- Email: [email protected]
Double-Blind Study on the Effects of Topical Anesthesia on Laryngeal Secretions
Double-Blind Study on the Effects of Topical Anesthesia on Laryngeal Secretions *Joy Walsh, †‡Ryan C. Branski, and †‡Katherine Verdolini *Boston, Massachusetts, and †‡Pittsburgh, Pennsylvania
Summary: The application of topical anesthesia to the oropharynx is a common clinical practice during oral and nasal laryngoscopy. Clinically, questions have been raised about whether topical anesthesia alters laryngeal secretions, which distorts clinical impressions. A double-blind, placebo controlled design was employed to address this issue. Ten premenopausal women with healthy vocal folds and 10 premenopausal women with phonotraumatic lesions underwent oral videolaryngoscopic examinations on subsequent days under both anesthesia and placebo conditions, in counterbalanced order. Video segments were rated by three judges. Dependent variables were balling and pooling of secretions, as previously described in the literature. Statistical analyses failed to reveal any clear effect of topical anesthesia on either secretion balling or pooling for the collapsed data set, but one cannot exclude changes in individual cases. Moreover, there was no evidence that secretions were differentially affected by anesthesia across subject groups. Null results in this data set replicate and extend previously reported findings by other authors.1 An incidental but potentially interesting finding was that the order of treatment condition (anesthesia versus placebo first) seemed relevant for secretions: Subjects who received the anesthesia condition first tended to show more secretion balling in general, as compared with subjects who received the placebo condition first. Speculation is entertained regarding possible physiological pathways for these incidental findings, which could be relevant for some clinical practice. Key Words: Anesthesia—Vocal folds—Secretions.
Supported in part by the National Institute on Deafness and Other Communication Disorders (RO1 DC005643). Address correspondence and reprint requests to Katherine Verdolini, Communication Science and Disorders, School of Health and Rehabilitation Sciences, 4033 Forbes Tower, Pittsburgh, PA 15260. E-mail: [email protected] Journal of Voice, Vol. 20, No. 2, pp. 282–290 0892-1997/$32.00 쑕 2006 The Voice Foundation doi:10.1016/j.jvoice.2005.03.009
Accepted for publication March 11, 2005. Presented at the Voice Foundation’s 30th Annual Symposium: Care of the Professional Voice, June 2001, Philadelphia, Pennsylvania. From the *Beth Israel Deaconess Medical Center, Voice/ Speech/Swallowing Service, Rehabilitation Services Department, Boston, Massachusetts; †Communication Science and Disorders, School of Health and Rehabilitation Sciences, University of Pittsburgh, Pittsburgh, Pennsylvania; and the ‡University of Pittsburgh Voice Center, Pittsburgh, Pennsylvania.
282
TOPICAL ANESTHESIA ON LARYNGEAL SECRETIONS INTRODUCTION Endoscopic imaging procedures are commonly used by otolaryngologists and speech-language pathologists to examine the larynx and its function. Topical anesthesia sprayed into the oro- or nasopharynx is used routinely by physicians for both flexible and rigid endoscopy to diminish the gag reflex and enhance patient comfort.1,2 As a class, topical anesthetics inhibit the conduction of nerve impulses at the site of application by temporarily blocking the permeation of sodium ions into the nerve cell.3 Anesthesia not only reduces sensation, but it also inhibits efferent output from the central nervous system thereby reducing motor control.4 Despite the clear benefits, controversy exists over the possible adverse side effects of anesthesia. This study addressed clinical reports about the effects of topical anesthesia on the larynx. Specifically, claims were investigated that anesthesia may influence the quantity or quality of vocal tract secretions and thereby confound the diagnostic process. Background is as follows. The most common anesthetics used by physicians for laryngeal examinations include Xylocaine (3% lidocaine hydrochloride and 0.25% phenylephrine) (AstraZeneca, Wilmington, DE), Hurricaine (20% lidocaine) (Beutlich, Waukkegan, IL), and Cetacaine (14% benzocaine, 2% butyl aminobenzoate, 2% tetracaine hydrochloride, 0.5% benzalkonium chloride, and 0.005% [acetyl] dimethyl ethyl ammonium bromide) (Cetylite Industries, Inc., Pennsauken, NJ). The root “-caine” refers to a synthetic alkaloid anesthetic. In laryngeal visualization, these agents are sprayed into the oropharynx or nasal passages before the insertion of a rigid or flexible endoscope. The numbing effect occurs within 1 to 2 minutes and persists for approximately 10 to 15 minutes. In some cases, the use of topical anesthetics by physicians is elective. In other cases, the gag response associated with endoscope placement may be pronounced, which makes local anesthetics necessary for adequate visualization of the larynx and pharynx. One caution regarding topical anesthetics is the potential for compromised patient safety. Relative to benzocaine, the anesthetic of interest in this study, the Physician’s Desk Reference lists possible side effects as light-headedness, drowsiness, tinnitus,
283
blurred or double vision, vomiting, confusion, and dizziness. More serious sequelae include possible cardiovascular responses such as bradycardia, hypotension, or cardiovascular collapse. In extremely rare cases, benzocaine can cause an allergic reaction producing edema or even anaphylaxis. Similar responses may be found with other topical anesthetic agents suitable for laryngeal endoscopy. In addition to the aforementioned rare, dramatic clinical sequelae, other negative side effects from topical anesthetics have been reported with specific reference to the laryngeal examination. Because of reduced proprioception induced by topical anesthetics, the use of these agents by physicians in laryngeal endoscopy may alter the vocalization and/or swallowing functions targeted by the examination.5,6 Clinical reports also indicate that topical anesthetics may cause changes in the quantity or quality of laryngeal secretions, specifically by generating secretion “balling” or “pooling” operationalized shortly.1,7 Each of these reactions can confound interpretation of laryngeal images, which potentially misleads diagnosis. Laryngeal secretion “balling” and “pooling” have been described as follows. Balling is the appearance of discrete accumulations of secretions on the vocal fold surface.1 Balling is thought to be a possible complication of topical anesthesia in laryngeal endoscopy1 or a sign of dehydration and thus increased secretion viscosity. The primary concern relative to endoscopy is that secretion balls can be mistaken for mass lesions of the vocal folds. In contrast to balling, secretion pooling involves the diffuse appearance of increased, thinned vocal tract secretions. Anesthesia-induced pooling may be caused by the temporary lack of sensory feedback to the region and thus a reduction in normal secretion clearance by swallowing.8 Also, laryngeal pooling may be caused by an abnormal physiological response to the anesthetic agent, which causes both an increase in secretion amount as well as reduced clearance.9 Regardless of the source, the obvious problem is that pooling may compromise the view of the larynx, which limits diagnostic accuracy. Another problem is that some clinicians consider secretion pooling an indicator of laryngeal irritation—a nonspecific term that refers to inflammation characterized by swelling, redness, or pain.3,10 Thus, Journal of Voice, Vol. 20, No. 2, 2006
284
JOY WALSH ET AL
pooling may create the impression of laryngeal inflammation where none exists. Although the issue has not been experimentally examined, it also seems possible that the effect of topical anesthesia on laryngeal secretions may depend on patient variables such as gender, hormonal status, and health of the laryngeal tissue. Particularly this last variable is of conceptual interest in this study. Phonotraumatic mass lesions such as nodules and polyps are sometimes associated with a characteristic “bridging” or “stranding” of secretions from one vocal fold to the other, typically at the midpoint of the musculomembranous vocal fold. Secretions are also observed to amass at the site of the lesions (see, for example, Jiang and Titze11). Regardless of the reason, it is clear that laryngeal secretions at the cordal midpoint are sometimes taken to signal the presence of phonotrauma. A hypothetical anesthesia-induced increase in secretions in this location could increase the likelihood of falsely diagnosing lesions. The only published study of which we are aware that has directly examined changes in laryngeal secretions subsequent to the application of topical anesthesia was conducted by Peppard and Bless.1 In that study, 10 adults (five men and five women) with normal larynges were examined before and after the application of Xylocaine (lidocaine 10%) to the oropharynx. No systematic effects of anesthesia were noted on laryngeal secretions. The present study built on the previous design, extending it in several ways. First, the number of subjects in this study was based on a priori power analyses. Second, the experimental design in this study included the evaluation of both healthy subjects and subjects with phonotraumatic lesions. Third, a placebo condition was incorporated into the study. Fourth, possible effects from gender and hormones were limited by examining only premenopausal, adult women. Fifth, the secretion rating scale was extended from a dichotomous one to a trichotomous one, to increase resolution in the data. The specific experimental questions were (1) does the application of Cetacaine administered to the oropharynx in typical clinical doses before rigid endoscopy influence the presence or amount of balling or pooling of laryngeal secretions? and (2) has any evidence been observed Journal of Voice, Vol. 20, No. 2, 2006
