Clinical Endocrinology (2009) 70, 18–25
doi: 10.1111/j.1365-2265.2008.03347.x
ORIGINAL ARTICLE
Voice characteristics in women with congenital adrenal hyperplasia due to 21-hydroxylase deficiency Blackwell Publishing Ltd
Ulrika Nygren*, Maria Södersten*, Henrik Falhammar†, Marja Thorén†, Kerstin Hagenfeldt‡ and Agneta Nordenskjöld‡,§ *Department of Clinical Science, Intervention and Technology, Division of Logopedics and Phoniatrics, †Department of Endocrinology, Metabolism and Diabetes, ‡Department of Women and Child Health and §Department of Paediatric Surgery, Astrid Lindgren Children Hospital, Karolinska Institutet, Stockholm, Sweden
Introduction Summary Objective Congenital adrenal hyperplasia (CAH) is an autosomal recessive inherited disorder resulting in lack of cortisol and often aldosterone and overproduction of androgens. Raised levels of androgens can lead to increased mass in the laryngeal tissues, which may lower the fundamental frequency (F0) of the voice. Design/patients A follow-up study of voice characteristics in 38 women with CAH and 24 age-matched controls between 18 and 63 years of age using subjective self-ratings of voice symptoms and acoustic and perceptual analyses. The results were correlated to mutation and treatment. Measurements The subjects rated degree of hoarseness, dark voice, voice problems and vocal fatigue on visual analogue scales. Audio recordings were made of a standard reading passage in a sound treated booth. Acoustic analyses were made of the subjects’ mean, minimum and maximum F0 in the habitual speech range. Three voice clinicians made perceptual evaluations of dark voice in terms of timbre. Results Women with CAH rated higher values than controls with regard to the statement ‘my voice is a problem in my daily life’. They spoke with significantly lower mean, lower minimum and lower maximum F0, as well as darker voice quality as compared with the controls. Affected voice symptoms were associated with a late diagnosis or problems with medication. Conclusion The voices in women with CAH can be virilized because of late diagnosis or suboptimal treatment with glucocorticoids. These voice problems may need referral for voice assessment and should be considered when treating CAH. (Received 8 April 2008; accepted 24 June 2008)
Correspondence: Agneta Nordenskjöld, CMM 02, Karolinska University Hospital, SE-171 76 Stockholm, Sweden. Tel.: +46 8517 77705; Fax: +46 8517 73620; E-mail:
[email protected]
18
Congenital adrenal hyperplasia (CAH) due to 21-hydroxylase deficiency which accounts for more than 95% of all CAH patients, is an autosomal recessive inherited disorder resulting in cortisol and often aldosterone deficiency and excess production of precursor steroids with androgen effect.1,2 Incidence is 1/10,500 in Sweden according to the neonatal screening programme.3 CAH is clinically subdivided in three forms depending on the mutations and the severity of the disease.4 Classic CAH includes the forms salt wasting (SW) and simple virilizing (SV). SW is the most severe form with deficiency in both cortisol and aldosterone. These phenotypes are often caused either by null-null mutations with no activity of the enzyme or by I2 splice mutations with few percentage of the normal activity only.4 SV is a milder phenotype with mainly cortisol deficiency. Life-long glucocorticoid treatment is necessary to prevent adrenal crisis and increased production of androgens.5 Non classic (NC) CAH is a milder form with a later age of onset and symptoms such as early onset of puberty, excessive body hair or infertility.6 Increased levels of androgens are known to lower the voice pitch. This occurs normally in boys during puberty when the larynx and vocal tract increase in dimensions and the vocal folds become longer and thicker. Increased mass of the vocal folds makes them vibrate in a slower tempo, which is reflected in lower values of the fundamental frequency (F0), a measure of vocal fold vibrations per second, specified in Hertz (Hz). The boys’ voices drop about one octave and girls about a third during normal voice change in puberty. Mean F0 measured for running speech is 195 Hz (± 20 Hz) for Swedish women in the age between 20 and 40 years, and for Swedish men about 110 Hz (± 10 Hz) in the same age range.7 Mean F0 decreases with age and is about 181 Hz for women 50–70 years of age. The higher F0 (pitch) in women as compared with men is mainly due to their shorter vocal folds. The quality of the voice in terms of dark voice timbre is also affected by the length of the vocal tract, i.e. the resonance tube from the vocal folds to the lips. The shorter vocal tract in females results in higher resonance (formant) frequencies and the voice is therefore perceived as ‘brighter’ as compared with the ‘darker’ male voice. Androgens change the structure of the vocal fold tissues so that the muscle mass increases, although the changes are hardly visible © 2009 The Authors Journal compilation © 2009 Blackwell Publishing Ltd
