conservative treatment . Swallow-ing improved following closure of the fistula, andthe patient did well on 6 small tube feedings per day,without symptomatic regurgitation . The patient re-fused postoperative radiation therapy . Throughouthis recovery period, he showed signs of being able todevelop "gastric" sounds consisting of occasionalgrunts, but a greater degree of voice restoration wasdesired . This prompted us to attempt atracheogastric fistula procedure to be stented with avoice prosthesis.

Preoperative testingPrior to attempting tracheogastric fistulatization,

the patient was tested for the ability to producegastric voice. This was accomplished by a 2-stageinsufflation test procedure . During stage one (ex-ternal insufflation) a #16 French red rubber catheterwas passed into the cervical gastric space . Positiveair pressure/air flow was applied via the catheter asthe patient sustained varied oral cavity postures thatsimulated the position for various sounds . Sustainedphonation and speech lasting for the duration ofinsufflation were produced by this method . Toestimate the amount of air pressure/air flow neededto generate voicing, insufflation by mouth was alsoperformed. Air was blown into the catheter (#16French) as the patient performed sustained soundand connected speech gestures, which resulted inacceptable phonation and/or speech . Airpressure/air flow requirements were deemed to benot excessive.

To verify these findings, and to assure that thepatient would be able to produce voicing using hisown respiratory air flow, the stage two insufflationtest was performed . It involved passing a #16 Frenchred rubber catheter in the same fashion as above,but this time the proximal housing of the insuf-flating catheter was attached to the stoma of thepatient . Occluding the housing with the investiga-tor's finger allowed for passage of pulmonary airinto the cervical gastric region that produced sus-

35

IZDEBSKI et al . : Speech Restoration Using Tracheogastric Fistula

Figure 1.A videoradiograph showing positioning of the insufflation catheter (arrow) high in the cervical gastric air space, resulting in afriction-like sound and limited vibration of the gastric mucosa.

tamed sound and connected speech . This procedurewas videotaped, using a flexible fiberopticbronchoscopy unit . Contraction of the gastriccervical stomach was confirmed during productionof sustained sound and connected speech.

Surgical procedureUnder general anesthesia, a secondary TG fistula

was created, using a Groningen tracheal punctureforcep . With the patient in a supine position, a rigidesophagoscope was placed into the cervical stomachvia the oral cavity . The esophagoscope was turned180 degrees to expose the beveled edge toward thepenetration site, which was within the visible lumenof the stoma . Considerable force was required topuncture through the gastric wall . Immediatelyfollowing the puncture, the voice prosthesis wasinserted, using the procedure described by the

Groningen group (5) . The fistula site was placed atthe level of the inferior margin of the stoma.

RESULTS

Immediately following surgery, the patient wasable to produce faint phonation and speech withocclusion of the stoma performed by one of theinvestigators . The patient was reexamined within 48hours of surgical recovery, at which time the stomawas cleaned and the button was ventilated withexternal pressurized air . This was achieved byinserting a #14 French catheter into the lumen of theGroningen voice prosthesis . During this ventilation,strong voicing and speech were produced . But thequality of voice decreased significantly (weakened)when the patient attempted to speak using his ownrespiratory air supply while covering the stoma . This

36


Figure 2.A videoradiograph showing deep positioning of the insufflation catheter (arrow) in the cervical gastric air space, resulting inimproved sound generation and creation of a neoglottis above the tip of the catheter.

occurred despite the absence of air leakage fromaround the occluding finger. The procedure ofexternal ventilation was repeated and good voicewas achieved, though again, it could not be dupli-cated by the patient's own attempts . However, thisdifference could perhaps reflect residual postopera-tive swelling, requiring higher air pressure/air flowsto set the gastric mucosa into vibration.

To verify the preoperative findings, the patientwas reinsufflated at 6 days post-puncture, using thestage two procedure . As previously, this resulted ina production of good voicing, with soft but ade-quate loudness for speech . The insufflation catheter(#16 French) was then marked at 1-inch intervalsand an opaque marker was attached to its distal end.Videoradiography was used to document the level ofthe catheter tip at which insufflation voicing wasoptimal . Raising the catheter from the optimalposition resulted in a weaker voice and decreasedvolume (Figure 1), while pushing the catheter deeper

into the gastric space increased the volume andimproved vibrations (Figure 2). When the positionof the distal tip of the catheter was matched to thein situ depth of the Groningen prosthesis, it becameapparent that the depth at which the prosthesis wasinserted accounted for the poor voice quality, ascompared to the insufflation voice quality . Optimalvoice quality was produced at a point 15 mm belowthe placement of the prosthesis, where the lumen ofgastric space had narrowed sufficiently to providevibratory capability.