for the possibility that Cetacaine differentially influences secretion balling or pooling depending on the presence or absence of diagnosed phonotraumatic mass lesions? METHODS Subjects After approval of the protocol by the Institutional Review Board at Massachusetts Eye and Ear Infirmary, 20 premenopausal adult women (20–43 years of age) were recruited with flyers to local singing and acting teachers’ associations, local music and other colleges, and voice clinics. Although vocal expertise was irrelevant to the experimental questions, professionally oriented voice users were targeted for two reasons. First, these persons may be more motivated than others to undergo a laryngeal examination and thus to volunteer for the study. Second, the incidence of laryngeal pathology may be greater in the performing vocalist population than in the population at large.12–14 Thus, recruitment of these subjects would be likely to attract some persons with phonotraumatic lesions. The number of subjects was determined a priori assuming an effect size of “1” (change from “no secretions” to “slightly remarkable” secretions or from slightly remarkable” to “remarkable” secretions) between anesthesia and placebo conditions, and a standard deviation of 0.5 (same scale), for both balling and pooling. Those analyses indicated that 10 subjects would be sufficient per group (healthy versus pathology) to yield 99.9% power to detect a potential effect of anesthesia on balling or pooling, which was the primary research question. Power analyses were not conducted to determine the subject number that would be required to assess the secondary question regarding a potential difference in the effect of anesthesia on laryngeal secretions for subjects with healthy larynges versus subjects with a diagnosis of phonotraumatic lesions. Subject recruitment continued until 10 subjects had completed participation in each of two experimental groups: subjects with phonotraumatic lesions and subjects without lesions. Exclusion criteria included known or suspected allergy to any anesthesia, current illness including upper respiratory infection, and inability to tolerate
TOPICAL ANESTHESIA ON LARYNGEAL SECRETIONS the endoscopic examination without anesthetic. To minimize variation in the data because of possible across-and within-gender hormonal effects, the study was further limited to premenopausal women. Satisfaction of inclusion and exclusion criteria was determined during a preexamination interview. Subjects were unaware of the study’s purpose. Equipment We used a Kay Elemetrics (Kay Elemetrics Corporation, Lincoln Park, NJ) rigid endoscope, model 9105, connected to a Kay Elemetrics stroboscopic light source to visualize the vocal folds. We used a Panasonic SCCD video camera (Panasonic Corporation, Tokyo, Japan), coupled to the endoscope, to record audio and video signals of the full procedure on a Panasonic LF-D103 DVD-recorder. The endoscopic view of the procedure was simultaneously displayed on a Sony PVM-1440 QM model monitor (Sony Corporation, Tokyo, Japan). Electroglottography electrodes, part of the Kay system, were positioned on the thyroid lamina to trigger the strobe. Overview of the experiment Endoscopic procedures were conducted at the Voice and Speech Laboratory in the Massachusetts Eye and Ear Infirmary, by a certified speech-language pathologist experienced in the procedure. This clinician was blind to the experimental hypotheses. Subjects were scheduled for two endoscopic examinations on consecutive days and were randomly assigned to one of two subgroups, group A or group B. Group A underwent the no-anesthesia (placebo) examination first, followed by the anesthesia examination the next day, approximately 24 hours after the original examination. Group B received the same examinations in reverse order. The 24-hour interval was chosen because this window is sufficiently long for medication effects to dissipate, yet sufficiently brief to minimize the likelihood of phonotraumatic, hormonal, or health changes affecting the larynx between examinations. Furthermore, the 24-hour interval between examinations limited any systematic influence of time of day on the data. Procedures All subjects participated individually in the experiment. Each subject was first screened for satisfaction of inclusion and exclusion criteria and provided
285
informed consent. Then, the subject proceeded to receive a baseline examination. The purpose of the baseline examination was twofold: (1) to determine if subjects could tolerate the presence of the endoscope without anesthesia, and (2) to orient subjects to the experimental task. For the baseline examination and for all subsequent examinations in the experiment, subjects were seated in a leaned-forward position, with the forearms bearing the torso’s weight on the thighs and the chin pointed forward. The rigid endoscope was then introduced into the oral cavity while the examining speech-language pathologist held the subject’s protruded tongue with a 4 × 4 piece of gauze and the subject produced a sustained /i/ at a comfortable pitch and loudness. We noted fundamental frequency and intensity from the first examination, and we used them for all subsequent visualizations in the experiment. Digital recording was initiated once the folds were in full view during sustained phonation. A minimum of three 3-second segments of that view was recorded. After baseline recordings, subjects received either the anesthesia or the placebo condition, depending on group assignment. Treatment bottles for both conditions were covered to mask identifying properties from both the subject and the endoscopist. A research associate handed the appropriate bottle to the endoscopist depending on the condition. For the anesthesia condition, four, 0.5-second (approximately) sprays of Cetacaine were delivered to the faucal pillars, uvula, and soft palate, for a total of 2 seconds. For the placebo condition, a clean, empty glass bottle with a squeeze-bulb applicator was sprayed air to the identical locations, for the same duration. It is acknowledged that although every attempt was made to keep the examiner “blinded” to the experimental condition, patient response to the aerated solution may have broken the blind in several cases. After a latency of 1 minute, the subject was instructed to clear her throat and swallow. The examination then proceeded exactly as in the baseline condition just described. After recordings for day 1 were completed, the subject received instructions regarding voice hygiene, to be followed between the two examinations. Specifically, the subject was instructed to minimize the use of diuretic agents including alcohol, and to Journal of Voice, Vol. 20, No. 2, 2006
286
JOY WALSH ET AL
limit or avoid behaviors that might be phonotraumatic. The purpose of those instructions was to limit the probability of extraneous changes in laryngeal tissue between days 1 and 2 of the experiment. Confirmation of adherence to these instructions was obtained during the follow-up examination, described next. The subject then returned approximately 24 hours later to receive the second examination—anesthesia or placebo, whichever condition had not been used by the clinician on day 1. Baseline views of the larynx were obtained as for day 1, followed by the experimental examination, depending on group assignment. After completion of all data capture, the subject was debriefed regarding the experimental hypotheses. Data management After data collection for each subject, a voicespecialized otolaryngologist was shown the video of the full baseline laryngeal examination from the first experimental day for each subject. The otolaryngologist classified each subject into three mutually exclusive categories: healthy, phonotraumatic lesions (only), or other abnormal findings. Only data from subjects in the first two groups were included in data analysis. In some cases, more than three /i/’s had been recorded for a given condition. When this situation occurred, the three “best images” (vocal folds in full view from anterior to posterior commissure, larynx, and phonation stable) were chosen by the endoscopist, who was uninformed about the experimental hypotheses. In the editing process, images that had been selected from the anesthesia and placebo conditions for each subject were dubbed back to back, as a set. Presentation of condition order was randomly varied across subjects. Twenty percent of all images (4), or 10% of data sets (2), were included twice in the data set so that measurement reliability could be assessed. Data scoring Two speech-language pathologists and one otolaryngologist experienced in videolaryngoscopy interpretation served as judges for perceptual ratings of secretions in the laryngeal images. Judges were uninformed about the experimental hypotheses and about Journal of Voice, Vol. 20, No. 2, 2006
