585867
research-article2015
AORXXX10.1177/0003489415585867Annals of Otology, Rhinology & LaryngologyHaft et al
Original Article
Anticholinergic Use Is a Major Risk Factor for Dysphonia
Annals of Otology, Rhinology & Laryngology 1–6 © The Author(s) 2015 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/0003489415585867 aor.sagepub.com
Sunny Haft, MD1, Douglas Farquhar, BA1, Ryan Carey, BSE1, and Natasha Mirza, MD1
Abstract Objective: We hypothesize that many cases of dysphonia of unclear etiology are a form of sicca caused by anticholinergic medication use, and we aim to determine their association. Study Design: This was a cross-sectional study conducted over a 6-month time period. Participants were drawn from a tertiary care laryngology practice within an academic institution. Methods: One hundred forty-nine patients met inclusion criteria. Patients rated the symptom of chronic hoarseness; scores were compared with participants’ medication lists, comorbidities, age, and sex, and a multivariate logistic regression model was developed. Significance was set at P < .05. As a secondary analysis, participants rated a variety of other symptoms using the Voice Handicap Index-10, Reflux Symptom Index, and the GRBAS scale, which were likewise compared to anticholinergic use. Results: Any patient taking at least 1 anticholinergic medication had a 2.32 increased odds (P = .03) of experiencing hoarseness. If the patient was taking 2 or more anticholinergic medications, those odds rose to 4.52 (P = .009). Conclusion: This is the first study, to our knowledge, that implicates medication use as a major risk factor for dysphonia of unclear etiology. An awareness of this association is invaluable when attributing cause to hoarseness and when considering treatment options. Keywords anticholinergic, dysphonia, GRBAS, hoarseness, laryngitis, medications, RSI, sicca, VHI, voice
Background Hoarseness is a common complaint within an otolaryngology practice. It is defined as a coarse or rough quality to the voice, which may be due to a wide range of underlying pathology.1 The work-up for dysphonia within a laryngology practice will often suggest an underlying diagnosis, which can include neoplasm, infection, phonotrauma, smoking, laryngopharyngeal reflux (LPR), or neurologic pathology, among others. Even with a thorough investigation, however, clinicians may still struggle to find a clear cause. Many cases of dysphonia have increasingly been ascribed to LPR within the laryngology literature, but the extent of causation is still not entirely clear. In cases of dysphonia of unclear etiology, recent studies have highlighted the importance of further laryngeal investigation rather than continued empiric reflux treatment or ongoing reflux investigation.2 Others have likewise shown that LPR is overdiagnosed within the laryngology clinic.3 We hypothesize that a significant degree of hoarseness of unclear etiology may be due to medication use—namely, anticholinergic medications—and is underdiagnosed within laryngology clinics.
No one has yet studied the correlation between medication use and dysphonia, and we aimed to establish and quantify this association. Particularly, we focused on patients who had persistent hoarseness without a clear cause despite work-up within the laryngology clinic. Anticholinergic medications may be contributing to dysphonia due to a drying effect. Vocal fold dehydration has been shown to cause a decrease in mucosal wave frequency and amplitude, increasing phonation threshold pressure and, thus, strain.4 In addition, a loss of mucosal lubrication may theoretically subject the vocal folds to higher degrees of friction and heat during vibration, leading to quicker molecular breakdown.5,6 Several lines of evidence support this possible causal relationship. Parasympathetic nerve fibers in the larynx 1
Department of Otorhinolaryngology–Head and Neck Surgery, University of Pennsylvania, Philadelphia, Pennsylvania, USA Corresponding Author: Sunny Haft, MD, Department of Otorhinolaryngology–Head and Neck Surgery, University of Pennsylvania, 3400 Spruce Street, 5 Ravdin, Philadelphia, PA 19104, USA. Email: [email protected]
Downloaded from aor.sagepub.com at Bobst Library, New York University on May 13, 2015
2
Annals of Otology, Rhinology & Laryngology
outnumber sympathetic fibers by 20 to 1,7 penetrating the basal lamina and directly stimulating secretion of mucus glands. In turn, the majority of anticholinergic medications are antimuscarinic, with the ability to block the release of acetylcholine at nerve junctions that would otherwise stimulate these mucus glands. Thus, anticholinergic medications may be decreasing glandular secretions, directly leading to sicca within the larynx. Second, studies have shown the negative phonatory effects of systemic dehydration on the larynx.4,8 Heavy coffee and tea drinking has been implicated in voice disorders,4 presumably due to its diuretic effect, although this has yet to be definitively proven. The laryngologist often advises the patient to stay hydrated and avoid coffee in order to combat hoarseness. Last, laryngitis sicca may be pathophysiologically very similar to xerostomia—the larynx being a mucosal extension of the oral cavity. Xerostomia has been extensively studied, with the top cause of dry mouth indeed being medication use. Scully et al9 found that anticholinergic medications are the most common reason for reduction in the amount of saliva, more so than even dehydration. Gunes et al10 collected saliva of patients on anticholinergic medications and found that they had 3.64 times less saliva than patients not on these medications. We hypothesize that the effect of anticholinergic medications on glandular tissue in the oral cavity is causing the same degree of dryness within the larynx, ultimately leading to hoarseness. An increased awareness of this link may be an important step in guiding physicians down a diagnostic and treatment regimen for dysphonia of unclear etiology.
