Descriptive Epidemiology of Voice Disorders in Rheumatoid Arthritis: Prevalence, Risk Factors, and Quality of Life Burden *Nelson Roy, †Kristine Tanner, ‡Ray M. Merrill, *Charisse Wright, §Karla L. Miller, and kKatherine A. Kendall, *xkSalt Lake City, yzProvo, Utah
Summary: Objectives. Rheumatoid arthritis (RA) is an autoimmune inflammatory disease which may adversely affect phonatory function. This study aimed to establish the prevalence, risks, and quality of life effects of voice disorders in RA. Study Design. This is a cross-sectional, descriptive epidemiology study. Methods. One hundred individuals with RA underwent a telephone interview to determine the frequency, severity, risks associated with, and quality of life burden of voice disorders. The results were analyzed using summary statistics, frequencies, chi-square tests, regression analysis, and risk ratios (P < 0.05). Results. Thirty-five percent of participants with RA reported a current voice disorder which was chronic and longstanding in most cases. The prevalence of a current voice disorder did not significantly differ across age, sex, medication use, voice use patterns, medical history, or RA severity. These chronic voice disorders produced significant adverse effects on both voice-related quality of life and short form 36 health-related quality of life scales. Specific voice symptoms such as ‘‘voice-related discomfort’’ and ‘‘chronic throat dryness’’ contributed disproportionately to the quality of life burden. Of those participants with a voice disorder, only 37% had ever sought professional help to improve their voice. Conclusions. These results indicate that voice disorders are common in RA and produce significant adverse effects on quality of life. Further research is necessary to better understand the origin of these disorders and their potential response to treatment. Key Words: Voice disorders–Epidemiology–Rheumatoid arthritis–Quality of life.
INTRODUCTION Autoimmune diseases are associated with the body’s inappropriate defense against its own healthy tissue. Rheumatoid arthritis (RA) is a chronic, disabling, autoimmune condition of unknown cause, characterized by inflammation of synovial joint tissue. Joints commonly affected in RA include the hands, feet, ankles, and wrists, but any joint may be involved, including within the larynx.1,2 RA affects approximately 2– 3% of the adult population,3,4 and women are more than twice as likely as men to develop the disease.1 Although the average age of onset is 55 years, the prevalence of RA increases with age, affecting 6% of white adults over age 65.3 RA within the larynx may produce voice problems; however, the true prevalence of such disorders and their burden on quality of life remains undetermined.4–9 Laryngeal signs of RA reportedly include ‘‘bamboo’’ nodules and cricoarytenoid joint fixation with dysphonia and stridor. Bamboo nodules are bilateral, yellow, midmembranous vocal fold swellings that have a calloused appearance and may occur in as many as 25% of individuals with RA.4,10–13 Stridor occurs presumably because Accepted for publication February 20, 2015. From the *Department of Communication Sciences and Disorders, The University of Utah, Salt Lake City, Utah; yDepartment of Communication Disorders, Brigham Young University, Provo, Utah; zDepartment of Health Science, Brigham Young University, Provo, Utah; xDivision of Rheumatology, Department of Internal Medicine, The University of Utah, Salt Lake City, Utah; and the kDivision of Otolaryngology—Head and Neck Surgery, Department of Surgery, The University of Utah, Salt Lake City, Utah. Address correspondence and reprint requests to Nelson Roy, Department of Communication Sciences and Disorders, The University of Utah, 390 South 1530 East, Room 1310, Salt Lake City, UT 84112-0252. E-mail: [email protected] Journal of Voice, Vol. -, No. -, pp. 1-14 0892-1997/$36.00 Ó 2015 The Voice Foundation http://dx.doi.org/10.1016/j.jvoice.2015.02.011
of fixation of the vocal folds via the cricoarytenoid joint, with upper airway obstruction reportedly occurring in 16% of patients.2,14–16 In addition to these laryngeal changes, many patients with RA report feeling a foreign body sensation in the throat, dysphagia, pain with speaking, and vocal fatigue.2,14,15,17 Table 1 summarizes the extant literature examining laryngeal/ voice involvement in RA. Inspection of Table 1 reveals considerable variability regarding the prevalence of laryngeal signs and phonatory symptoms in RA. Some of these discrepancies are related to inconsistent criteria of what precisely constitutes a ‘‘voice disorder,’’ methodological differences related to sampling procedures, and sample populations and sizes, among other issues. For instance, some studies relied exclusively on visual inspection of the larynx and reported a very high prevalence of laryngeal changes (ie, 80%) in patients with RA,22,23 whereas Lofgren and Montgomery18 reported laryngeal involvement in only 26% of participants on the basis of patient histories and extralaryngeal palpation, in addition to indirect laryngoscopy. Lawry et al24 compared indirect laryngoscopy versus computed tomography and documented laryngeal manifestations of RA in 32% and 54% of participants, respectively. More recently, investigators have used the voice handicap index (VHI)25 and other self-report methodologies to establish the presence of a voice disorder and reported prevalence rates in RA ranging from 5% to 48%.6,7,9,19,20 Although RA has the potential to adversely affect phonatory function, it is clear from the previously described review that significant disagreement exists regarding (1) the prevalence of voice disorders in RA, (2) patterns and symptoms of voice disorders in RA, (3) risk factors associated with these disorders, and (4) the consequences of voice disorders and their associated symptoms on social, physical, and emotional
2
TABLE 1. Studies Examining Voice Dysfunction in RA Study
# of Subjects
Gender Represented
Age (y)
‘‘Incidence of laryngeal involvement in rheumatoid arthritis’’ Lofgren and Montgomery18 ‘‘Rheumatoid nodules of the larynx’’ Woo et al4
100 participants with RA
68 women, 32 men
N/A
64-year-old woman
One woman with RA
64 y
‘‘Airway obstruction and rheumatoid arthritis’’ Vergnenegre et al15
100 participants with RA, 88 participants with other rheumatological diseases 1 woman with RA
77 women, 23 men in RA group and 67 women, 21 men in control group
External palpation and indirect laryngoscopy.
Findings 26% of RA participants examined were found to have involvement of cricoarytenoid joint.