Based on this result, a decision was made to createa lower TG fistula . The second fistula was punc-tured 15 mm below the primary site, and a newGroningen voice prosthesis was put into place 25mm below the inferior edge of the stoma (Figure 3).Following this procedure, the primary Groningenvoice prosthesis was removed . Closure of the pri-mary fistula was achieved spontaneously within 24hours . Next morning, the patient was able to

37


Figure 3.A videoradiograph showing optimal positioning of the Groningen voice prosthesis (lower arrow) in the cervical gastric air space and acreation of a clearly visible neoglottis (upper arrow).

swallow without evidence of leakage through theprimary fistula . The function of the new prosthesiswas retested 48 hours postoperatively . Stomal occlu-sion performed manually by one of the investigatorsresulted in adequate voicing and speech (high lungvolume) ; but voicing and speech quality deterioratedsignificantly when occlusion was done by the pa-tient, because of his inability to synchronize respira-tory drive, stoma occlusion, and articulation . Thismay have been the result of his apraxia-like behav-ior, which had developed following the pull-upsurgery.

To verify the gastric and tracheal pressure require-ments needed to generate TG sound, a series ofaerodynamic measures were taken . These consistedof measuring simultaneously : a) gastric air pressure(using an esophageal latex balloon) ; b) trachealpressure (using a custom-made housing with a leadto the differential pressure transducer enclosedwithin this housing) ; and, c) oral air flow via a

pneumotachograph . Air pressures were measureddirectly, using a calibrated differential air-pressuretransducer system (model CD 18 Validine Corpora-tion, Northridge, CA) . The outputs from the trans-ducer were fed into biologic amplifiers (Model VR6, Electronics for Medicine, Pleasantville, NY) andwere written out on a paper strip (running at thespeed of 25 mm/sec for further measurements) . Oralair-flow signal from a pneumotachograph (Fleisch#6) was calibrated from external air-source at roomtemperature and humidity. Phonation loudness wasmeasured at a distance of 1 m in a free field, using asound pressure level meter . Sound pressure measure-ments were taken when the pneumotachographmask was removed from patient's face.

To produce soft phonation (< 60dB) duringinsufflation (external method), an average of 25 cmH2O of gastric pressure was needed . For mid-loudness of about 60dB, 75 cm H 2O gastric pressurewas needed ; for loud phonation (>65dB), between

38


INSUFFLATION

100 —

0

0

75 —

I

I

IS

M

L

PHONATION LEVEL

Figure 4.Plot of gastric air pressure (P G) in cm H 2 Ogenerated for soft (S), mid (M) and loud (L)phonation during external insufflation test.(See text).

80 and 90 cm H 2O gastric pressure was necessary(see Figure 4) . However, when the tracheal andgastric pressures and air flow were measured simul-taneously, as the patient attempted to producephonation, the generated gastric pressure did notexceed 50 cm H2O; tracheal pressure was onlyslightly higher (less than 60 cm H20), and the airflow at the mouth was low (an average of 30 cc/sec).(This is illustrated in Figure 5 .) This aerodynamicprofile (tracheal versus gastric pressure differenceless than 10 cm H 2O ; low air-flow) was inadequatefor production of functional phonation despite theoptimal positioning of the prosthesis under thevibrating segment.

CONCLUSIONS AND DISCUSSION

Failure to restore speech has been associated withpharyngolaryngoesophagectomies . The technique ofTE puncture combined with a voice prosthesis was

modified in this study to restore speech in a patientwith gastric pull-up via a TG puncture procedure.Crucial to the prosthetic restoration of voice wereseveral factors, including: 1) identification of anoptimal location for placement of the fistula andprosthesis ; 2) pressure requirements ; 3) prostheticdesign; and, 4) patient status . A Groningen voicebutton was used because it can be placed within thetrachea without the need for external attachment.Therefore, it could be placed deeper in the tracheathan any other available prosthesis.