Methods Study Population A cross-sectional study was performed to determine the relationship between anticholinergic medication use and laryngitis symptoms in adults. The sample population chosen was a tertiary-care laryngology practice. All consecutive patients presenting to a clinic over a 6-month period in 2013 were included. Patients typically presented with voice or other laryngeal complaints after referral from a primary care practice. Patients were excluded from analysis if they had 1 or more significant risk factors for laryngitis. These included persistent reflux refractory to medical treatment (patients currently taking an H2 antagonist or proton-pump inhibitor [PPI] with a concurrent high score [4 or 5] on questions on the Reflux Symptoms Index [RSI] about “heartburn” or “coughing after lying down”), professional voice overuse (singers, performers, and teachers who complained of worsening hoarseness in the context of their profession), current smokers who smoke more than a
half pack per day for greater than 1 year, a history of prior neck radiation, Sjogren’s syndrome, amyloidosis, recent laryngeal surgery or injection within the past month, tracheostomy dependency, and structural abnormalities seen on exam (papillomatosis, nodules, granulomas, or subglottic stenosis). Patients who met inclusion criteria were then divided into those positive for dysphonia and those without dysphonia, as determined by a self-reported hoarseness score, as described below. The Institutional Committee on Research Involving Human Subjects approved the study protocol; a written consent was obtained from all participants.
Medication Sampling and Classification Patient medication lists were obtained from chart review and reaffirmed on patient interview. All medications, dosing, and compliance were confirmed by the study authors during the clinical visit. Vitamins were excluded from analysis. To help establish a cause–effect relationship, all medications that were started following the onset of hoarseness were also excluded. Each medication taken by a participant was classified as an anticholinergic, diuretic, sympathomimetic, opioid, or “other.” Anticholinergics were identified according to 2 large studies that categorized anticholinergic burden.11,12 Independent assessments of anticholinergic properties of drugs were used to categorize medications not examined by these studies.
Measurement of Dysphonia To ensure reliability, laryngitis and dysphonia were measured with 2 subjective scales: the Voice Handicap Index-10 (VHI-10) and the RSI, as well as 1 objective scale, the GRBAS. The VHI-10 is a validated 10-question survey designed to measure dysphonia, and the RSI is a 9-item scale developed to measure laryngitis related to reflux.13,14 These were completed by each participant prior to seeing the physician in order to reduce bias. The GRBAS is a clinician-based voice assessment scale that evaluates Grade, Roughness, Breathiness, Asthenia, and Strain. The GRBAS was determined by a trained laryngologist during the patient encounter. We chose to use the symptom of hoarseness, as gathered from the RSI, as our primary measure of dysphonia.1 A score of 3 to 5 was chosen to signify moderate to severe hoarseness, and a score of 0 to 2 to signify mild to none. As a secondary analysis to improve reliability, the total scores on the VHI-10, RSI, and GRBAS scales were also compared to medication use. Scores ≥ 13 on the RSI14 and ≥ 5 on the GRBAS were considered positive for laryngitis. As a post-hoc assessment, each subscore on the VHI-10 was also evaluated.
Downloaded from aor.sagepub.com at Bobst Library, New York University on May 13, 2015
3
Haft et al
Assessment of Comorbidities Comorbidities for vocal dysfunction were assessed to control for potential confounders. A chart review was conducted on each patient, noting age, sex, rhinitis, sinusitis, asthma, seasonal allergies, chronic obstructive pulmonary disease (COPD), gastroesophageal reflux disease (GERD), and recent laryngeal surgery, if applicable.
Statistical Analyses Descriptive statistics were used to analyze baseline characteristics. Statistical comparisons were performed using the Wilcox rank-sum test, Student t test, or Fisher exact test, as appropriate. To assess the effect of potential confounders, 2 multivariate logistic regression models were constructed, using the hoarseness scale and the total RSI score as the outcome, and including age, sex, comorbidities for vocal dysfunction (sinusitis, rhinitis, asthma, COPD, allergies, GERD, previous head and neck surgery), and other classes of drugs (sympathomimetics, diuretics, and opioids) as independent variables. A P value of .05 was considered significant. All statistical analysis was performed using STATA/IC 13.0 software (Stata Corporation, College Station, Texas, USA).
Results Of 243 patients enrolled in the study, 149 were eligible for inclusion. The mean age was 59 years and 44% were male; 49% of the population took at least 1 anticholinergic medication, and 27% took 2 or more. The most common medications taken by patients are listed in Table 1. Seventy-five patients (50%) were considered positive for laryngitis, scoring ≥ 3 on the hoarseness scale, and 74 were considered negative (Figure 1). There were no significant differences between the 2 cohorts in demographics or vocal comorbidities.
Anticholinergic Medication Use Is a Risk Factor for Hoarseness The use of at least 1 anticholinergic medication was significantly associated with hoarseness (35% of patients negative for hoarseness vs 57% of positive patients; P = .036). There was a stronger association with taking at least 2 anticholinergics (9% vs 27%; P = .003). Patients positive for hoarseness were also more likely to take more total medications than their non-hoarse counterparts (means of 5.5 vs 4.3; P = .03). However, when the number of anticholinergic drugs were excluded from this analysis, the association was no longer significant (3.7 vs 4.5; P = .08). No other classes of drugs had any associations with laryngitis, including diuretics (n = 43, 27% vs 31%;
Table 1. Common Anticholinergic Medications Being Taken by Study Participants. Anticholinergic Medication Diazepam Prednisone Furosemide Alprazolam Oxycodone Amitriptyline Diphenhydramine Clonazepam Paroxetine Nortriptyline Temazepam Bupropion Isosorbide Atropine Meclizine Oxybutynin Diltiazem Tramadol
No. of Participants Taking Medication 14 14 14 13 12 11 11 8 7 6 5 4 4 3 3 3 3 3
P = .72), sympathomimetics (n = 6, 7% vs 1%; P = .11), and opioids (n = 11, 8% vs 7%; P = .77). When potential confounders such as GERD, asthma, allergies, and rhinitis were accounted for in a multivariate analysis, patients with hoarseness were 2.32 times as likely to be taking at least 1 anticholinergic as patients without (P = .031) and were 4.52 times as likely to be taking 2 or more anticholinergic medications (P = .009). Overall, the odds for hoarseness increased by 1.76 with each anticholinergic medication that a patient was taking (P = .014). The results from the multivariate analysis are summarized in Table 2. No other demographic factors or comorbidities were significantly associated with hoarseness in the multivariate analysis except for rhinitis, which was actually protective against hoarseness (odds ratio = 0.09, P = .03). It is notable that the rhinitis population had an n of 5.