subjects’ conditions. For training, judges were provided black-and-white images of larynges, including a normal larynx (without obvious secretions), a larynx that previous authors had described as showing pooling, and a larynx that those authors had described as showing balling.1 Judges were then shown video images that the investigators considered to reflect “slightly remarkable pooling,” “remarkable pooling,” “slightly remarkable balling,” and “remarkable balling.” These images were selected from the database at the Voice and Speech Laboratory of the Massachusetts Eye and Ear Infirmary, based on prior descriptions by Peppard and Bless.1 Before initiating scoring, judges were asked if they considered that the samples were acceptable anchors for the target phenomena. Agreement was unanimous. After training and anchor presentation, judges rated the target laryngeal images. Judges were seated in the same room and viewed the examinations simultaneously. However, judges made their ratings individually and silently. Judges were asked to determine whether the video clip under review was normal (ie, no secretion formation) or whether “slightly remarkable pooling,” “remarkable pooling,” “slightly remarkable balling,” or “remarkable balling” was noted. This scale was conceptually similar to the one used by Peppard and Bless in the previous study,1 extending the dichotomous scale in the earlier study (“remarkable” versus “unremarkable” secretions) to a scale with slightly greater resolution. Once all three judges recorded their ratings for an individual examination, they announced their ratings aloud. If all three judges’ ratings were identical, they proceeded to rating of the next video clip. If any disagreement was discovered, the judges came to a verbal, unanimous consensus. Judges were asked to rate each segment as it was shown. However, any examination or any portion of an examination could be repeated or played frame by frame at any judge’s request. This particular type of consensus perceptual analysis has been described previously in the literature.15 The methods employed have specific advantages with regard to both time and reliability. However, rater bias may confound results. Every attempt was made to limit this potential confound.
TOPICAL ANESTHESIA ON LARYNGEAL SECRETIONS
RESULTS Data reliability Test–retest data on 10% of data sets indicated 100% agreement across scores. Thus, the scores’ reliability was obviously high, which renders continued statistical evaluation of it irrelevant.
strated secretion balling than pooling. Because these data do not relate to any experimental hypotheses, no statistical tests were conducted on them and they are reported descriptively only. Effect of anesthesia on secretions Findings regarding the effects of anesthesia on secretion pooling and balling, independent of subject type or treatment order, are shown in Figures 1A and B. In those figures, “proportion of subjects” refers to the proportion of persons having either “slightly remarkable” or “remarkable” secretion scores. In brief, this data set failed to reveal differences in the presence or severity of either balling or pooling as a function of treatment type (anesthesia versus placebo; P ⫽ 0.317 and P ⫽ 0.564 for pooling and balling, respectively). Therefore, there was no clear evidence that anesthesia influenced secretion presence within the laryngeal area, for the collapsed subject group.
Proportion of Subjects with Notable Secretion Findings
A
Balling Across All Subjects 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Anesthesia
Placebo Treatment Condition
B Proportion of Subjects with Notable Secretion Findings
Statistical management The judges’ ratings of 0 (no secretion formation), 1 (slight balling or pooling), and 2 (remarkable balling or pooling) constituted trinomial data for each dependent variable (balling or pooling). Thus, nonparametric statistical analyses were required. Wilcoxon matched-pairs signed-ranks tests were employed to address both experimental questions (does Cetacaine increase the presence of laryngeal secretion balling or pooling, and does Cetacaine differentially affect secretion balling or pooling differently for subjects with healthy versus phonotraumatically injured vocal folds?). Post hoc analysis used the Fisher exact test to explore the effect of treatment order (anesthesia first versus placebo first) on balling and pooling, collapsing across both treatment and subject type. This test employs a contingency table to display how the order of presentation produces different outcomes. More specifically, to address the two experimental questions, six statistical tests were conducted. One test assessed differences in secretion balling as a function of anesthesia versus placebo conditions, and a second test assessed the same differences for pooling. Four tests were then employed to look at the effects of anesthesia on secretion pooling and balling in both subject groups (healthy subjects and those subjects with phonotraumatic lesions). To optimize the likelihood of finding effects if they exist, the investigation-wide α-level was set at 0.10 and then adjusted to account for the six tests. As a result, the criterion level for each statistical test was set to α ⫽ 0.017. Post hoc analysis was not accounted for in the α adjustment. Reliability of the data would be assessed by examining the proportion of test–retest scores for which exact agreement was obtained, with a Student t test.
287
Pooling Across All Subjects 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Anesthesia
Placebo Treatment Condition
Secretion results General results Collapsing across examination conditions and subject type, numerically more subjects demon-
FIGURE 1. A. The proportion of all subjects with a balling ranking of “1” or greater in the anesthesia versus placebo condition. B. The proportion of all subjects with a pooling ranking of “1” or greater in the anesthesia versus placebo condition. Journal of Voice, Vol. 20, No. 2, 2006
288
JOY WALSH ET AL A
Balling in Healthy Subjects 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Anesthesia
Placebo
Proportion of Subjects with Notable Findings
Proportion of Subjects with Notable Findings
A
Balling in Subjects with Phonotrauma 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Anesthesia
Treatment Condition
Pooling in Healthy Subjects 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Anesthesia
Placebo Treatment Condition
FIGURE 2. A. The proportion of healthy subjects with a balling ranking of “1” or greater in the anesthesia versus placebo condition. B. The proportion of healthy subjects with phonotrauma with a pooling ranking of “1” of greater in the anesthesia versus placebo condition.
Effect of anesthesia on secretions as a function of subject type The occurrence and severity of balling and pooling as a function of treatment condition (anesthesia versus placebo) is shown in Figures 2A and B for healthy subjects and in Figures 3A and B for subjects with phonotrauma. Numerically, one more subject with lesions had balling in the anesthesia condition than in the placebo condition (Figure 3A), and the same finding was seen for the pooling condition (Figure 3B). In contrast, for healthy subjects, one less subject showed pooling in the anesthesia condition as compared with the placebo (Figure 2B), and the same number of subjects showed balling in both conditions (Figure 2A). None of these numeric observations approached anything like statistical robustness (for balling in healthy subjects, P ⫽ 0.500; for pooling in healthy subjects, P ⫽ 0.375; for balling in subjects with lesions, P ⫽ 0.500; for pooling in subjects with lesions, P ⫽ 0.500). Journal of Voice, Vol. 20, No. 2, 2006
B Proportion of Subjects with Notable Findings
Proportion of Subjects with Notable Secretion Findings
B
Placebo Treatment Condition
Pooling in Subjects with Phonotrauma
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Anesthesia
Placebo Treatment Condition
FIGURE 3. A. The proportion of subjects with phonotrauma with a balling ranking of “1” or greater in the anesthesia versus placebo condition. B. The proportion of subjects with phonotrauma with a pooling ranking of “1” or greater in the anesthesia versus placebo condition.