Rheumatoid nodules can Flexible laryngoscopic affect the vocal folds, examination; which causes hoarseness laryngostroboscopy to or dysphonia; RA in the assess vibratory function; larynx is found in as microdirect laryngoscopy; many as 25% of RA and evacuation of participants. subcordal masses from each vocal fold. Number of obstructive RA group ¼ 60 ± 12 y Questionnaire, syndromes was higher in control group ¼ 57 ± 21 y spirographic participants with RA than measurements taken, other rheumatologic chest radiograph. conditions.
1 woman
57-year-old with 10-year history of RA
N/A
N/A
‘‘Prevalence and relative risk of dysphonia in rheumatoid arthritis’’ Speyer et al9
47 men, 119 women
19–89 y, Mean age ¼ 61
Arytenoid adduction surgery performed to normalize voice.
In cases of bilateral cricoarytenoid joint involvement in RA, airway obstruction can occur. VHI-10 and reflux symptom Findings suggest a prevalence of voice index to assess impairment of 5% based symptomatology of on VHI-10 score >15. laryngopharyngeal reflux; Participants with RA had a joint assessment, erythrocyte sedimentation rate (ESR), and visual analog to calculate disease activity score (DAS). Prevalence of voice Visual analog scale of disorders 15% in RA severity by participants, compared rheumatologist and with 6% in controls based patient; VHI and threeon VHI >15. item outcome scale on RA 2.9–3.7 times more likely perception of to have dysphonia. impairment.
Journal of Voice, Vol. -, No. -, 2015
‘‘Arytenoid adduction to treat impaired adduction of the vocal fold due to rheumatoid arthritis’’ Kumai et al16 ‘‘Prevalence of subjective 73 with RA voice impairment in 73 controls rheumatoid arthritis’’ Fisher et al6
166 RA 148 control
Method
109 participants with autoimmune disease (60% with RA) 41 controls
74 women and Average age similar 35 men in for autoimmune and autoimmune group control groups
80 S’s with autoimmune conditions 60 controls (vocally normal, non rheumatic)
44 subjects with RA, RA group ¼ 58.2 y control group ¼ 48.3 y 32 with systemic lupus, 4 with Sjogren syndrome.
47 participants with RA, 40 controls
40 women, 7 men in RA group 26 women, 14 men in control group
‘‘Laryngeal assessment by videolaryngealstroboscopy in patients with rheumatoid arthritis’’ Gomez-Puerta et al21
36 participants with RA
33 females, 3 males
RA group ¼ 56.5 ± 12.6 y Evaluated participants with Prevalence of dysphonia was 12.8% and 47.9 ± 13.6 in control RA according to disease videolaryngoscopy group activity score, laryngeal changes 72.4% in symptoms, VHI and participants with RA; videolaryngoscopy Posterior laryngitis was compared with a control the most common group. diagnosis. Organic involvement was 56; 3 ± 14; 2 y Reflux symptom index uncommon, one patient (RSI), voice handicap had bamboo nodules, no index-10 (VHI-10) to cricoarytenoid evaluate functional voice impairment was found, impairment, but pharyngeal-laryngeal videolaryngostroboreflux was found in 64% scopy to determine of participants. laryngeal involvement.
3
Abbreviation: N/A, not applicable.
Prevalence of dysphonia VHI, three-item VAS, was 38.6% in RA group biochemical and Cbased on VHI >15. No reative protein study in 40 clear association disease-active between disease activity participants. and VHI total score.
Epidemiology of Voice Disorders in Rheumatoid Arthritis
‘‘Voice symptoms in patients with autoimmune disease: a Cross-sectional epidemiological study’’ Liu et al7 ‘‘The prevalence of dysphonia, its association with immunomediated diseases and correlation with biochemical markers’’ Sanz et al19 ‘‘Laryngeal involvement in rheumatoid arthritis’’ Beirith et al20
Endoscopic laryngoscopy Bamboo joint-like nodes were identified in three and laryngostroboscopy female participants. and phonetogram documentation for logopedic voice investigation. Laryngoscopic evaluation, Almost 2/3s of RA participants had loss of acoustic analysis, range, and 66% had computerized hoarseness compared tomography; hoarseness, with 25% in control loss of range, vocal group; 9.1% had decrease fatigue were vocal in vocal fold mobility and symptoms inquired. 27% had moderate/ severe edema of vocal folds/arytenoids compared with none in control group. VHI-10, xerostomia scale, 16.9% of RA group reported acid reflux inquiry, and significant voice anxiety/depression scale. symptoms based on VHI10 > 7.
Nelson Roy, et al
29, 31, and 50 y ‘‘Hoarseness due to 3 female participants 3 women bamboo nodes in with bamboo patients with nodes autoimmune diseases: a review of literature’’ Hilgert et al11 ‘‘Cricoarytenoid joint 2 males 22–65; mean age 48.5 y 11 participants abnormalities in patients 9 females in RA group with RA with rheumatoid 8 controls arthritis’’ Berjawi et al5
4 functioning. This cross-sectional, descriptive epidemiological investigation aimed to improve understanding of the prevalence, risk factors, and quality of life burden of voice disorders in patients with RA. METHODS Participants To identify potential participants with RA, a chart review was undertaken of patients in the Division of Rheumatology at The University of Utah Hospital over a 3-year period. The medical records of all eligible participants were subsequently reviewed in detail by C.W. to verify diagnosis. Conventional criteria for the diagnosis of RA were used including the presence of rheumatoid factor (RF), anti-cyclic citrullinated peptide igG antibodies, and one or more of the following symptoms: morning joint stiffness, arthritis of three or more joint areas, arthritis of hand joints, symmetric arthritis, rheumatoid nodules, and/or radiographic changes.1,26,27 RF, which is an autoimmune antibody, is considered a marker for RA, with approximately 80% of individuals with RA having elevated RF levels.28 In addition to RF, another marker of RA is the presence of antibodies directed to citrullinated peptides or proteins, as measured by an anti-cyclic citrullinated peptide test. Consecutive patients with confirmed RA diagnosis (using the previously described criteria) were identified and invited to participate until 100 participants were enrolled. Inclusion criteria were diagnosis of RA, age 18 years, Englishspeaking, adequate hearing for purposes of telephone interview, and no known or reported cognitive deficits. The medical record review identified 235 eligible participants with RA who were subsequently invited to participate. Of these, 110 participants did not return the researcher’s phone calls and 25 participants declined; therefore, the final group of participants included 100 patients with RA or 42% of eligible participants. Once written consent was obtained, a telephone interview using a structured questionnaire involving primarily forced-choice responses was completed. The questionnaire required approximately 75 minutes to complete. Interviews were conducted over a 9-month period from 2012 to 2013 by trained students in the Department of Communication Sciences and Disorders at The University of Utah (IRB#00058438). The final pool of participants ranged in age from 29.3 to 95.1 years, with mean age 61.1 years (standard deviation [SD], 13.1). A description of the 100 participants is provided in Table 2. Most participants were women, were nonHispanic whites, had a current household yearly gross income of $60 000, and had at least some college education. RA disease duration ranged from 1 to 50 years, with a mean of 19.6 (SD, 13.4). The number of participants currently taking medication for RA was 95 (95%), over time ranging in years from 1 to 47 (Mean [M], 15.6; SD, 12.4). Years with RA and whether medication was currently being used for RA were not significantly associated with age, sex, race/ethnicity, income, or education. Among the 100 participants with RA, only eight had another coexisting autoimmune condition in addition to RA; four had Sj€ ogren syndrome, one of whom also had mixed con-