Anticholinergic Medication Use and VHI-10 Score The VHI-10 scores significantly differed when comparing patients taking 2 or more anticholinergics versus no anticholinergics (mean scores of 16.8 vs 11.9; P = .023). There was no difference when comparing patients on 1 anticholinergic versus none. However, when analyzing each of the 10 questions separately as a post-hoc analysis, patients taking at least 1 anticholinergic scored significantly higher on 5 of the 10 questions, and patients taking at least 2 anticholinergics scored significantly higher on 9 of the 10 questions (Table 3).
Downloaded from aor.sagepub.com at Bobst Library, New York University on May 13, 2015
4
Annals of Otology, Rhinology & Laryngology
Figure 1. Study flowchart with patient characteristics listed. Patients were categorized as having laryngitis if they scored a 3 or above on a 0-5 severity scale for hoarseness. Table 2. Multivariate Analysis Results Comparing Hoarseness Scores to Multiple Variables.a Variable
No.
Odds Ratio
1 or more anticholinergic 2 or more anticholinergics Each anticholinergic Age Male sex Sinusitis Rhinitis Asthma COPD Allergies GERD Diuretics Opioids
69 27 69 149 149 15 5 6 7 20 75 43 11
2.32 4.52 1.76 1.00 0.67 1.36 0.09 0.17 2.54 5.32 0.49 1.12 0.53
P Value .03* .01* .01* .98 .30 .76 .03* .12 .35 .08 .06 .80 .43
Abbreviations: COPD, chronic obstructive pulmonary disease; GERD, gastroesophageal reflux disease. a Asterisks indicate statistical significance.
Anticholinergic Medication Use and RSI and GRBAS Scores On the RSI, 95 patients (63%) had an RSI score ≥ 13, and 54 (36%) had a score < 13. Of the 115 patients assessed with the GRBAS scale, 79 (69%) had a score ≥ 5 and 36 (31%) had a score < 5. The use of at least 1 anticholinergic medication was also significantly associated with a high RSI score (P = .001) and a high GRBAS score (P = .005). Likewise, the use of at least 2 anticholinergic medications was significantly
associated with a high RSI score (P < .001), although the association with a high GRBAS score was borderline significant (P = .06). Smoking was associated with a high GRBAS score (P = .01), but no other demographic factors, vocal comorbidities, or other classes of drugs were significantly associated with high scores on these measures. When using the RSI as the dependent variable in the multivariate model, every additional anticholinergic medication was associated with a 2.32 increased odds of the patient having a high score (P = .0001). A patient taking 1 or more anticholinergics had a 2.98 increased odds of having a high score (P = .006). If a patient was taking 2 or more anticholinergics, he or she had an 8.62 increased odds of having a high score (P = .0001).
Discussion This is the first study, to our knowledge, that correlates anticholinergic medication use with dysphonia. We found that the odds of being hoarse when taking a single anticholinergic medication were nearly twice that of a patient taking no anticholinergic drug. The odds of hoarseness jumped to nearly 5 when a patient was taking 2 or more anticholinergics, suggesting an additive effect. Whereas the laryngology literature has suggested that anticholinergic as well as diuretic use may cause dysphonia,15 we have shown for the first time their association. Although anticholinergics and hoarseness are highly correlated, we unexpectedly found no positive association between dysphonia and diuretics, suggesting that direct medication effects may play a much larger role in dysphonia than systemic dehydration.
Downloaded from aor.sagepub.com at Bobst Library, New York University on May 13, 2015
5
Haft et al Table 3. P Values Comparing the Individual VHI-10 Questions for Patients Not Taking Anticholinergic Medications Versus Those Taking 1 or More.a VHI-10 Questions My voice makes it difficult for people to hear me People have difficulty understanding me in a noisy room My voice difficulties restrict my personal and social life I feel left out of conversation because of my voice My voice problem causes me to lose income I feel as though I have to strain to produce voice The clarity of my voice is unpredictable My voice problem upsets me My voice makes me feel handicapped People ask, “What’s wrong with your voice?” Total VHI-10 score
1 Ach
≥ 2 Ach
.04*
< .01*
.01*
< .01*
.03*
< .01*
.04*
< .01*
.62
.81
.02*
< .01*
.04* .27 .03* .18
< .01* .03* .03* < .01*
.28
.02*
Abbreviation: Ach, anticholinergic. a Asterisks indicate statistical significance.