Effect of treatment order on secretions The order of treatments was counterbalanced across subjects to limit the likelihood of order effects in the data. However, to assess a possible interaction of treatment type and treatment order, potential order effects were examined both descriptively and statistically. Results are shown in Figures 4 and 5. Figure 4 shows that more subjects who received the anesthesia condition first had balling across sessions than did subjects who received the placebo first. This effect was significant (P ⫽ 0.005). Figure 5 shows that more subjects who received the placebo condition first showed pooling than did subjects who received anesthesia first. However, that difference did not achieve significance (P ⫽ 0.65). Examination of other potential interactions of the data with order effects failed to elucidate any other findings. DISCUSSION The central experimental questions were (1) does the application of Cetacaine administered to the
TOPICAL ANESTHESIA ON LARYNGEAL SECRETIONS 1 Proportion of all subjects with Notable Secretion Pooling
1 Proportion of all subjects with Notable Secretion Balling
289
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
0 Anesth 1st
Placebo 1st Treatment condition
Anesth 1st
Placebo 1st Treatment condition
FIGURE 4. The proportion of all subjects with a secretion ranking of “1” or greater for secretion balling as a function of the order of treatment presentation.
FIGURE 5. The proportion of all subjects with a secretion ranking of “1” or greater for secretion pooling as a function of the order of treatment presentation.
oropharynx in typical clinical doses before rigid endoscopy influence the presence or quantity balling or pooling of laryngeal secretions? and (2) has any evidence been found for the possibility that Cetacaine may differentially influence secretion balling or pooling depending on the presence or absence of phonotraumatic mass lesions? Data from this study failed to find evidence of an affirmative answer to either experimental question, for the premenopausal women examined. That is, no clear differences in either secretion balling or pooling were observed with the administration of anesthesia, as compared with a placebo treatment, in this subject set. This finding constitutes a replication of reports by previous authors who also examined men,1 despite our use of extended experimental controls in this study. Relative to the second question, data from this study failed to find any clear evidence of a differential effect of anesthesia on laryngeal secretions for healthy subjects versus subjects with phonotrauma. Statistically, no evidence was found for an increase in either balling or pooling in either subject group, with anesthesia. Numerically, one more subject with lesioned vocal folds showed both balling and pooling in the anesthesia condition, as compared with the placebo condition. For the healthy subjects, one less subject showed pooling in the anesthesia condition compared with the placebo, and the same number of subjects showed balling in the two conditions for that group. A reasonable conclusion is that currently, no robust evidence is available that topical anesthesia
during rigid endoscopy affects the presence or quantity of laryngeal secretions, in any systematic way. Clearly, significance cannot be imputed to null results. However, confidence in the findings is increased by their replication of previous reports.1 The data do not imply that secretion changes do not occur with anesthesia. Individual differences in response to anesthesia could well provoke changes in some cases.9 The data simply say that any effects of topical anesthesia on secretions do not seem systematic. A final observation based on our findings is related to possible order effects in the data. Specifically, subjects who received the anesthesia condition first tended to display balling with greater frequency, across all examinations, as compared with subjects who received the placebo condition first. This finding is not altogether straightforward to interpret, and frankly, it may be spurious. However, in the interest of thoroughness, some speculations can be made. One possible candidate to explain the apparent effect might be emotional or physiological arousal associated with the anesthetic. Explanations around this somewhat convoluted hypothesis require a brief discussion of possible origins of secretion balling. Arguably, balling may sometimes be governed by the sympathetic nervous system (SNS)’s “fight or flight” response, which induces secretion drying. If the anesthetic condition was more aversive than the placebo for some subjects, SNS activity could have increased in those subjects, thus drying vocal fold secretions and increasing the frequency of balling on day 1 of the procedures. In that event, increases in balling on day 2 might Journal of Voice, Vol. 20, No. 2, 2006
290
JOY WALSH ET AL
be seen because of expectations of an aversive experience. Exploration of this hypothesis might be pursued in future studies by assessing emotional and physiological variables associated with endoscopy with and without topical anesthesia.
the third author. The authors thank Elaine Rubinstein, Ph.D., for statistical consulting.
CONCLUSIONS
1. Peppard RC, Bless DM. The use of topical anesthesia in videostroboscopic examination of the larynx. J Voice. 1991; 5:57–63. 2. Sataloff RT. Physical examination. In: Sataloff RT, ed. Professional Voice-The Science and Art of Clinical Care. San Diego, CA: Singular Publishing Group; 1997:208–209. 3. O’Toole M. Miller-Keane Encyclopedia and Dictionary of Medicine, Nursing, and Allied Health. 6th ed. Philadelphia, PA: W.B. Saunders Company; 1997. 4. Martin FG. Drugs and vocal function. J Voice. 1988;2: 338–344. 5. Selkin SG. Other clinical applications of flexible fiberoptic endoscopy. Cleft Palate J. 1984;21:29–32. 6. Yanagisawa E. Fiberscopic and telescopic videolaryngoscopy—a comparative study. In: Baer T, Sasaki C, Harris K, eds. Laryngeal Function in Phonation and Respiration. Boston, MA: College Hill Press; 1987:475–484. 7. Verdolini K, Hess MM, Titze IR, Bierhals W, Gross M. Investigation of vocal fold impact stress in human subjects. J Voice. 1999;13:184–202. 8. Leonard RJ, Ringel RL. Vocal shadowing under conditions of normal and altered laryngeal sensation. J Speech Hear Res. 1979;22:794–817. 9. Hess MM, Verdolini K, Bierhals W, Mansmann U, Gross M. Endolaryngeal contact pressures. J Voice. 1998;12:50–67. 10. Brodnitz F. Vocal Rehabilitation. Rochester, MN: American Academy of Ophthalmology and Otolaryngology; 1971. 11. Jiang JJ, Titze IR. Measurement of vocal fold intraglottal pressure and impact stress. J Voice. 1994;8:132–144. 12. Miller MK, Verdolini K. Frequency and risk factors for voice problems in teachers of singing and control subjects. J Voice. 1995;9:348–362. 13. Titze IR, Lemke J, Montequin D. Populations in the U.S. workforce who rely on voice as a primary tool for trade. J Voice. 1997;11:254–259. 14. Verdolini K, Ramig LO. Review: occupational risks for voice problems. Logoped Phoniatr Vocol. 2001;26:37–46. 15. Hartelius L, Theodoros D, Cahill L, Lilvik M. Comparability of perceptual analysis of speech characteristics in Australian and Swedish speakers with multiple sclerosis. Folia Phoniatr Logop. 2003;55:177–188.