Journal of Voice, Vol. -, No. -, 2015
TABLE 2. Description of Participants Demographic Variable Age (y) 20–39 40–59 60–69 70–79 >80 Sex Male Female Ethnicity/race NH—white NH—Asian NH—Native American/ Alaska Native Hispanic Household yearly gross income < $20 000 $20 000–$39 999 $40 000–$59 999 >$60 000 Missing Education 80 years; P < 0.0344). RA severity also decreased with increased education level (M, 100 for < high school [HS]; M, 87.8 for HS; M, 72.2 for some college; M, 67.1 for college degree; and M, 50.9 for some graduate/graduate degree). Questionnaire description The interviewer used a comprehensive, 45-item questionnaire which probed multiple domains including six primary content areas—voice disorder prevalence, history, and symptoms/signs; medical history; social/lifestyle factors; psychosocial factors; occupational factors; and RA disease severity. For the purpose of this study, we considered a voice disorder to be ‘‘any time the voice does not work, perform, or sound as it normally should, so that it interferes with communication.’’30 This case definition of a voice disorder was previously used in the largest National Institutes of Health epidemiology study of voice disorders in the United States, and therefore, permits comparisons with the general population and other high-risk groups.30 The six primary domains assessed by the questionnaire are detailed in the following. Voice disorder history. The questionnaire examined details of the participant’s frequency, onset, and duration of current and/or previous voice concerns with the intention of clearly defining correlates of voice concerns, patterns of change across the life span, typical disease-related features, rates of treatment seeking, and outcomes. The presence of voice disorders and related symptomatology were assessed in relation to the other domains, which are detailed in the following sections. Medical history. Voice production can be influenced by a variety of medical conditions. We selected 21 conditions or disorders with the potential to influence the structure and/or function of voice and considered the presence or absence of these conditions in the participant’s lifetime and, if appropriate, mechanisms of medical treatment. The inclusion of the 21 disorders was based on previously used and validated questionnaires, a review of the present literature, and the clinical experience of the investigators.30–32 This section also included more detailed questions regarding upper respiratory concerns and the use of medications. Psychosocial history and quality of life. Psychosocial factors may contribute to, and be affected by, voice disorders.33 This section of the questionnaire probed two main areas within the psychosocial domain, including psychosocial risk factors, such as stress/anxiety and activity level and the psychosocial impact of voice disorders. In addition, two standardized batteries, the voice-related quality of life (V-RQOL)34 and the short form 36 health survey (SF-36),35 were included to examine the quality of life impact of voice disorders in RA and the relation of these disorders to general health and well-being. The VRQOL is a validated 10-item instrument that has been used to determine the impact of voice impairment on quality of life and as an outcome measure in participants with voice disorders. The V-RQOL is considered a disease-specific quality of life instrument. In contrast, the SF-36 is a psychometrically validated questionnaire developed to assess eight different measures of functional health and well-being. The SF-36 is a generic tool in that it does not target a specific age, disease, or treatment
5
group (ie, it is not disease-specific). However, it can establish the relative health burden of eight profiles (physical functioning, bodily pain, role limitations due to physical health problems, role limitations due to emotional problems, emotional wellbeing, social functioning, energy/fatigue, and general health perceptions). In addition, two component summary measures of physical and mental health are generated with four scales reflecting different dimensions of physical health and four scales reflecting different dimensions of mental health. The raw scores from the questionnaire are converted into a 0–100 scale, with 100 representing the highest state of health. General normative or ‘‘average’’ scores for each of the scale items across all participants have been reported previously, ranging from the 70s for general health, bodily pain, and energy/fatigue, to the 80s for social functioning, physical functioning, role limitations— physical health, and emotional well-being, to the 90s for role limitations—emotional stresses. To help the reader interpret the results, the highest possible score for physical functioning is obtained by performing all types of physical activity including the most vigorous without limitations due to health. For rolephysical, the highest score is obtained by having no problems with work or other daily activities. For bodily pain, the highest score is obtained by having no pain or limitations due to pain. For general health, the highest score is obtained by evaluating personal health as excellent. Likewise, the highest scores for energy/fatigue are obtained by feeling full of pep and energy all the time; for social functioning by performing normal social activities without interference due to physical or emotional problems; for role-emotional by having no problems with work or other daily activities; and for mental health by feeling peaceful, happy, and calm all the time. A brief description of each of the eight scales is provided in Appendix 1. Occupation/employment history. Research has identified a link between occupation and the risk of voice disorders.30–32 The questionnaire examined aspects of the participant’s occupational history to determine its potential role in voice disorder development and symptomatology. This section examined work history, job-related voice use patterns, and the participant’s perception of the impact of job-related voice use on voice production. Social and lifestyle history. Social and lifestyle patterns and choices may influence voice.30,33 Diet may influence these processes via its impact on systemic hydration and on levels of stomach acid refluxing into the larynx and pharynx. The use of tobacco products, alcohol, and recreational drugs may also influence function by altering the integrity of tissues in the aerodigestive tract. Thus, this section probed the participant’s dietary history, such as caffeine intake, consumption of dairy products, or the intake of spicy foods, as well as their history of using tobacco products, alcohol, and recreational drugs. Disease severity. Numerous items in the questionnaire rated the severity of RA, specific symptoms, and overall general health. Along with these items included in the main body of the questionnaire, several appendices were administered to assess the severity of comorbid autoimmune diagnoses, if present. As discussed previously, the severity of RA was assessed
6 using the RAPS, a psychometrically validated instrument developed to measure pain in adult patients (>18 years) with RA.29 Pain is a major complaint in individuals with RA and often the greatest limiting factor related to function and mobility. The 24 items on the scale are scored using a sevenpoint Likert scale ranging from ‘‘0-never’’ to ‘‘6-always’’ with a higher score representing a greater severity of pain. These scores were compared with voice symptoms and disorders (among other variables) to determine possible relationships.