In this study, we used a multifaceted approach in assessing laryngeal complaints as they relate to medication use. Increasing scores on the hoarseness scale, VHI-10, RSI, and GRBAS all showed significant association with anticholinergic use, supporting the reliability of our assessment. By having rigid inclusion criteria that includes only dysphonia of questionable cause, we hoped to reduce confounding by a considerable amount. We found that about half of our clinic patients with hoarseness did not have a clear etiology, and of these, there was a very significant association with anticholinergic use. This suggests that a fair number of patients presenting to a laryngology practice may be suffering from symptoms secondary to specific medication use. An awareness of this important association is valuable to the clinician when attributing cause to hoarseness and when considering treatment options. Even though we attempted to reduce confounding by excluding symptomatic reflux patients, we did not exclude patients who were currently taking a PPI. Empiric long-term PPI therapy has become the standard of care for most patients presenting with chronic hoarseness without a clear organic cause. A recent study found that 56% of patients presenting to a tertiary voice clinic had already taken, or were currently taking, an extended PPI regimen for their persistent hoarseness.16 Traditionally, the resolution of hoarseness following PPI therapy has been considered diagnostic of LPR,17 but without resolution, there is still tremendous controversy surrounding the diagnosis and treatment of LPR given a significant lack of standardization and evidence-based support.17
Indeed, the majority of participants in our study have also been on extended PPI therapy, and with the persistence of voice complaints despite being on PPI therapy, it was important to include this large majority whose dysphonia may be caused by other, yet unstudied reasons for hoarseness, such as anticholinergic use. Furthermore, we included a previous diagnosis of GERD in our multivariate analysis, which was statistically unrelated to hoarseness. This initially seems surprising, however, these were only patients with successfully treated GERD (and presumably LPR) who were currently asymptomatic; patients with treatment refractory GERD were excluded from analysis. Although this was not the aim of our study, this suggests that successfully treating symptoms of GERD correlates with successful treatment of hoarseness likely due to LPR. In addition, patients with seasonal allergies (n = 20) had a 5.32 increased odds of being hoarse, although the correlation was borderline nonsignificant (P = .08). Whereas rhinitis was protective against hoarseness in our results, we believe that this is likely a statistical artifact due to a very low n of 5 participants. Of note, the association with rhinitis was not seen when using the VHI-10, RSI, and GRBAS scales. Our study also showed a borderline nonsignificant association (P = .08) between hoarseness and polypharmacy when excluding anticholinergics from the analysis. Large studies of xerostomia have likewise found positive correlations with polypharmacy, supporting a similar pathologic mechanism. Dry mouth is a common complaint in the elderly, with the main reason indeed being high frequency of polypharmacy.9 It may be possible that polypharmacy itself can exacerbate dysphonia, in addition to the effects of anticholinergic medications. There are mentionable limitations to this study. As a cohort study, we have shown an association as opposed to a cause and effect relationship. Second, we did not include additional instrumental measures in our analysis, such as stroboscopy, which only a minority of our patients underwent in the course of their routine clinical care. Two patients had esophagogastroduodenoscopy, which was not included in analysis due to a low n. We also chose not to evaluate pH probe testing results as studies have shown marked variability in accuracy with the study, with diagnostic yield reported to range widely, anywhere from 14% to 83%.17 We instead relied on validated patient surveys in conjunction with clinician performed laryngoscopy and assessment to exclude organic causes. Finally, even though we included many potential confounders within our multivariate analysis, we excluded others due to chart review constraints, such as muscle tension dysphonia, whose main symptom is hoarseness. Further study is necessary to elucidate the relationship of anticholinergic use with hoarseness beyond mere correlation. Research is needed to elucidate whether a safe decrease in the number of total medications will improve dysphonia
Downloaded from aor.sagepub.com at Bobst Library, New York University on May 13, 2015
6
Annals of Otology, Rhinology & Laryngology
in this patient population. Follow-up studies should focus on showing whether a targeted reduction of anticholinergic medications in particular will improve hoarseness of unclear etiology.
Conclusion To our knowledge, this is the first study that implicates medication use as a risk factor for hoarseness. The striking association with anticholinergic use suggests that dysphonia of unclear etiology may be due to sicca, or a drying effect. An awareness of this important association is invaluable when attributing cause to dysphonia and when considering treatment options. Authors’ Note This article was presented as a poster at the 2013 AAO-HNSF meeting in Vancouver, Canada.
Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.
References 1. Schwartz SR, Cohen SM, Dailey SH, et al. Clinical practice guideline: hoarseness (dysphonia). Otolaryngol Head Neck Surg. 2009;141:S1-S31. 2. Sulica L. Hoarseness misattributed to reflux: sources and patterns of error. Ann Otol Rhinol Laryngol. 2014;123(6):442-445. 3. Thomas JP, Zubiaur FM. Overdiagnosis of laryngopharyngeal reflux as the cause of hoarseness. Eur Arch Otorhinolaryngol. 2013;270:995-999.
4. Witt RE, Taylor LN, Regner MF, et al. Effects of surface dehydration on mucosal wave amplitude and frequency in excised canine larynges. Otolaryngol Head Neck Surg. 2011;141:108-113. 5. Dworkin J. Laryngitis: types, causes, and treatments. Otolaryngol Clin North Am. 2008;41:419-436. 6. Verdolini K, Min Y, Titze I, et al. Biological mechanisms underlying voice changes due to dehydration. J Speech Lang Hear Res. 2002;45:268-281. 7. Yoshida Y, Shimazaki T, Tanaka Y, et al. Ganglions and ganglionic neurons in the cat’s larynx. Acta Otolaryngol. 1993;113:415-420. 8. Woodson G. The Larynx. Philadelphia, PA: Lippincott Williams & Wilkins; 2003. 9. Scully C. Drug effect on salivary glands: dry mouth. Oral Dis. 2003;9:165-176. 10. Gunes Z, Denat Y, Muezzinoglu M, et al. The risk factors effecting dry mouth in inpatients in hospital in West Anatolia. J Clin Nurs. 2011;21:408-414. 11. Carnahan RM, Lund BC, Perry PJ, et al. The anticholinergic drug scale as a measure of drug related anticholinergic burden: associations with serum anticholinergic activity. J Clin Pharmacol. 2006;46:1481-1486. 12. Rudolph JL, Salow MJ, Angelini MC, et al. The anticholinergic risk scale and anticholinergic adverse effects in older persons. Arch Intern Med. 2008;168(5):508-513. 13. Rosen C, Lee A, Osborne J, et al. Development and validation of the Voice Handicap Index-10. Laryngoscope. 2004;114(9):1549-1556. 14. Belafsky PC, Postma GN, Koufman JA. Validity and reliability of the Reflux Symptom Index. J Voice. 2002;16: 274-277. 15. Ylitalo R, Merati AL, Bielamowicz SA. Textbook of Laryngology. San Diego, CA: Plural Publishing; 2007. 16. Cohen SM, Garrett CG. Hoarseness: is it really laryngopharyngeal reflux? Laryngoscope. 2008;117:363-366. 17. Gupta R, Sataloff RT. Laryngopharyngeal reflux: current concepts and questions. Curr Opin Otolaryngol Head Neck Surg. 2009;17:143-148.