As for a previous report,1 data from this study failed to find any clear effect of topical anesthesia on the presence or quantity of laryngeal secretions, despite the inclusion of extended experimental controls. Descriptive examination of the data found extremely weak evidence that the frequency of both secretion balling and pooling increased in subjects with phonotraumatic lesions, in the anesthesia condition, but this evidence did not even vaguely approach statistical significance. No evidence whatsoever was observed of an effect of anesthesia on either balling or pooling in healthy subjects. However, individual differences in response to topical anesthesia are not excluded. Findings may suggest an order effect, which indicates greater frequency of secretion balling in general when topical anesthesia is used by a physician during an initial laryngeal examination. However, those findings are not straightforward to interpret and may be spurious. A possible explanation—to be pursued in future studies—is that order effects may relate to “emotional” responses to anesthesia, and are played out by physiological pathways in the autonomic nervous system. If future pursuit of this finding confirms it, clinical implications could suggest (1) restricting topical anesthetic in laryngeal examination to those cases clearly requiring it and (2) interpreting the findings of laryngeal examinations with anesthetic considering the “emotional” hypothesis. Acknowledgments: This manuscript was prepared as the first author’s Master’s thesis in Communication Sciences and Disorders at the Massachusetts General Hospital Institute of Health Professions, under the direction of
Journal of Voice, Vol. 20, No. 2, 2006
REFERENCES
Summary: The application of topical anesthesia to the oropharynx is a common clinical practice during oral and nasal laryngoscopy. Clinically, questions have been raised about whether topical anesthesia alters laryngeal secretions, which distorts clinical impressions. A double-blind, placebo controlled design was employed to address this issue. Ten premenopausal women with healthy vocal folds and 10 premenopausal women with phonotraumatic lesions underwent oral videolaryngoscopic examinations on subsequent days under both anesthesia and placebo conditions, in counterbalanced order. Video segments were rated by three judges. Dependent variables were balling and pooling of secretions, as previously described in the literature. Statistical analyses failed to reveal any clear effect of topical anesthesia on either secretion balling or pooling for the collapsed data set, but one cannot exclude changes in individual cases. Moreover, there was no evidence that secretions were differentially affected by anesthesia across subject groups. Null results in this data set replicate and extend previously reported findings by other authors.1 An incidental but potentially interesting finding was that the order of treatment condition (anesthesia versus placebo first) seemed relevant for secretions: Subjects who received the anesthesia condition first tended to show more secretion balling in general, as compared with subjects who received the placebo condition first. Speculation is entertained regarding possible physiological pathways for these incidental findings, which could be relevant for some clinical practice. Key Words: Anesthesia—Vocal folds—Secretions.
Supported in part by the National Institute on Deafness and Other Communication Disorders (RO1 DC005643). Address correspondence and reprint requests to Katherine Verdolini, Communication Science and Disorders, School of Health and Rehabilitation Sciences, 4033 Forbes Tower, Pittsburgh, PA 15260. E-mail: [email protected] Journal of Voice, Vol. 20, No. 2, pp. 282–290 0892-1997/$32.00 쑕 2006 The Voice Foundation doi:10.1016/j.jvoice.2005.03.009
Accepted for publication March 11, 2005. Presented at the Voice Foundation’s 30th Annual Symposium: Care of the Professional Voice, June 2001, Philadelphia, Pennsylvania. From the *Beth Israel Deaconess Medical Center, Voice/ Speech/Swallowing Service, Rehabilitation Services Department, Boston, Massachusetts; †Communication Science and Disorders, School of Health and Rehabilitation Sciences, University of Pittsburgh, Pittsburgh, Pennsylvania; and the ‡University of Pittsburgh Voice Center, Pittsburgh, Pennsylvania.
282
TOPICAL ANESTHESIA ON LARYNGEAL SECRETIONS INTRODUCTION Endoscopic imaging procedures are commonly used by otolaryngologists and speech-language pathologists to examine the larynx and its function. Topical anesthesia sprayed into the oro- or nasopharynx is used routinely by physicians for both flexible and rigid endoscopy to diminish the gag reflex and enhance patient comfort.1,2 As a class, topical anesthetics inhibit the conduction of nerve impulses at the site of application by temporarily blocking the permeation of sodium ions into the nerve cell.3 Anesthesia not only reduces sensation, but it also inhibits efferent output from the central nervous system thereby reducing motor control.4 Despite the clear benefits, controversy exists over the possible adverse side effects of anesthesia. This study addressed clinical reports about the effects of topical anesthesia on the larynx. Specifically, claims were investigated that anesthesia may influence the quantity or quality of vocal tract secretions and thereby confound the diagnostic process. Background is as follows. The most common anesthetics used by physicians for laryngeal examinations include Xylocaine (3% lidocaine hydrochloride and 0.25% phenylephrine) (AstraZeneca, Wilmington, DE), Hurricaine (20% lidocaine) (Beutlich, Waukkegan, IL), and Cetacaine (14% benzocaine, 2% butyl aminobenzoate, 2% tetracaine hydrochloride, 0.5% benzalkonium chloride, and 0.005% [acetyl] dimethyl ethyl ammonium bromide) (Cetylite Industries, Inc., Pennsauken, NJ). The root “-caine” refers to a synthetic alkaloid anesthetic. In laryngeal visualization, these agents are sprayed into the oropharynx or nasal passages before the insertion of a rigid or flexible endoscope. The numbing effect occurs within 1 to 2 minutes and persists for approximately 10 to 15 minutes. In some cases, the use of topical anesthetics by physicians is elective. In other cases, the gag response associated with endoscope placement may be pronounced, which makes local anesthetics necessary for adequate visualization of the larynx and pharynx. One caution regarding topical anesthetics is the potential for compromised patient safety. Relative to benzocaine, the anesthetic of interest in this study, the Physician’s Desk Reference lists possible side effects as light-headedness, drowsiness, tinnitus,
283
blurred or double vision, vomiting, confusion, and dizziness. More serious sequelae include possible cardiovascular responses such as bradycardia, hypotension, or cardiovascular collapse. In extremely rare cases, benzocaine can cause an allergic reaction producing edema or even anaphylaxis. Similar responses may be found with other topical anesthetic agents suitable for laryngeal endoscopy. In addition to the aforementioned rare, dramatic clinical sequelae, other negative side effects from topical anesthetics have been reported with specific reference to the laryngeal examination. Because of reduced proprioception induced by topical anesthetics, the use of these agents by physicians in laryngeal endoscopy may alter the vocalization and/or swallowing functions targeted by the examination.5,6 Clinical reports also indicate that topical anesthetics may cause changes in the quantity or quality of laryngeal secretions, specifically by generating secretion “balling” or “pooling” operationalized shortly.1,7 Each of these reactions can confound interpretation of laryngeal images, which potentially misleads diagnosis. Laryngeal secretion “balling” and “pooling” have been described as follows. Balling is the appearance of discrete accumulations of secretions on the vocal fold surface.1 Balling is thought to be a possible complication of topical anesthesia in laryngeal endoscopy1 or a sign of dehydration and thus increased secretion viscosity. The primary concern relative to endoscopy is that secretion balls can be mistaken for mass lesions of the vocal folds. In contrast to balling, secretion pooling involves the diffuse appearance of increased, thinned vocal tract secretions. Anesthesia-induced pooling may be caused by the temporary lack of sensory feedback to the region and thus a reduction in normal secretion clearance by swallowing.8 Also, laryngeal pooling may be caused by an abnormal physiological response to the anesthetic agent, which causes both an increase in secretion amount as well as reduced clearance.9 Regardless of the source, the obvious problem is that pooling may compromise the view of the larynx, which limits diagnostic accuracy. Another problem is that some clinicians consider secretion pooling an indicator of laryngeal irritation—a nonspecific term that refers to inflammation characterized by swelling, redness, or pain.3,10 Thus, Journal of Voice, Vol. 20, No. 2, 2006
284
JOY WALSH ET AL
pooling may create the impression of laryngeal inflammation where none exists. Although the issue has not been experimentally examined, it also seems possible that the effect of topical anesthesia on laryngeal secretions may depend on patient variables such as gender, hormonal status, and health of the laryngeal tissue. Particularly this last variable is of conceptual interest in this study. Phonotraumatic mass lesions such as nodules and polyps are sometimes associated with a characteristic “bridging” or “stranding” of secretions from one vocal fold to the other, typically at the midpoint of the musculomembranous vocal fold. Secretions are also observed to amass at the site of the lesions (see, for example, Jiang and Titze11). Regardless of the reason, it is clear that laryngeal secretions at the cordal midpoint are sometimes taken to signal the presence of phonotrauma. A hypothetical anesthesia-induced increase in secretions in this location could increase the likelihood of falsely diagnosing lesions. The only published study of which we are aware that has directly examined changes in laryngeal secretions subsequent to the application of topical anesthesia was conducted by Peppard and Bless.1 In that study, 10 adults (five men and five women) with normal larynges were examined before and after the application of Xylocaine (lidocaine 10%) to the oropharynx. No systematic effects of anesthesia were noted on laryngeal secretions. The present study built on the previous design, extending it in several ways. First, the number of subjects in this study was based on a priori power analyses. Second, the experimental design in this study included the evaluation of both healthy subjects and subjects with phonotraumatic lesions. Third, a placebo condition was incorporated into the study. Fourth, possible effects from gender and hormones were limited by examining only premenopausal, adult women. Fifth, the secretion rating scale was extended from a dichotomous one to a trichotomous one, to increase resolution in the data. The specific experimental questions were (1) does the application of Cetacaine administered to the oropharynx in typical clinical doses before rigid endoscopy influence the presence or amount of balling or pooling of laryngeal secretions? and (2) has any evidence been observed Journal of Voice, Vol. 20, No. 2, 2006