Statistical analyses Conventional data analyses procedures used in descriptive epidemiology research were used.33,36 Medical, familial, environmental, and social history data were examined using contingency tables, summary statistics, chi-square tests, and risk ratios (RRs). In the following section, data analysis procedures are described in detail. The prevalence and severity of voice problems were evaluated using descriptive statistics. Variables were examined using contingency tables and multiple logistic regression. Bivariate analyses of association were evaluated for statistical significance using the chi-square (c2), the Mantel-Haenszel chisquare (MH c2), and the Cochran-Mantel-Haenszel statistic (CMH). The c2 was used to test for independence between variables, the MH c2 was used to test for differences in trends between variables, and the CMH was used to test for independence after adjusting for potential confounding factors. Tests that beta coefficients in the logistic regression models equal zero were based on the Wald test, which follows an approximate c2 distribution. Cross-tabulations were used to perform bivariate analyses between selected variables, with statistical significance on the basis of the chi-square test for independence (c2). The t statistic was used for testing the null hypothesis of equality of means between groups (ie, those with voice disorders vs those without), with the t statistic computed using approximate degrees of freedom from Satterthwaite approximation when the variances between the two groups were unequal. In addition, unadjusted and adjusted (controlling for the presence of other variables) RRs were estimated using logistic regression to establish specific risk factors for reporting voice disorders. The association between the history of a voice disorder and the presence of selected risk factors (eg, employment classification, workrelated voice use issues, history of esophageal reflux disease, frequent colds/sinus infections, chemical exposures, and many other disease entities) was assessed for homogeneity across the levels of age and sex using the Breslow-Day test for homogeneity. Multiple logistic regression modeling was used, with interaction terms among significant main effects evaluated in the model. Finally, whether the relationship between the severity of voice disorders and selected dichotomous risk factors depended on the duration, frequency, sex, or age was addressed by assessing interaction terms among the significant main effects in multiple regression models. Analyses were performed using Statistical Analysis Software (SAS) Version 9.1, SAS Institute, Inc., Cary, North Carolina37
Journal of Voice, Vol. -, No. -, 2015
RESULTS Prevalence of voice disorders Thirty-five participants with RA (ie, 35%) reported having a current voice disorder, and 20 participants (20%) reported having a voice disorder in the past. Of the participants who currently had a voice disorder, 29 (82.9%) reported that it began gradually and only six (17.1%) said it began suddenly; 30 (85.7%) reported it as a chronic problem (ie, duration of 4 weeks or longer). For those with a voice disorder, 20% first noticed the disorder within the past year, 30% 1–3 years earlier, 26.7% 4– 9 years earlier, and 23.3% 10 years earlier. The prevalence of a current voice disorder did not significantly differ across the levels of age, sex, race/ethnicity, income, or education. Voice symptoms The presence and frequency of selected voice-related symptoms identified in the sample population are presented in Table 4. Participants were asked to quantify the frequency of selected voice symptoms with ‘‘daily’’ ¼ 1, ‘‘weekly’’ ¼ 2, ‘‘monthly’’ ¼ 3, ‘‘several times a year’’ ¼ 4, and ‘‘yearly or less’’ ¼ 5. The mean frequency of the selected voice symptoms ranged from 1.7 to 3.8 (M, 3.0 or monthly; SD, 0.6). Almost one-third of participants reported current hoarseness (31%) and chronic throat dryness (36%), and almost one-half of participants (48%) complained of frequent throat-clearing behavior that occurred on a daily basis. Ten of the 15 voice-related symptoms were significantly associated with the participants currently reporting a voice disorder including hoarseness, voice tiring after using it for a short time, trouble speaking or singing softly, difficulty projecting the voice, discomfort while using the voice, increased effort to talk, chronic dryness in the throat, wet/gurgly voice quality, frequent throat clearing, and an airy or breathy voice. General health, lifestyle, voice use patterns, and voice disorders To assess the relationship between general health/medical conditions, lifestyle and personal habits, and voice use patterns and experiencing a current voice disorder, participants were asked numerous questions related to each of these areas. RRs for all variables were determined. Table 5 lists many (but not all) of the areas addressed, including the number of participants with each condition who also reported a current voice disorder and the related RRs. Inspection of Table 5 reveals that despite an exhaustive review of medical conditions, lifestyle, and vocal habits, only a small number of factors emerged as being associated with a current voice disorder. The factors associated with significant RRs included current neck/throat (RR, 2.4) and jaw tension (RR, 2.3), past history of vocally demanding occupation (RR, 2.3) and related voice problems (RR, 2.4), past history of solo singing (RR, 1.90), current cough (RR, 2.23) and throat clearing (RR, 1.78), frequent ingestion of spicy foods (RR, 1.87) and colas (RR, 1.71), current sleep disorder (RR, 1.8), and being quiet (RR, 1.7) and physically active (RR, 0.5). It is also worth noting that numerous medical conditions and other potential risk factors not listed in Table 5 failed to be significantly associated with reporting a current voice
Nelson Roy, et al
7
Epidemiology of Voice Disorders in Rheumatoid Arthritis
TABLE 4. Levels of Selected Voice-Related Symptoms Voice Disordery in Those With a Current Symptom
c2
Current
Frequency*
Symptom
%
M (SD)
n
%
P Value
Hoarseness Voice tires or changes quality after using it for a short time Trouble speaking or singing softly Difficulty projecting your voice Loss of singing range Discomfort while using your voice A monotone voice (monopitch) Takes effort to talk Chronic dryness in your throat Wet, gurgly voice quality Chronic soreness in your throat Frequently clear your throat A bitter or acid taste Wobbly or shaky voice An ‘‘airy’’ or ‘‘breathy’’ voice
31 19
3.8 (1.5) 2.7 (1.4)
20 16
65 84
Summary: Objectives. Rheumatoid arthritis (RA) is an autoimmune inflammatory disease which may adversely affect phonatory function. This study aimed to establish the prevalence, risks, and quality of life effects of voice disorders in RA. Study Design. This is a cross-sectional, descriptive epidemiology study. Methods. One hundred individuals with RA underwent a telephone interview to determine the frequency, severity, risks associated with, and quality of life burden of voice disorders. The results were analyzed using summary statistics, frequencies, chi-square tests, regression analysis, and risk ratios (P < 0.05). Results. Thirty-five percent of participants with RA reported a current voice disorder which was chronic and longstanding in most cases. The prevalence of a current voice disorder did not significantly differ across age, sex, medication use, voice use patterns, medical history, or RA severity. These chronic voice disorders produced significant adverse effects on both voice-related quality of life and short form 36 health-related quality of life scales. Specific voice symptoms such as ‘‘voice-related discomfort’’ and ‘‘chronic throat dryness’’ contributed disproportionately to the quality of life burden. Of those participants with a voice disorder, only 37% had ever sought professional help to improve their voice. Conclusions. These results indicate that voice disorders are common in RA and produce significant adverse effects on quality of life. Further research is necessary to better understand the origin of these disorders and their potential response to treatment. Key Words: Voice disorders–Epidemiology–Rheumatoid arthritis–Quality of life.