Downloaded from aor.sagepub.com at Bobst Library, New York University on May 13, 2015
research-article2015
AORXXX10.1177/0003489415585867Annals of Otology, Rhinology & LaryngologyHaft et al
Original Article
Anticholinergic Use Is a Major Risk Factor for Dysphonia
Annals of Otology, Rhinology & Laryngology 1–6 © The Author(s) 2015 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/0003489415585867 aor.sagepub.com
Sunny Haft, MD1, Douglas Farquhar, BA1, Ryan Carey, BSE1, and Natasha Mirza, MD1
Abstract Objective: We hypothesize that many cases of dysphonia of unclear etiology are a form of sicca caused by anticholinergic medication use, and we aim to determine their association. Study Design: This was a cross-sectional study conducted over a 6-month time period. Participants were drawn from a tertiary care laryngology practice within an academic institution. Methods: One hundred forty-nine patients met inclusion criteria. Patients rated the symptom of chronic hoarseness; scores were compared with participants’ medication lists, comorbidities, age, and sex, and a multivariate logistic regression model was developed. Significance was set at P < .05. As a secondary analysis, participants rated a variety of other symptoms using the Voice Handicap Index-10, Reflux Symptom Index, and the GRBAS scale, which were likewise compared to anticholinergic use. Results: Any patient taking at least 1 anticholinergic medication had a 2.32 increased odds (P = .03) of experiencing hoarseness. If the patient was taking 2 or more anticholinergic medications, those odds rose to 4.52 (P = .009). Conclusion: This is the first study, to our knowledge, that implicates medication use as a major risk factor for dysphonia of unclear etiology. An awareness of this association is invaluable when attributing cause to hoarseness and when considering treatment options. Keywords anticholinergic, dysphonia, GRBAS, hoarseness, laryngitis, medications, RSI, sicca, VHI, voice
Background Hoarseness is a common complaint within an otolaryngology practice. It is defined as a coarse or rough quality to the voice, which may be due to a wide range of underlying pathology.1 The work-up for dysphonia within a laryngology practice will often suggest an underlying diagnosis, which can include neoplasm, infection, phonotrauma, smoking, laryngopharyngeal reflux (LPR), or neurologic pathology, among others. Even with a thorough investigation, however, clinicians may still struggle to find a clear cause. Many cases of dysphonia have increasingly been ascribed to LPR within the laryngology literature, but the extent of causation is still not entirely clear. In cases of dysphonia of unclear etiology, recent studies have highlighted the importance of further laryngeal investigation rather than continued empiric reflux treatment or ongoing reflux investigation.2 Others have likewise shown that LPR is overdiagnosed within the laryngology clinic.3 We hypothesize that a significant degree of hoarseness of unclear etiology may be due to medication use—namely, anticholinergic medications—and is underdiagnosed within laryngology clinics.
No one has yet studied the correlation between medication use and dysphonia, and we aimed to establish and quantify this association. Particularly, we focused on patients who had persistent hoarseness without a clear cause despite work-up within the laryngology clinic. Anticholinergic medications may be contributing to dysphonia due to a drying effect. Vocal fold dehydration has been shown to cause a decrease in mucosal wave frequency and amplitude, increasing phonation threshold pressure and, thus, strain.4 In addition, a loss of mucosal lubrication may theoretically subject the vocal folds to higher degrees of friction and heat during vibration, leading to quicker molecular breakdown.5,6 Several lines of evidence support this possible causal relationship. Parasympathetic nerve fibers in the larynx 1
Department of Otorhinolaryngology–Head and Neck Surgery, University of Pennsylvania, Philadelphia, Pennsylvania, USA Corresponding Author: Sunny Haft, MD, Department of Otorhinolaryngology–Head and Neck Surgery, University of Pennsylvania, 3400 Spruce Street, 5 Ravdin, Philadelphia, PA 19104, USA. Email: [email protected]
Downloaded from aor.sagepub.com at Bobst Library, New York University on May 13, 2015
2
Annals of Otology, Rhinology & Laryngology
outnumber sympathetic fibers by 20 to 1,7 penetrating the basal lamina and directly stimulating secretion of mucus glands. In turn, the majority of anticholinergic medications are antimuscarinic, with the ability to block the release of acetylcholine at nerve junctions that would otherwise stimulate these mucus glands. Thus, anticholinergic medications may be decreasing glandular secretions, directly leading to sicca within the larynx. Second, studies have shown the negative phonatory effects of systemic dehydration on the larynx.4,8 Heavy coffee and tea drinking has been implicated in voice disorders,4 presumably due to its diuretic effect, although this has yet to be definitively proven. The laryngologist often advises the patient to stay hydrated and avoid coffee in order to combat hoarseness. Last, laryngitis sicca may be pathophysiologically very similar to xerostomia—the larynx being a mucosal extension of the oral cavity. Xerostomia has been extensively studied, with the top cause of dry mouth indeed being medication use. Scully et al9 found that anticholinergic medications are the most common reason for reduction in the amount of saliva, more so than even dehydration. Gunes et al10 collected saliva of patients on anticholinergic medications and found that they had 3.64 times less saliva than patients not on these medications. We hypothesize that the effect of anticholinergic medications on glandular tissue in the oral cavity is causing the same degree of dryness within the larynx, ultimately leading to hoarseness. An increased awareness of this link may be an important step in guiding physicians down a diagnostic and treatment regimen for dysphonia of unclear etiology.