For sound to be produced, it was necessary toobtain a critical gastric space to create vibratingneoglottis . Placement of the prosthesis high in thetrachea was unsuccessful due to the wide gastricspace that was unable to contract sufficiently tocreate vibrations but generated only short friction-like segments . This friction was of inadequate powerto produce acceptable speech and resulted in rapidlosses of phonatory air reserves, which necessitatedfrequent inhalations . Lowering the voice prosthesis15 mm satisfied the criteria for the critical

39


AERODYNAMICS OF TGF PHONATION

P (cm H2 0)

'' (ml/sec)

75 —

— 75

50 —

25 —

— 50

— 25

Tp

Gp

AVERAGE MAXIMUM

Figure 5.Histogram of achieved maximum average gastric air pressure(P G) and tracheal air pressure (P T) and air-flow at the mouthduring patient's own attempts to phonate . Pressure measuredin cm H 2 O air flow (V) in ml/sec.

gastric space needed to create vibrations . This isconsistent with recent observations that the develop-ment of neoesophageal speech in gastric pull-uppatients is possible if the cervical portion of thegastric pouch diminishes in size secondary to fibrosis(7) . Similar to prior observations of voicing produc-tion by patients with colon transplant, in whomnarrowing of colon to form pseudoglottis wasobserved (4,9), it is difficult to explain the actioncausing active gastric contraction during speech inthis case. Unfortunately, poor aerodynamics of thispatient imposed difficulty to generate adequatepressure differences to overcome prosthesis resis-tance, and to consistently vibrate the gastricmucosa . Apraxia resulted in problems with coordi-nating respiratory driving force and occlusion of thestoma with his finger . This combination of factorsprevented him from producing voicing adequate forgenerating functional communication . Therefore,the voice prosthesis was removed, and the patientwas provided with an oral electrolarynx. Theelectrolarynx speech was, however, also of poorquality due to coordination problems.

Although from a clinical standpoint this case maybe considered a failure, the experience suggests that

creating gastrotracheal puncture stented with aprosthesis may be encouraging for speech restora-tion if the patient is carefully selected . However,several basic parameters must be achieved . Theseinclude adequate air support, critical gastric space,optimal location of the fistula, proper voice prosthe-sis, adequate tracheal and gastric pressure differ-ence, and the mental status of the patient . If theseparameters exist, it can be assumed that phonationand speech will be achieved using the voice prosthe-sis approach in gastric pull-up patients with minimalpostoperative risks and complications.

ACKNOWLEDGMENT

This manuscript was prepared by Ms . Debra Calogero.

REFERENCES

1 . Harrison DFN : Surgical repair in hypopharyngeal andcervical esophageal cancer . Analysis of 162 patients . AnnOtol 90 :372-375, 1981 .

40


2. Harrison DFN, Thompson AE : Pharyngolaryngoesophag-ectomy with pharyngogastric anastomosis for cancer of

the hypopharynx : Review of 101 operations . Head Neck

7.Surgery 8 :418-428, 1986.

3. Holden H : Personal communication, 1987.4. Lai M, Evison G: Voice production following laryngo-

8.pharyngo-esophagectomy with colon transplant. JLaryngol Otol 80 :1208-1217, 1966.

5. Nijdam HE, Annyas AA, Schutte HK, et al . : A new

prosthesis for voice restoration after laryngectomy . Arch

9.Otorhinolaryngol 237 :27-33, 1982.

6. Salamoun W, Swartz WM, Johnson JT, et al . : Free

jejunal transfer for reconstruction of the laryngopharynx.Otolaryngol Head Neck Surg 96 :149-150, 1987.Schechter GL, Baker JW, Gilbert DA . Functional evalua-tion of pharyngoesophageal reconstructive techniques.Arch Otolaryngol Head Neck Surg 113 :40-44, 1987.Wei WI, Lam KH, Choi S, Wong J : Late problems afterpharyngolaryngoesophagectomy and pharyngogastricanastomosis for cancer of the larynx and hypopharynx.Am J Surg 148 :509-513, 1984.Wertz RT, Keith RL, DeSanto LW : Speech after laryngo-esophagectomy with colon transplant . J Speech Hear Dis38 :495-501, 1973 .

speech restoration post-pharyngolaryngoesophagectomy using ... · speech restoration...

Documents