INTRODUCTION Autoimmune diseases are associated with the body’s inappropriate defense against its own healthy tissue. Rheumatoid arthritis (RA) is a chronic, disabling, autoimmune condition of unknown cause, characterized by inflammation of synovial joint tissue. Joints commonly affected in RA include the hands, feet, ankles, and wrists, but any joint may be involved, including within the larynx.1,2 RA affects approximately 2– 3% of the adult population,3,4 and women are more than twice as likely as men to develop the disease.1 Although the average age of onset is 55 years, the prevalence of RA increases with age, affecting 6% of white adults over age 65.3 RA within the larynx may produce voice problems; however, the true prevalence of such disorders and their burden on quality of life remains undetermined.4–9 Laryngeal signs of RA reportedly include ‘‘bamboo’’ nodules and cricoarytenoid joint fixation with dysphonia and stridor. Bamboo nodules are bilateral, yellow, midmembranous vocal fold swellings that have a calloused appearance and may occur in as many as 25% of individuals with RA.4,10–13 Stridor occurs presumably because Accepted for publication February 20, 2015. From the *Department of Communication Sciences and Disorders, The University of Utah, Salt Lake City, Utah; yDepartment of Communication Disorders, Brigham Young University, Provo, Utah; zDepartment of Health Science, Brigham Young University, Provo, Utah; xDivision of Rheumatology, Department of Internal Medicine, The University of Utah, Salt Lake City, Utah; and the kDivision of Otolaryngology—Head and Neck Surgery, Department of Surgery, The University of Utah, Salt Lake City, Utah. Address correspondence and reprint requests to Nelson Roy, Department of Communication Sciences and Disorders, The University of Utah, 390 South 1530 East, Room 1310, Salt Lake City, UT 84112-0252. E-mail: [email protected] Journal of Voice, Vol. -, No. -, pp. 1-14 0892-1997/$36.00 Ó 2015 The Voice Foundation http://dx.doi.org/10.1016/j.jvoice.2015.02.011
of fixation of the vocal folds via the cricoarytenoid joint, with upper airway obstruction reportedly occurring in 16% of patients.2,14–16 In addition to these laryngeal changes, many patients with RA report feeling a foreign body sensation in the throat, dysphagia, pain with speaking, and vocal fatigue.2,14,15,17 Table 1 summarizes the extant literature examining laryngeal/ voice involvement in RA. Inspection of Table 1 reveals considerable variability regarding the prevalence of laryngeal signs and phonatory symptoms in RA. Some of these discrepancies are related to inconsistent criteria of what precisely constitutes a ‘‘voice disorder,’’ methodological differences related to sampling procedures, and sample populations and sizes, among other issues. For instance, some studies relied exclusively on visual inspection of the larynx and reported a very high prevalence of laryngeal changes (ie, 80%) in patients with RA,22,23 whereas Lofgren and Montgomery18 reported laryngeal involvement in only 26% of participants on the basis of patient histories and extralaryngeal palpation, in addition to indirect laryngoscopy. Lawry et al24 compared indirect laryngoscopy versus computed tomography and documented laryngeal manifestations of RA in 32% and 54% of participants, respectively. More recently, investigators have used the voice handicap index (VHI)25 and other self-report methodologies to establish the presence of a voice disorder and reported prevalence rates in RA ranging from 5% to 48%.6,7,9,19,20 Although RA has the potential to adversely affect phonatory function, it is clear from the previously described review that significant disagreement exists regarding (1) the prevalence of voice disorders in RA, (2) patterns and symptoms of voice disorders in RA, (3) risk factors associated with these disorders, and (4) the consequences of voice disorders and their associated symptoms on social, physical, and emotional
2
TABLE 1. Studies Examining Voice Dysfunction in RA Study
# of Subjects
Gender Represented
Age (y)
‘‘Incidence of laryngeal involvement in rheumatoid arthritis’’ Lofgren and Montgomery18 ‘‘Rheumatoid nodules of the larynx’’ Woo et al4
100 participants with RA
68 women, 32 men
N/A
64-year-old woman
One woman with RA
64 y
‘‘Airway obstruction and rheumatoid arthritis’’ Vergnenegre et al15
100 participants with RA, 88 participants with other rheumatological diseases 1 woman with RA
77 women, 23 men in RA group and 67 women, 21 men in control group
External palpation and indirect laryngoscopy.
Findings 26% of RA participants examined were found to have involvement of cricoarytenoid joint.
Rheumatoid nodules can Flexible laryngoscopic affect the vocal folds, examination; which causes hoarseness laryngostroboscopy to or dysphonia; RA in the assess vibratory function; larynx is found in as microdirect laryngoscopy; many as 25% of RA and evacuation of participants. subcordal masses from each vocal fold. Number of obstructive RA group ¼ 60 ± 12 y Questionnaire, syndromes was higher in control group ¼ 57 ± 21 y spirographic participants with RA than measurements taken, other rheumatologic chest radiograph. conditions.