Methods Study Population A cross-sectional study was performed to determine the relationship between anticholinergic medication use and laryngitis symptoms in adults. The sample population chosen was a tertiary-care laryngology practice. All consecutive patients presenting to a clinic over a 6-month period in 2013 were included. Patients typically presented with voice or other laryngeal complaints after referral from a primary care practice. Patients were excluded from analysis if they had 1 or more significant risk factors for laryngitis. These included persistent reflux refractory to medical treatment (patients currently taking an H2 antagonist or proton-pump inhibitor [PPI] with a concurrent high score [4 or 5] on questions on the Reflux Symptoms Index [RSI] about “heartburn” or “coughing after lying down”), professional voice overuse (singers, performers, and teachers who complained of worsening hoarseness in the context of their profession), current smokers who smoke more than a
half pack per day for greater than 1 year, a history of prior neck radiation, Sjogren’s syndrome, amyloidosis, recent laryngeal surgery or injection within the past month, tracheostomy dependency, and structural abnormalities seen on exam (papillomatosis, nodules, granulomas, or subglottic stenosis). Patients who met inclusion criteria were then divided into those positive for dysphonia and those without dysphonia, as determined by a self-reported hoarseness score, as described below. The Institutional Committee on Research Involving Human Subjects approved the study protocol; a written consent was obtained from all participants.
Medication Sampling and Classification Patient medication lists were obtained from chart review and reaffirmed on patient interview. All medications, dosing, and compliance were confirmed by the study authors during the clinical visit. Vitamins were excluded from analysis. To help establish a cause–effect relationship, all medications that were started following the onset of hoarseness were also excluded. Each medication taken by a participant was classified as an anticholinergic, diuretic, sympathomimetic, opioid, or “other.” Anticholinergics were identified according to 2 large studies that categorized anticholinergic burden.11,12 Independent assessments of anticholinergic properties of drugs were used to categorize medications not examined by these studies.
Measurement of Dysphonia To ensure reliability, laryngitis and dysphonia were measured with 2 subjective scales: the Voice Handicap Index-10 (VHI-10) and the RSI, as well as 1 objective scale, the GRBAS. The VHI-10 is a validated 10-question survey designed to measure dysphonia, and the RSI is a 9-item scale developed to measure laryngitis related to reflux.13,14 These were completed by each participant prior to seeing the physician in order to reduce bias. The GRBAS is a clinician-based voice assessment scale that evaluates Grade, Roughness, Breathiness, Asthenia, and Strain. The GRBAS was determined by a trained laryngologist during the patient encounter. We chose to use the symptom of hoarseness, as gathered from the RSI, as our primary measure of dysphonia.1 A score of 3 to 5 was chosen to signify moderate to severe hoarseness, and a score of 0 to 2 to signify mild to none. As a secondary analysis to improve reliability, the total scores on the VHI-10, RSI, and GRBAS scales were also compared to medication use. Scores ≥ 13 on the RSI14 and ≥ 5 on the GRBAS were considered positive for laryngitis. As a post-hoc assessment, each subscore on the VHI-10 was also evaluated.
Downloaded from aor.sagepub.com at Bobst Library, New York University on May 13, 2015
3
Haft et al
Assessment of Comorbidities Comorbidities for vocal dysfunction were assessed to control for potential confounders. A chart review was conducted on each patient, noting age, sex, rhinitis, sinusitis, asthma, seasonal allergies, chronic obstructive pulmonary disease (COPD), gastroesophageal reflux disease (GERD), and recent laryngeal surgery, if applicable.
Statistical Analyses Descriptive statistics were used to analyze baseline characteristics. Statistical comparisons were performed using the Wilcox rank-sum test, Student t test, or Fisher exact test, as appropriate. To assess the effect of potential confounders, 2 multivariate logistic regression models were constructed, using the hoarseness scale and the total RSI score as the outcome, and including age, sex, comorbidities for vocal dysfunction (sinusitis, rhinitis, asthma, COPD, allergies, GERD, previous head and neck surgery), and other classes of drugs (sympathomimetics, diuretics, and opioids) as independent variables. A P value of .05 was considered significant. All statistical analysis was performed using STATA/IC 13.0 software (Stata Corporation, College Station, Texas, USA).
Results Of 243 patients enrolled in the study, 149 were eligible for inclusion. The mean age was 59 years and 44% were male; 49% of the population took at least 1 anticholinergic medication, and 27% took 2 or more. The most common medications taken by patients are listed in Table 1. Seventy-five patients (50%) were considered positive for laryngitis, scoring ≥ 3 on the hoarseness scale, and 74 were considered negative (Figure 1). There were no significant differences between the 2 cohorts in demographics or vocal comorbidities.
Anticholinergic Medication Use Is a Risk Factor for Hoarseness The use of at least 1 anticholinergic medication was significantly associated with hoarseness (35% of patients negative for hoarseness vs 57% of positive patients; P = .036). There was a stronger association with taking at least 2 anticholinergics (9% vs 27%; P = .003). Patients positive for hoarseness were also more likely to take more total medications than their non-hoarse counterparts (means of 5.5 vs 4.3; P = .03). However, when the number of anticholinergic drugs were excluded from this analysis, the association was no longer significant (3.7 vs 4.5; P = .08). No other classes of drugs had any associations with laryngitis, including diuretics (n = 43, 27% vs 31%;
Table 1. Common Anticholinergic Medications Being Taken by Study Participants. Anticholinergic Medication Diazepam Prednisone Furosemide Alprazolam Oxycodone Amitriptyline Diphenhydramine Clonazepam Paroxetine Nortriptyline Temazepam Bupropion Isosorbide Atropine Meclizine Oxybutynin Diltiazem Tramadol
No. of Participants Taking Medication 14 14 14 13 12 11 11 8 7 6 5 4 4 3 3 3 3 3
P = .72), sympathomimetics (n = 6, 7% vs 1%; P = .11), and opioids (n = 11, 8% vs 7%; P = .77). When potential confounders such as GERD, asthma, allergies, and rhinitis were accounted for in a multivariate analysis, patients with hoarseness were 2.32 times as likely to be taking at least 1 anticholinergic as patients without (P = .031) and were 4.52 times as likely to be taking 2 or more anticholinergic medications (P = .009). Overall, the odds for hoarseness increased by 1.76 with each anticholinergic medication that a patient was taking (P = .014). The results from the multivariate analysis are summarized in Table 2. No other demographic factors or comorbidities were significantly associated with hoarseness in the multivariate analysis except for rhinitis, which was actually protective against hoarseness (odds ratio = 0.09, P = .03). It is notable that the rhinitis population had an n of 5.