for the possibility that Cetacaine differentially influences secretion balling or pooling depending on the presence or absence of diagnosed phonotraumatic mass lesions? METHODS Subjects After approval of the protocol by the Institutional Review Board at Massachusetts Eye and Ear Infirmary, 20 premenopausal adult women (20–43 years of age) were recruited with flyers to local singing and acting teachers’ associations, local music and other colleges, and voice clinics. Although vocal expertise was irrelevant to the experimental questions, professionally oriented voice users were targeted for two reasons. First, these persons may be more motivated than others to undergo a laryngeal examination and thus to volunteer for the study. Second, the incidence of laryngeal pathology may be greater in the performing vocalist population than in the population at large.12–14 Thus, recruitment of these subjects would be likely to attract some persons with phonotraumatic lesions. The number of subjects was determined a priori assuming an effect size of “1” (change from “no secretions” to “slightly remarkable” secretions or from slightly remarkable” to “remarkable” secretions) between anesthesia and placebo conditions, and a standard deviation of 0.5 (same scale), for both balling and pooling. Those analyses indicated that 10 subjects would be sufficient per group (healthy versus pathology) to yield 99.9% power to detect a potential effect of anesthesia on balling or pooling, which was the primary research question. Power analyses were not conducted to determine the subject number that would be required to assess the secondary question regarding a potential difference in the effect of anesthesia on laryngeal secretions for subjects with healthy larynges versus subjects with a diagnosis of phonotraumatic lesions. Subject recruitment continued until 10 subjects had completed participation in each of two experimental groups: subjects with phonotraumatic lesions and subjects without lesions. Exclusion criteria included known or suspected allergy to any anesthesia, current illness including upper respiratory infection, and inability to tolerate
TOPICAL ANESTHESIA ON LARYNGEAL SECRETIONS the endoscopic examination without anesthetic. To minimize variation in the data because of possible across-and within-gender hormonal effects, the study was further limited to premenopausal women. Satisfaction of inclusion and exclusion criteria was determined during a preexamination interview. Subjects were unaware of the study’s purpose. Equipment We used a Kay Elemetrics (Kay Elemetrics Corporation, Lincoln Park, NJ) rigid endoscope, model 9105, connected to a Kay Elemetrics stroboscopic light source to visualize the vocal folds. We used a Panasonic SCCD video camera (Panasonic Corporation, Tokyo, Japan), coupled to the endoscope, to record audio and video signals of the full procedure on a Panasonic LF-D103 DVD-recorder. The endoscopic view of the procedure was simultaneously displayed on a Sony PVM-1440 QM model monitor (Sony Corporation, Tokyo, Japan). Electroglottography electrodes, part of the Kay system, were positioned on the thyroid lamina to trigger the strobe. Overview of the experiment Endoscopic procedures were conducted at the Voice and Speech Laboratory in the Massachusetts Eye and Ear Infirmary, by a certified speech-language pathologist experienced in the procedure. This clinician was blind to the experimental hypotheses. Subjects were scheduled for two endoscopic examinations on consecutive days and were randomly assigned to one of two subgroups, group A or group B. Group A underwent the no-anesthesia (placebo) examination first, followed by the anesthesia examination the next day, approximately 24 hours after the original examination. Group B received the same examinations in reverse order. The 24-hour interval was chosen because this window is sufficiently long for medication effects to dissipate, yet sufficiently brief to minimize the likelihood of phonotraumatic, hormonal, or health changes affecting the larynx between examinations. Furthermore, the 24-hour interval between examinations limited any systematic influence of time of day on the data. Procedures All subjects participated individually in the experiment. Each subject was first screened for satisfaction of inclusion and exclusion criteria and provided
285
informed consent. Then, the subject proceeded to receive a baseline examination. The purpose of the baseline examination was twofold: (1) to determine if subjects could tolerate the presence of the endoscope without anesthesia, and (2) to orient subjects to the experimental task. For the baseline examination and for all subsequent examinations in the experiment, subjects were seated in a leaned-forward position, with the forearms bearing the torso’s weight on the thighs and the chin pointed forward. The rigid endoscope was then introduced into the oral cavity while the examining speech-language pathologist held the subject’s protruded tongue with a 4 × 4 piece of gauze and the subject produced a sustained /i/ at a comfortable pitch and loudness. We noted fundamental frequency and intensity from the first examination, and we used them for all subsequent visualizations in the experiment. Digital recording was initiated once the folds were in full view during sustained phonation. A minimum of three 3-second segments of that view was recorded. After baseline recordings, subjects received either the anesthesia or the placebo condition, depending on group assignment. Treatment bottles for both conditions were covered to mask identifying properties from both the subject and the endoscopist. A research associate handed the appropriate bottle to the endoscopist depending on the condition. For the anesthesia condition, four, 0.5-second (approximately) sprays of Cetacaine were delivered to the faucal pillars, uvula, and soft palate, for a total of 2 seconds. For the placebo condition, a clean, empty glass bottle with a squeeze-bulb applicator was sprayed air to the identical locations, for the same duration. It is acknowledged that although every attempt was made to keep the examiner “blinded” to the experimental condition, patient response to the aerated solution may have broken the blind in several cases. After a latency of 1 minute, the subject was instructed to clear her throat and swallow. The examination then proceeded exactly as in the baseline condition just described. After recordings for day 1 were completed, the subject received instructions regarding voice hygiene, to be followed between the two examinations. Specifically, the subject was instructed to minimize the use of diuretic agents including alcohol, and to Journal of Voice, Vol. 20, No. 2, 2006
286
JOY WALSH ET AL
limit or avoid behaviors that might be phonotraumatic. The purpose of those instructions was to limit the probability of extraneous changes in laryngeal tissue between days 1 and 2 of the experiment. Confirmation of adherence to these instructions was obtained during the follow-up examination, described next. The subject then returned approximately 24 hours later to receive the second examination—anesthesia or placebo, whichever condition had not been used by the clinician on day 1. Baseline views of the larynx were obtained as for day 1, followed by the experimental examination, depending on group assignment. After completion of all data capture, the subject was debriefed regarding the experimental hypotheses. Data management After data collection for each subject, a voicespecialized otolaryngologist was shown the video of the full baseline laryngeal examination from the first experimental day for each subject. The otolaryngologist classified each subject into three mutually exclusive categories: healthy, phonotraumatic lesions (only), or other abnormal findings. Only data from subjects in the first two groups were included in data analysis. In some cases, more than three /i/’s had been recorded for a given condition. When this situation occurred, the three “best images” (vocal folds in full view from anterior to posterior commissure, larynx, and phonation stable) were chosen by the endoscopist, who was uninformed about the experimental hypotheses. In the editing process, images that had been selected from the anesthesia and placebo conditions for each subject were dubbed back to back, as a set. Presentation of condition order was randomly varied across subjects. Twenty percent of all images (4), or 10% of data sets (2), were included twice in the data set so that measurement reliability could be assessed. Data scoring Two speech-language pathologists and one otolaryngologist experienced in videolaryngoscopy interpretation served as judges for perceptual ratings of secretions in the laryngeal images. Judges were uninformed about the experimental hypotheses and about Journal of Voice, Vol. 20, No. 2, 2006
subjects’ conditions. For training, judges were provided black-and-white images of larynges, including a normal larynx (without obvious secretions), a larynx that previous authors had described as showing pooling, and a larynx that those authors had described as showing balling.1 Judges were then shown video images that the investigators considered to reflect “slightly remarkable pooling,” “remarkable pooling,” “slightly remarkable balling,” and “remarkable balling.” These images were selected from the database at the Voice and Speech Laboratory of the Massachusetts Eye and Ear Infirmary, based on prior descriptions by Peppard and Bless.1 Before initiating scoring, judges were asked if they considered that the samples were acceptable anchors for the target phenomena. Agreement was unanimous. After training and anchor presentation, judges rated the target laryngeal images. Judges were seated in the same room and viewed the examinations simultaneously. However, judges made their ratings individually and silently. Judges were asked to determine whether the video clip under review was normal (ie, no secretion formation) or whether “slightly remarkable pooling,” “remarkable pooling,” “slightly remarkable balling,” or “remarkable balling” was noted. This scale was conceptually similar to the one used by Peppard and Bless in the previous study,1 extending the dichotomous scale in the earlier study (“remarkable” versus “unremarkable” secretions) to a scale with slightly greater resolution. Once all three judges recorded their ratings for an individual examination, they announced their ratings aloud. If all three judges’ ratings were identical, they proceeded to rating of the next video clip. If any disagreement was discovered, the judges came to a verbal, unanimous consensus. Judges were asked to rate each segment as it was shown. However, any examination or any portion of an examination could be repeated or played frame by frame at any judge’s request. This particular type of consensus perceptual analysis has been described previously in the literature.15 The methods employed have specific advantages with regard to both time and reliability. However, rater bias may confound results. Every attempt was made to limit this potential confound.