1 woman
57-year-old with 10-year history of RA
N/A
N/A
‘‘Prevalence and relative risk of dysphonia in rheumatoid arthritis’’ Speyer et al9
47 men, 119 women
19–89 y, Mean age ¼ 61
Arytenoid adduction surgery performed to normalize voice.
In cases of bilateral cricoarytenoid joint involvement in RA, airway obstruction can occur. VHI-10 and reflux symptom Findings suggest a prevalence of voice index to assess impairment of 5% based symptomatology of on VHI-10 score >15. laryngopharyngeal reflux; Participants with RA had a joint assessment, erythrocyte sedimentation rate (ESR), and visual analog to calculate disease activity score (DAS). Prevalence of voice Visual analog scale of disorders 15% in RA severity by participants, compared rheumatologist and with 6% in controls based patient; VHI and threeon VHI >15. item outcome scale on RA 2.9–3.7 times more likely perception of to have dysphonia. impairment.
Journal of Voice, Vol. -, No. -, 2015
‘‘Arytenoid adduction to treat impaired adduction of the vocal fold due to rheumatoid arthritis’’ Kumai et al16 ‘‘Prevalence of subjective 73 with RA voice impairment in 73 controls rheumatoid arthritis’’ Fisher et al6
166 RA 148 control
Method
109 participants with autoimmune disease (60% with RA) 41 controls
74 women and Average age similar 35 men in for autoimmune and autoimmune group control groups
80 S’s with autoimmune conditions 60 controls (vocally normal, non rheumatic)
44 subjects with RA, RA group ¼ 58.2 y control group ¼ 48.3 y 32 with systemic lupus, 4 with Sjogren syndrome.
47 participants with RA, 40 controls
40 women, 7 men in RA group 26 women, 14 men in control group
‘‘Laryngeal assessment by videolaryngealstroboscopy in patients with rheumatoid arthritis’’ Gomez-Puerta et al21
36 participants with RA
33 females, 3 males
RA group ¼ 56.5 ± 12.6 y Evaluated participants with Prevalence of dysphonia was 12.8% and 47.9 ± 13.6 in control RA according to disease videolaryngoscopy group activity score, laryngeal changes 72.4% in symptoms, VHI and participants with RA; videolaryngoscopy Posterior laryngitis was compared with a control the most common group. diagnosis. Organic involvement was 56; 3 ± 14; 2 y Reflux symptom index uncommon, one patient (RSI), voice handicap had bamboo nodules, no index-10 (VHI-10) to cricoarytenoid evaluate functional voice impairment was found, impairment, but pharyngeal-laryngeal videolaryngostroboreflux was found in 64% scopy to determine of participants. laryngeal involvement.
3
Abbreviation: N/A, not applicable.
Prevalence of dysphonia VHI, three-item VAS, was 38.6% in RA group biochemical and Cbased on VHI >15. No reative protein study in 40 clear association disease-active between disease activity participants. and VHI total score.
Epidemiology of Voice Disorders in Rheumatoid Arthritis
‘‘Voice symptoms in patients with autoimmune disease: a Cross-sectional epidemiological study’’ Liu et al7 ‘‘The prevalence of dysphonia, its association with immunomediated diseases and correlation with biochemical markers’’ Sanz et al19 ‘‘Laryngeal involvement in rheumatoid arthritis’’ Beirith et al20
Endoscopic laryngoscopy Bamboo joint-like nodes were identified in three and laryngostroboscopy female participants. and phonetogram documentation for logopedic voice investigation. Laryngoscopic evaluation, Almost 2/3s of RA participants had loss of acoustic analysis, range, and 66% had computerized hoarseness compared tomography; hoarseness, with 25% in control loss of range, vocal group; 9.1% had decrease fatigue were vocal in vocal fold mobility and symptoms inquired. 27% had moderate/ severe edema of vocal folds/arytenoids compared with none in control group. VHI-10, xerostomia scale, 16.9% of RA group reported acid reflux inquiry, and significant voice anxiety/depression scale. symptoms based on VHI10 > 7.
Nelson Roy, et al
29, 31, and 50 y ‘‘Hoarseness due to 3 female participants 3 women bamboo nodes in with bamboo patients with nodes autoimmune diseases: a review of literature’’ Hilgert et al11 ‘‘Cricoarytenoid joint 2 males 22–65; mean age 48.5 y 11 participants abnormalities in patients 9 females in RA group with RA with rheumatoid 8 controls arthritis’’ Berjawi et al5
4 functioning. This cross-sectional, descriptive epidemiological investigation aimed to improve understanding of the prevalence, risk factors, and quality of life burden of voice disorders in patients with RA. METHODS Participants To identify potential participants with RA, a chart review was undertaken of patients in the Division of Rheumatology at The University of Utah Hospital over a 3-year period. The medical records of all eligible participants were subsequently reviewed in detail by C.W. to verify diagnosis. Conventional criteria for the diagnosis of RA were used including the presence of rheumatoid factor (RF), anti-cyclic citrullinated peptide igG antibodies, and one or more of the following symptoms: morning joint stiffness, arthritis of three or more joint areas, arthritis of hand joints, symmetric arthritis, rheumatoid nodules, and/or radiographic changes.1,26,27 RF, which is an autoimmune antibody, is considered a marker for RA, with approximately 80% of individuals with RA having elevated RF levels.28 In addition to RF, another marker of RA is the presence of antibodies directed to citrullinated peptides or proteins, as measured by an anti-cyclic citrullinated peptide test. Consecutive patients with confirmed RA diagnosis (using the previously described criteria) were identified and invited to participate until 100 participants were enrolled. Inclusion criteria were diagnosis of RA, age 18 years, Englishspeaking, adequate hearing for purposes of telephone interview, and no known or reported cognitive deficits. The medical record review identified 235 eligible participants with RA who were subsequently invited to participate. Of these, 110 participants did not return the researcher’s phone calls and 25 participants declined; therefore, the final group of participants included 100 patients with RA or 42% of eligible participants. Once written consent was obtained, a telephone interview using a structured questionnaire involving primarily forced-choice responses was completed. The questionnaire required approximately 75 minutes to complete. Interviews were conducted over a 9-month period from 2012 to 2013 by trained students in the Department of Communication Sciences and Disorders at The University of Utah (IRB#00058438). The final pool of participants ranged in age from 29.3 to 95.1 years, with mean age 61.1 years (standard deviation [SD], 13.1). A description of the 100 participants is provided in Table 2. Most participants were women, were nonHispanic whites, had a current household yearly gross income of $60 000, and had at least some college education. RA disease duration ranged from 1 to 50 years, with a mean of 19.6 (SD, 13.4). The number of participants currently taking medication for RA was 95 (95%), over time ranging in years from 1 to 47 (Mean [M], 15.6; SD, 12.4). Years with RA and whether medication was currently being used for RA were not significantly associated with age, sex, race/ethnicity, income, or education. Among the 100 participants with RA, only eight had another coexisting autoimmune condition in addition to RA; four had Sj€ ogren syndrome, one of whom also had mixed con-
Journal of Voice, Vol. -, No. -, 2015