Anticholinergic Medication Use and VHI-10 Score The VHI-10 scores significantly differed when comparing patients taking 2 or more anticholinergics versus no anticholinergics (mean scores of 16.8 vs 11.9; P = .023). There was no difference when comparing patients on 1 anticholinergic versus none. However, when analyzing each of the 10 questions separately as a post-hoc analysis, patients taking at least 1 anticholinergic scored significantly higher on 5 of the 10 questions, and patients taking at least 2 anticholinergics scored significantly higher on 9 of the 10 questions (Table 3).
Downloaded from aor.sagepub.com at Bobst Library, New York University on May 13, 2015
4
Annals of Otology, Rhinology & Laryngology
Figure 1. Study flowchart with patient characteristics listed. Patients were categorized as having laryngitis if they scored a 3 or above on a 0-5 severity scale for hoarseness. Table 2. Multivariate Analysis Results Comparing Hoarseness Scores to Multiple Variables.a Variable
No.
Odds Ratio
1 or more anticholinergic 2 or more anticholinergics Each anticholinergic Age Male sex Sinusitis Rhinitis Asthma COPD Allergies GERD Diuretics Opioids
69 27 69 149 149 15 5 6 7 20 75 43 11
2.32 4.52 1.76 1.00 0.67 1.36 0.09 0.17 2.54 5.32 0.49 1.12 0.53
P Value .03* .01* .01* .98 .30 .76 .03* .12 .35 .08 .06 .80 .43
Abbreviations: COPD, chronic obstructive pulmonary disease; GERD, gastroesophageal reflux disease. a Asterisks indicate statistical significance.
Anticholinergic Medication Use and RSI and GRBAS Scores On the RSI, 95 patients (63%) had an RSI score ≥ 13, and 54 (36%) had a score < 13. Of the 115 patients assessed with the GRBAS scale, 79 (69%) had a score ≥ 5 and 36 (31%) had a score < 5. The use of at least 1 anticholinergic medication was also significantly associated with a high RSI score (P = .001) and a high GRBAS score (P = .005). Likewise, the use of at least 2 anticholinergic medications was significantly
associated with a high RSI score (P < .001), although the association with a high GRBAS score was borderline significant (P = .06). Smoking was associated with a high GRBAS score (P = .01), but no other demographic factors, vocal comorbidities, or other classes of drugs were significantly associated with high scores on these measures. When using the RSI as the dependent variable in the multivariate model, every additional anticholinergic medication was associated with a 2.32 increased odds of the patient having a high score (P = .0001). A patient taking 1 or more anticholinergics had a 2.98 increased odds of having a high score (P = .006). If a patient was taking 2 or more anticholinergics, he or she had an 8.62 increased odds of having a high score (P = .0001).
Discussion This is the first study, to our knowledge, that correlates anticholinergic medication use with dysphonia. We found that the odds of being hoarse when taking a single anticholinergic medication were nearly twice that of a patient taking no anticholinergic drug. The odds of hoarseness jumped to nearly 5 when a patient was taking 2 or more anticholinergics, suggesting an additive effect. Whereas the laryngology literature has suggested that anticholinergic as well as diuretic use may cause dysphonia,15 we have shown for the first time their association. Although anticholinergics and hoarseness are highly correlated, we unexpectedly found no positive association between dysphonia and diuretics, suggesting that direct medication effects may play a much larger role in dysphonia than systemic dehydration.
Downloaded from aor.sagepub.com at Bobst Library, New York University on May 13, 2015
5
Haft et al Table 3. P Values Comparing the Individual VHI-10 Questions for Patients Not Taking Anticholinergic Medications Versus Those Taking 1 or More.a VHI-10 Questions My voice makes it difficult for people to hear me People have difficulty understanding me in a noisy room My voice difficulties restrict my personal and social life I feel left out of conversation because of my voice My voice problem causes me to lose income I feel as though I have to strain to produce voice The clarity of my voice is unpredictable My voice problem upsets me My voice makes me feel handicapped People ask, “What’s wrong with your voice?” Total VHI-10 score
1 Ach
≥ 2 Ach
.04*
< .01*
.01*
< .01*
.03*
< .01*
.04*
< .01*
.62
.81
.02*
< .01*
.04* .27 .03* .18
< .01* .03* .03* < .01*
.28
.02*
Abbreviation: Ach, anticholinergic. a Asterisks indicate statistical significance.