TOPICAL ANESTHESIA ON LARYNGEAL SECRETIONS
RESULTS Data reliability Test–retest data on 10% of data sets indicated 100% agreement across scores. Thus, the scores’ reliability was obviously high, which renders continued statistical evaluation of it irrelevant.
strated secretion balling than pooling. Because these data do not relate to any experimental hypotheses, no statistical tests were conducted on them and they are reported descriptively only. Effect of anesthesia on secretions Findings regarding the effects of anesthesia on secretion pooling and balling, independent of subject type or treatment order, are shown in Figures 1A and B. In those figures, “proportion of subjects” refers to the proportion of persons having either “slightly remarkable” or “remarkable” secretion scores. In brief, this data set failed to reveal differences in the presence or severity of either balling or pooling as a function of treatment type (anesthesia versus placebo; P ⫽ 0.317 and P ⫽ 0.564 for pooling and balling, respectively). Therefore, there was no clear evidence that anesthesia influenced secretion presence within the laryngeal area, for the collapsed subject group.
Proportion of Subjects with Notable Secretion Findings
A
Balling Across All Subjects 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Anesthesia
Placebo Treatment Condition
B Proportion of Subjects with Notable Secretion Findings
Statistical management The judges’ ratings of 0 (no secretion formation), 1 (slight balling or pooling), and 2 (remarkable balling or pooling) constituted trinomial data for each dependent variable (balling or pooling). Thus, nonparametric statistical analyses were required. Wilcoxon matched-pairs signed-ranks tests were employed to address both experimental questions (does Cetacaine increase the presence of laryngeal secretion balling or pooling, and does Cetacaine differentially affect secretion balling or pooling differently for subjects with healthy versus phonotraumatically injured vocal folds?). Post hoc analysis used the Fisher exact test to explore the effect of treatment order (anesthesia first versus placebo first) on balling and pooling, collapsing across both treatment and subject type. This test employs a contingency table to display how the order of presentation produces different outcomes. More specifically, to address the two experimental questions, six statistical tests were conducted. One test assessed differences in secretion balling as a function of anesthesia versus placebo conditions, and a second test assessed the same differences for pooling. Four tests were then employed to look at the effects of anesthesia on secretion pooling and balling in both subject groups (healthy subjects and those subjects with phonotraumatic lesions). To optimize the likelihood of finding effects if they exist, the investigation-wide α-level was set at 0.10 and then adjusted to account for the six tests. As a result, the criterion level for each statistical test was set to α ⫽ 0.017. Post hoc analysis was not accounted for in the α adjustment. Reliability of the data would be assessed by examining the proportion of test–retest scores for which exact agreement was obtained, with a Student t test.
287
Pooling Across All Subjects 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Anesthesia
Placebo Treatment Condition
Secretion results General results Collapsing across examination conditions and subject type, numerically more subjects demon-
FIGURE 1. A. The proportion of all subjects with a balling ranking of “1” or greater in the anesthesia versus placebo condition. B. The proportion of all subjects with a pooling ranking of “1” or greater in the anesthesia versus placebo condition. Journal of Voice, Vol. 20, No. 2, 2006
288
JOY WALSH ET AL A
Balling in Healthy Subjects 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Anesthesia
Placebo
Proportion of Subjects with Notable Findings
Proportion of Subjects with Notable Findings
A
Balling in Subjects with Phonotrauma 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Anesthesia
Treatment Condition
Pooling in Healthy Subjects 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Anesthesia
Placebo Treatment Condition
FIGURE 2. A. The proportion of healthy subjects with a balling ranking of “1” or greater in the anesthesia versus placebo condition. B. The proportion of healthy subjects with phonotrauma with a pooling ranking of “1” of greater in the anesthesia versus placebo condition.
Effect of anesthesia on secretions as a function of subject type The occurrence and severity of balling and pooling as a function of treatment condition (anesthesia versus placebo) is shown in Figures 2A and B for healthy subjects and in Figures 3A and B for subjects with phonotrauma. Numerically, one more subject with lesions had balling in the anesthesia condition than in the placebo condition (Figure 3A), and the same finding was seen for the pooling condition (Figure 3B). In contrast, for healthy subjects, one less subject showed pooling in the anesthesia condition as compared with the placebo (Figure 2B), and the same number of subjects showed balling in both conditions (Figure 2A). None of these numeric observations approached anything like statistical robustness (for balling in healthy subjects, P ⫽ 0.500; for pooling in healthy subjects, P ⫽ 0.375; for balling in subjects with lesions, P ⫽ 0.500; for pooling in subjects with lesions, P ⫽ 0.500). Journal of Voice, Vol. 20, No. 2, 2006
B Proportion of Subjects with Notable Findings
Proportion of Subjects with Notable Secretion Findings
B
Placebo Treatment Condition
Pooling in Subjects with Phonotrauma
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Anesthesia
Placebo Treatment Condition
FIGURE 3. A. The proportion of subjects with phonotrauma with a balling ranking of “1” or greater in the anesthesia versus placebo condition. B. The proportion of subjects with phonotrauma with a pooling ranking of “1” or greater in the anesthesia versus placebo condition.