TABLE 2. Description of Participants Demographic Variable Age (y) 20–39 40–59 60–69 70–79 >80 Sex Male Female Ethnicity/race NH—white NH—Asian NH—Native American/ Alaska Native Hispanic Household yearly gross income < $20 000 $20 000–$39 999 $40 000–$59 999 >$60 000 Missing Education 80 years; P < 0.0344). RA severity also decreased with increased education level (M, 100 for < high school [HS]; M, 87.8 for HS; M, 72.2 for some college; M, 67.1 for college degree; and M, 50.9 for some graduate/graduate degree). Questionnaire description The interviewer used a comprehensive, 45-item questionnaire which probed multiple domains including six primary content areas—voice disorder prevalence, history, and symptoms/signs; medical history; social/lifestyle factors; psychosocial factors; occupational factors; and RA disease severity. For the purpose of this study, we considered a voice disorder to be ‘‘any time the voice does not work, perform, or sound as it normally should, so that it interferes with communication.’’30 This case definition of a voice disorder was previously used in the largest National Institutes of Health epidemiology study of voice disorders in the United States, and therefore, permits comparisons with the general population and other high-risk groups.30 The six primary domains assessed by the questionnaire are detailed in the following. Voice disorder history. The questionnaire examined details of the participant’s frequency, onset, and duration of current and/or previous voice concerns with the intention of clearly defining correlates of voice concerns, patterns of change across the life span, typical disease-related features, rates of treatment seeking, and outcomes. The presence of voice disorders and related symptomatology were assessed in relation to the other domains, which are detailed in the following sections. Medical history. Voice production can be influenced by a variety of medical conditions. We selected 21 conditions or disorders with the potential to influence the structure and/or function of voice and considered the presence or absence of these conditions in the participant’s lifetime and, if appropriate, mechanisms of medical treatment. The inclusion of the 21 disorders was based on previously used and validated questionnaires, a review of the present literature, and the clinical experience of the investigators.30–32 This section also included more detailed questions regarding upper respiratory concerns and the use of medications. Psychosocial history and quality of life. Psychosocial factors may contribute to, and be affected by, voice disorders.33 This section of the questionnaire probed two main areas within the psychosocial domain, including psychosocial risk factors, such as stress/anxiety and activity level and the psychosocial impact of voice disorders. In addition, two standardized batteries, the voice-related quality of life (V-RQOL)34 and the short form 36 health survey (SF-36),35 were included to examine the quality of life impact of voice disorders in RA and the relation of these disorders to general health and well-being. The VRQOL is a validated 10-item instrument that has been used to determine the impact of voice impairment on quality of life and as an outcome measure in participants with voice disorders. The V-RQOL is considered a disease-specific quality of life instrument. In contrast, the SF-36 is a psychometrically validated questionnaire developed to assess eight different measures of functional health and well-being. The SF-36 is a generic tool in that it does not target a specific age, disease, or treatment
5
group (ie, it is not disease-specific). However, it can establish the relative health burden of eight profiles (physical functioning, bodily pain, role limitations due to physical health problems, role limitations due to emotional problems, emotional wellbeing, social functioning, energy/fatigue, and general health perceptions). In addition, two component summary measures of physical and mental health are generated with four scales reflecting different dimensions of physical health and four scales reflecting different dimensions of mental health. The raw scores from the questionnaire are converted into a 0–100 scale, with 100 representing the highest state of health. General normative or ‘‘average’’ scores for each of the scale items across all participants have been reported previously, ranging from the 70s for general health, bodily pain, and energy/fatigue, to the 80s for social functioning, physical functioning, role limitations— physical health, and emotional well-being, to the 90s for role limitations—emotional stresses. To help the reader interpret the results, the highest possible score for physical functioning is obtained by performing all types of physical activity including the most vigorous without limitations due to health. For rolephysical, the highest score is obtained by having no problems with work or other daily activities. For bodily pain, the highest score is obtained by having no pain or limitations due to pain. For general health, the highest score is obtained by evaluating personal health as excellent. Likewise, the highest scores for energy/fatigue are obtained by feeling full of pep and energy all the time; for social functioning by performing normal social activities without interference due to physical or emotional problems; for role-emotional by having no problems with work or other daily activities; and for mental health by feeling peaceful, happy, and calm all the time. A brief description of each of the eight scales is provided in Appendix 1. Occupation/employment history. Research has identified a link between occupation and the risk of voice disorders.30–32 The questionnaire examined aspects of the participant’s occupational history to determine its potential role in voice disorder development and symptomatology. This section examined work history, job-related voice use patterns, and the participant’s perception of the impact of job-related voice use on voice production. Social and lifestyle history. Social and lifestyle patterns and choices may influence voice.30,33 Diet may influence these processes via its impact on systemic hydration and on levels of stomach acid refluxing into the larynx and pharynx. The use of tobacco products, alcohol, and recreational drugs may also influence function by altering the integrity of tissues in the aerodigestive tract. Thus, this section probed the participant’s dietary history, such as caffeine intake, consumption of dairy products, or the intake of spicy foods, as well as their history of using tobacco products, alcohol, and recreational drugs. Disease severity. Numerous items in the questionnaire rated the severity of RA, specific symptoms, and overall general health. Along with these items included in the main body of the questionnaire, several appendices were administered to assess the severity of comorbid autoimmune diagnoses, if present. As discussed previously, the severity of RA was assessed
6 using the RAPS, a psychometrically validated instrument developed to measure pain in adult patients (>18 years) with RA.29 Pain is a major complaint in individuals with RA and often the greatest limiting factor related to function and mobility. The 24 items on the scale are scored using a sevenpoint Likert scale ranging from ‘‘0-never’’ to ‘‘6-always’’ with a higher score representing a greater severity of pain. These scores were compared with voice symptoms and disorders (among other variables) to determine possible relationships.