In this study, we used a multifaceted approach in assessing laryngeal complaints as they relate to medication use. Increasing scores on the hoarseness scale, VHI-10, RSI, and GRBAS all showed significant association with anticholinergic use, supporting the reliability of our assessment. By having rigid inclusion criteria that includes only dysphonia of questionable cause, we hoped to reduce confounding by a considerable amount. We found that about half of our clinic patients with hoarseness did not have a clear etiology, and of these, there was a very significant association with anticholinergic use. This suggests that a fair number of patients presenting to a laryngology practice may be suffering from symptoms secondary to specific medication use. An awareness of this important association is valuable to the clinician when attributing cause to hoarseness and when considering treatment options. Even though we attempted to reduce confounding by excluding symptomatic reflux patients, we did not exclude patients who were currently taking a PPI. Empiric long-term PPI therapy has become the standard of care for most patients presenting with chronic hoarseness without a clear organic cause. A recent study found that 56% of patients presenting to a tertiary voice clinic had already taken, or were currently taking, an extended PPI regimen for their persistent hoarseness.16 Traditionally, the resolution of hoarseness following PPI therapy has been considered diagnostic of LPR,17 but without resolution, there is still tremendous controversy surrounding the diagnosis and treatment of LPR given a significant lack of standardization and evidence-based support.17
Indeed, the majority of participants in our study have also been on extended PPI therapy, and with the persistence of voice complaints despite being on PPI therapy, it was important to include this large majority whose dysphonia may be caused by other, yet unstudied reasons for hoarseness, such as anticholinergic use. Furthermore, we included a previous diagnosis of GERD in our multivariate analysis, which was statistically unrelated to hoarseness. This initially seems surprising, however, these were only patients with successfully treated GERD (and presumably LPR) who were currently asymptomatic; patients with treatment refractory GERD were excluded from analysis. Although this was not the aim of our study, this suggests that successfully treating symptoms of GERD correlates with successful treatment of hoarseness likely due to LPR. In addition, patients with seasonal allergies (n = 20) had a 5.32 increased odds of being hoarse, although the correlation was borderline nonsignificant (P = .08). Whereas rhinitis was protective against hoarseness in our results, we believe that this is likely a statistical artifact due to a very low n of 5 participants. Of note, the association with rhinitis was not seen when using the VHI-10, RSI, and GRBAS scales. Our study also showed a borderline nonsignificant association (P = .08) between hoarseness and polypharmacy when excluding anticholinergics from the analysis. Large studies of xerostomia have likewise found positive correlations with polypharmacy, supporting a similar pathologic mechanism. Dry mouth is a common complaint in the elderly, with the main reason indeed being high frequency of polypharmacy.9 It may be possible that polypharmacy itself can exacerbate dysphonia, in addition to the effects of anticholinergic medications. There are mentionable limitations to this study. As a cohort study, we have shown an association as opposed to a cause and effect relationship. Second, we did not include additional instrumental measures in our analysis, such as stroboscopy, which only a minority of our patients underwent in the course of their routine clinical care. Two patients had esophagogastroduodenoscopy, which was not included in analysis due to a low n. We also chose not to evaluate pH probe testing results as studies have shown marked variability in accuracy with the study, with diagnostic yield reported to range widely, anywhere from 14% to 83%.17 We instead relied on validated patient surveys in conjunction with clinician performed laryngoscopy and assessment to exclude organic causes. Finally, even though we included many potential confounders within our multivariate analysis, we excluded others due to chart review constraints, such as muscle tension dysphonia, whose main symptom is hoarseness. Further study is necessary to elucidate the relationship of anticholinergic use with hoarseness beyond mere correlation. Research is needed to elucidate whether a safe decrease in the number of total medications will improve dysphonia
Downloaded from aor.sagepub.com at Bobst Library, New York University on May 13, 2015
6
Annals of Otology, Rhinology & Laryngology
in this patient population. Follow-up studies should focus on showing whether a targeted reduction of anticholinergic medications in particular will improve hoarseness of unclear etiology.
Conclusion To our knowledge, this is the first study that implicates medication use as a risk factor for hoarseness. The striking association with anticholinergic use suggests that dysphonia of unclear etiology may be due to sicca, or a drying effect. An awareness of this important association is invaluable when attributing cause to dysphonia and when considering treatment options. Authors’ Note This article was presented as a poster at the 2013 AAO-HNSF meeting in Vancouver, Canada.
Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.
References 1. Schwartz SR, Cohen SM, Dailey SH, et al. Clinical practice guideline: hoarseness (dysphonia). Otolaryngol Head Neck Surg. 2009;141:S1-S31. 2. Sulica L. Hoarseness misattributed to reflux: sources and patterns of error. Ann Otol Rhinol Laryngol. 2014;123(6):442-445. 3. Thomas JP, Zubiaur FM. Overdiagnosis of laryngopharyngeal reflux as the cause of hoarseness. Eur Arch Otorhinolaryngol. 2013;270:995-999.
4. Witt RE, Taylor LN, Regner MF, et al. Effects of surface dehydration on mucosal wave amplitude and frequency in excised canine larynges. Otolaryngol Head Neck Surg. 2011;141:108-113. 5. Dworkin J. Laryngitis: types, causes, and treatments. Otolaryngol Clin North Am. 2008;41:419-436. 6. Verdolini K, Min Y, Titze I, et al. Biological mechanisms underlying voice changes due to dehydration. J Speech Lang Hear Res. 2002;45:268-281. 7. Yoshida Y, Shimazaki T, Tanaka Y, et al. Ganglions and ganglionic neurons in the cat’s larynx. Acta Otolaryngol. 1993;113:415-420. 8. Woodson G. The Larynx. Philadelphia, PA: Lippincott Williams & Wilkins; 2003. 9. Scully C. Drug effect on salivary glands: dry mouth. Oral Dis. 2003;9:165-176. 10. Gunes Z, Denat Y, Muezzinoglu M, et al. The risk factors effecting dry mouth in inpatients in hospital in West Anatolia. J Clin Nurs. 2011;21:408-414. 11. Carnahan RM, Lund BC, Perry PJ, et al. The anticholinergic drug scale as a measure of drug related anticholinergic burden: associations with serum anticholinergic activity. J Clin Pharmacol. 2006;46:1481-1486. 12. Rudolph JL, Salow MJ, Angelini MC, et al. The anticholinergic risk scale and anticholinergic adverse effects in older persons. Arch Intern Med. 2008;168(5):508-513. 13. Rosen C, Lee A, Osborne J, et al. Development and validation of the Voice Handicap Index-10. Laryngoscope. 2004;114(9):1549-1556. 14. Belafsky PC, Postma GN, Koufman JA. Validity and reliability of the Reflux Symptom Index. J Voice. 2002;16: 274-277. 15. Ylitalo R, Merati AL, Bielamowicz SA. Textbook of Laryngology. San Diego, CA: Plural Publishing; 2007. 16. Cohen SM, Garrett CG. Hoarseness: is it really laryngopharyngeal reflux? Laryngoscope. 2008;117:363-366. 17. Gupta R, Sataloff RT. Laryngopharyngeal reflux: current concepts and questions. Curr Opin Otolaryngol Head Neck Surg. 2009;17:143-148.
Downloaded from aor.sagepub.com at Bobst Library, New York University on May 13, 2015