Effect of treatment order on secretions The order of treatments was counterbalanced across subjects to limit the likelihood of order effects in the data. However, to assess a possible interaction of treatment type and treatment order, potential order effects were examined both descriptively and statistically. Results are shown in Figures 4 and 5. Figure 4 shows that more subjects who received the anesthesia condition first had balling across sessions than did subjects who received the placebo first. This effect was significant (P ⫽ 0.005). Figure 5 shows that more subjects who received the placebo condition first showed pooling than did subjects who received anesthesia first. However, that difference did not achieve significance (P ⫽ 0.65). Examination of other potential interactions of the data with order effects failed to elucidate any other findings. DISCUSSION The central experimental questions were (1) does the application of Cetacaine administered to the
TOPICAL ANESTHESIA ON LARYNGEAL SECRETIONS 1 Proportion of all subjects with Notable Secretion Pooling
1 Proportion of all subjects with Notable Secretion Balling
289
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
0 Anesth 1st
Placebo 1st Treatment condition
Anesth 1st
Placebo 1st Treatment condition
FIGURE 4. The proportion of all subjects with a secretion ranking of “1” or greater for secretion balling as a function of the order of treatment presentation.
FIGURE 5. The proportion of all subjects with a secretion ranking of “1” or greater for secretion pooling as a function of the order of treatment presentation.
oropharynx in typical clinical doses before rigid endoscopy influence the presence or quantity balling or pooling of laryngeal secretions? and (2) has any evidence been found for the possibility that Cetacaine may differentially influence secretion balling or pooling depending on the presence or absence of phonotraumatic mass lesions? Data from this study failed to find evidence of an affirmative answer to either experimental question, for the premenopausal women examined. That is, no clear differences in either secretion balling or pooling were observed with the administration of anesthesia, as compared with a placebo treatment, in this subject set. This finding constitutes a replication of reports by previous authors who also examined men,1 despite our use of extended experimental controls in this study. Relative to the second question, data from this study failed to find any clear evidence of a differential effect of anesthesia on laryngeal secretions for healthy subjects versus subjects with phonotrauma. Statistically, no evidence was found for an increase in either balling or pooling in either subject group, with anesthesia. Numerically, one more subject with lesioned vocal folds showed both balling and pooling in the anesthesia condition, as compared with the placebo condition. For the healthy subjects, one less subject showed pooling in the anesthesia condition compared with the placebo, and the same number of subjects showed balling in the two conditions for that group. A reasonable conclusion is that currently, no robust evidence is available that topical anesthesia
during rigid endoscopy affects the presence or quantity of laryngeal secretions, in any systematic way. Clearly, significance cannot be imputed to null results. However, confidence in the findings is increased by their replication of previous reports.1 The data do not imply that secretion changes do not occur with anesthesia. Individual differences in response to anesthesia could well provoke changes in some cases.9 The data simply say that any effects of topical anesthesia on secretions do not seem systematic. A final observation based on our findings is related to possible order effects in the data. Specifically, subjects who received the anesthesia condition first tended to display balling with greater frequency, across all examinations, as compared with subjects who received the placebo condition first. This finding is not altogether straightforward to interpret, and frankly, it may be spurious. However, in the interest of thoroughness, some speculations can be made. One possible candidate to explain the apparent effect might be emotional or physiological arousal associated with the anesthetic. Explanations around this somewhat convoluted hypothesis require a brief discussion of possible origins of secretion balling. Arguably, balling may sometimes be governed by the sympathetic nervous system (SNS)’s “fight or flight” response, which induces secretion drying. If the anesthetic condition was more aversive than the placebo for some subjects, SNS activity could have increased in those subjects, thus drying vocal fold secretions and increasing the frequency of balling on day 1 of the procedures. In that event, increases in balling on day 2 might Journal of Voice, Vol. 20, No. 2, 2006
290
JOY WALSH ET AL
be seen because of expectations of an aversive experience. Exploration of this hypothesis might be pursued in future studies by assessing emotional and physiological variables associated with endoscopy with and without topical anesthesia.
the third author. The authors thank Elaine Rubinstein, Ph.D., for statistical consulting.
CONCLUSIONS
1. Peppard RC, Bless DM. The use of topical anesthesia in videostroboscopic examination of the larynx. J Voice. 1991; 5:57–63. 2. Sataloff RT. Physical examination. In: Sataloff RT, ed. Professional Voice-The Science and Art of Clinical Care. San Diego, CA: Singular Publishing Group; 1997:208–209. 3. O’Toole M. Miller-Keane Encyclopedia and Dictionary of Medicine, Nursing, and Allied Health. 6th ed. Philadelphia, PA: W.B. Saunders Company; 1997. 4. Martin FG. Drugs and vocal function. J Voice. 1988;2: 338–344. 5. Selkin SG. Other clinical applications of flexible fiberoptic endoscopy. Cleft Palate J. 1984;21:29–32. 6. Yanagisawa E. Fiberscopic and telescopic videolaryngoscopy—a comparative study. In: Baer T, Sasaki C, Harris K, eds. Laryngeal Function in Phonation and Respiration. Boston, MA: College Hill Press; 1987:475–484. 7. Verdolini K, Hess MM, Titze IR, Bierhals W, Gross M. Investigation of vocal fold impact stress in human subjects. J Voice. 1999;13:184–202. 8. Leonard RJ, Ringel RL. Vocal shadowing under conditions of normal and altered laryngeal sensation. J Speech Hear Res. 1979;22:794–817. 9. Hess MM, Verdolini K, Bierhals W, Mansmann U, Gross M. Endolaryngeal contact pressures. J Voice. 1998;12:50–67. 10. Brodnitz F. Vocal Rehabilitation. Rochester, MN: American Academy of Ophthalmology and Otolaryngology; 1971. 11. Jiang JJ, Titze IR. Measurement of vocal fold intraglottal pressure and impact stress. J Voice. 1994;8:132–144. 12. Miller MK, Verdolini K. Frequency and risk factors for voice problems in teachers of singing and control subjects. J Voice. 1995;9:348–362. 13. Titze IR, Lemke J, Montequin D. Populations in the U.S. workforce who rely on voice as a primary tool for trade. J Voice. 1997;11:254–259. 14. Verdolini K, Ramig LO. Review: occupational risks for voice problems. Logoped Phoniatr Vocol. 2001;26:37–46. 15. Hartelius L, Theodoros D, Cahill L, Lilvik M. Comparability of perceptual analysis of speech characteristics in Australian and Swedish speakers with multiple sclerosis. Folia Phoniatr Logop. 2003;55:177–188.
As for a previous report,1 data from this study failed to find any clear effect of topical anesthesia on the presence or quantity of laryngeal secretions, despite the inclusion of extended experimental controls. Descriptive examination of the data found extremely weak evidence that the frequency of both secretion balling and pooling increased in subjects with phonotraumatic lesions, in the anesthesia condition, but this evidence did not even vaguely approach statistical significance. No evidence whatsoever was observed of an effect of anesthesia on either balling or pooling in healthy subjects. However, individual differences in response to topical anesthesia are not excluded. Findings may suggest an order effect, which indicates greater frequency of secretion balling in general when topical anesthesia is used by a physician during an initial laryngeal examination. However, those findings are not straightforward to interpret and may be spurious. A possible explanation—to be pursued in future studies—is that order effects may relate to “emotional” responses to anesthesia, and are played out by physiological pathways in the autonomic nervous system. If future pursuit of this finding confirms it, clinical implications could suggest (1) restricting topical anesthetic in laryngeal examination to those cases clearly requiring it and (2) interpreting the findings of laryngeal examinations with anesthetic considering the “emotional” hypothesis. Acknowledgments: This manuscript was prepared as the first author’s Master’s thesis in Communication Sciences and Disorders at the Massachusetts General Hospital Institute of Health Professions, under the direction of
Journal of Voice, Vol. 20, No. 2, 2006
REFERENCES