Statistical analyses Conventional data analyses procedures used in descriptive epidemiology research were used.33,36 Medical, familial, environmental, and social history data were examined using contingency tables, summary statistics, chi-square tests, and risk ratios (RRs). In the following section, data analysis procedures are described in detail. The prevalence and severity of voice problems were evaluated using descriptive statistics. Variables were examined using contingency tables and multiple logistic regression. Bivariate analyses of association were evaluated for statistical significance using the chi-square (c2), the Mantel-Haenszel chisquare (MH c2), and the Cochran-Mantel-Haenszel statistic (CMH). The c2 was used to test for independence between variables, the MH c2 was used to test for differences in trends between variables, and the CMH was used to test for independence after adjusting for potential confounding factors. Tests that beta coefficients in the logistic regression models equal zero were based on the Wald test, which follows an approximate c2 distribution. Cross-tabulations were used to perform bivariate analyses between selected variables, with statistical significance on the basis of the chi-square test for independence (c2). The t statistic was used for testing the null hypothesis of equality of means between groups (ie, those with voice disorders vs those without), with the t statistic computed using approximate degrees of freedom from Satterthwaite approximation when the variances between the two groups were unequal. In addition, unadjusted and adjusted (controlling for the presence of other variables) RRs were estimated using logistic regression to establish specific risk factors for reporting voice disorders. The association between the history of a voice disorder and the presence of selected risk factors (eg, employment classification, workrelated voice use issues, history of esophageal reflux disease, frequent colds/sinus infections, chemical exposures, and many other disease entities) was assessed for homogeneity across the levels of age and sex using the Breslow-Day test for homogeneity. Multiple logistic regression modeling was used, with interaction terms among significant main effects evaluated in the model. Finally, whether the relationship between the severity of voice disorders and selected dichotomous risk factors depended on the duration, frequency, sex, or age was addressed by assessing interaction terms among the significant main effects in multiple regression models. Analyses were performed using Statistical Analysis Software (SAS) Version 9.1, SAS Institute, Inc., Cary, North Carolina37
Journal of Voice, Vol. -, No. -, 2015
RESULTS Prevalence of voice disorders Thirty-five participants with RA (ie, 35%) reported having a current voice disorder, and 20 participants (20%) reported having a voice disorder in the past. Of the participants who currently had a voice disorder, 29 (82.9%) reported that it began gradually and only six (17.1%) said it began suddenly; 30 (85.7%) reported it as a chronic problem (ie, duration of 4 weeks or longer). For those with a voice disorder, 20% first noticed the disorder within the past year, 30% 1–3 years earlier, 26.7% 4– 9 years earlier, and 23.3% 10 years earlier. The prevalence of a current voice disorder did not significantly differ across the levels of age, sex, race/ethnicity, income, or education. Voice symptoms The presence and frequency of selected voice-related symptoms identified in the sample population are presented in Table 4. Participants were asked to quantify the frequency of selected voice symptoms with ‘‘daily’’ ¼ 1, ‘‘weekly’’ ¼ 2, ‘‘monthly’’ ¼ 3, ‘‘several times a year’’ ¼ 4, and ‘‘yearly or less’’ ¼ 5. The mean frequency of the selected voice symptoms ranged from 1.7 to 3.8 (M, 3.0 or monthly; SD, 0.6). Almost one-third of participants reported current hoarseness (31%) and chronic throat dryness (36%), and almost one-half of participants (48%) complained of frequent throat-clearing behavior that occurred on a daily basis. Ten of the 15 voice-related symptoms were significantly associated with the participants currently reporting a voice disorder including hoarseness, voice tiring after using it for a short time, trouble speaking or singing softly, difficulty projecting the voice, discomfort while using the voice, increased effort to talk, chronic dryness in the throat, wet/gurgly voice quality, frequent throat clearing, and an airy or breathy voice. General health, lifestyle, voice use patterns, and voice disorders To assess the relationship between general health/medical conditions, lifestyle and personal habits, and voice use patterns and experiencing a current voice disorder, participants were asked numerous questions related to each of these areas. RRs for all variables were determined. Table 5 lists many (but not all) of the areas addressed, including the number of participants with each condition who also reported a current voice disorder and the related RRs. Inspection of Table 5 reveals that despite an exhaustive review of medical conditions, lifestyle, and vocal habits, only a small number of factors emerged as being associated with a current voice disorder. The factors associated with significant RRs included current neck/throat (RR, 2.4) and jaw tension (RR, 2.3), past history of vocally demanding occupation (RR, 2.3) and related voice problems (RR, 2.4), past history of solo singing (RR, 1.90), current cough (RR, 2.23) and throat clearing (RR, 1.78), frequent ingestion of spicy foods (RR, 1.87) and colas (RR, 1.71), current sleep disorder (RR, 1.8), and being quiet (RR, 1.7) and physically active (RR, 0.5). It is also worth noting that numerous medical conditions and other potential risk factors not listed in Table 5 failed to be significantly associated with reporting a current voice
Nelson Roy, et al
7
Epidemiology of Voice Disorders in Rheumatoid Arthritis
TABLE 4. Levels of Selected Voice-Related Symptoms Voice Disordery in Those With a Current Symptom
c2
Current
Frequency*
Symptom
%
M (SD)
n
%
P Value
Hoarseness Voice tires or changes quality after using it for a short time Trouble speaking or singing softly Difficulty projecting your voice Loss of singing range Discomfort while using your voice A monotone voice (monopitch) Takes effort to talk Chronic dryness in your throat Wet, gurgly voice quality Chronic soreness in your throat Frequently clear your throat A bitter or acid taste Wobbly or shaky voice An ‘‘airy’’ or ‘‘breathy’’ voice
31 19
3.8 (1.5) 2.7 (1.4)
20 16
65 84