Voice characteristics in women with CAH 19 in a laryngoscopic examination.8 The laryngeal changes due to virilization are irreversible.8–10 One would expect that such vocal changes are rare today among women with CAH because of early diagnosis and adequate treatment. These women are, however, in reality at risk of being periodically exposed to pathologically high levels of androgens such that their voices might be affected. The risk assessment is uncertain since there are only few reports in the literature addressing the voice function in females with CAH. In one case report, a 17-year-old girl with severe virilization due to CAH and insufficient medical treatment was described.11 Her voice was masculine with a mean F0 of 104 Hz, a frequency normal for an adult male voice. In a German study, 10 girls and 3 boys with CAH, who ranged in age from 5 to 16 years, were examined with laryngoscopy, perceptual analysis of pitch, voice range and voice quality. In this study, these junior voices were characterized by a remarkably dark voice timbre and in some cases also by hoarseness and low pitch.12 When assessing voice characteristics a factor to consider is smoking since it is a common cause of lower fundamental frequency of the voice due to oedema in the vocal folds.13 This so-called Reincke’s oedema is particularly common in women who have smoked heavily during many years. Women with Reincke’s oedema are often perceived as men in the telephone because of their deep and sometimes rough voices. Vocally demanding activities with extensive and loud talking may also affect the voice function and lead to subjective voice symptoms such as vocal fatigue and hoarseness. Many occupations are vocally demanding and women tend to develop functional voice problems more often than men.14 The purpose of this study was to investigate voice characteristics such as fundamental frequency and voice timbre in adult women with CAH as compared with control subjects. We also wanted to study if women with CAH have subjective voice problems and whether all these parameters were correlated to either mutation (severity of phenotype) or androgen load, i.e. late diagnosis or poorly controlled medication.
Methods
Fig. 1 Age distribution in cases (CAH, n = 38) and controls (CTR, n = 24).
Medical data from journals Information on medical history including age of diagnosis and adherence to glucocorticoid and mineral corticoid therapy was collected as well as data on body height, body mass index (BMI), and waist-hip ratio. Total body bone mineral density (BMD Z score), fat and lean mass were measured with dual energy X-ray absorptiometry (DXA) expressed as standard deviations from age-related normal mean (z-score). In addition, type of CYP21A2 mutation were analysed.3,15 Hence, common clinical data for the 38 women with CAH were compared with their respective 38 age-matched controls from the larger study. Questionnaire Both patients and controls answered a questionnaire concerning smoking and profession. All subjects also rated their voice function based on four statements ‘my voice is hoarse’, ‘my voice is dark’, ‘my voice is a problem in my daily life’ and ‘I get tired in my throat when speaking’. The ratings were given on 100-mm visual analogue scales, 0 mm indicated ‘not at all’ and 100 mm ‘high degree of ’. Recordings
Subjects 3,15,16
This report is part of a larger Swedish follow-up study. Thirty-eight women with CAH, ranging in age from 18 to 63 years (mean age 31), were recruited through their physicians or informed by the National CAH patient organization. The age distribution is shown in Fig. 1. The control subjects were recruited from the Swedish Population Registry and born on the same day as the patients. They were examined one year after the patients. Because of that it was only possible to recruit 24 age matched controls to the 38 patients. The control subjects were ranging in age from 19 to 63 years (mean age 32). All subjects were examined as outpatients between 2003 and 2006 at the Department of Speech Language Pathology, Karolinska University Hospital in Stockholm, Sweden. The investigator did not know the status of the patients with regard to clinical diagnosis and the mutations when the voice recordings and the analyses of these were performed.
Digital audio recordings of the voice were made in a sound treated booth following clinical routines.17 A head mounted electret microphone was placed at a distance of 15 cm from the mouth. The reading sample was a Swedish standard text of about 40 s duration. The subjects read the text aloud twice. The first reading was saved as a sound file for acoustic analysis of F0 and for perceptual evaluation. The second reading was directly recorded using a phonetogram program. Acoustic analysis of fundamental frequency Acoustic analyses were made of mean F0 using the computer-based programme Soundswell (Saven Hitech, Enhagsvägen, Sweden; http://www.savenhitech.se) and presented in histograms (Fig. 2a). The recordings of the reading were also presented in habitual speech range phonetograms using the computer phonetograph Phog (Saven Hitech). A phonetogram is a two-dimensional image of the voice
© 2009 The Authors Journal compilation © 2009 Blackwell Publishing Ltd, Clinical Endocrinology, 70, 18–25
20 U. Nygren et al.
Fig. 2 (a) An example of a histogram of fundamental frequency (F0) showing a mean value of 209 Hz. The x-axis shows fundamental frequency (Hz) and the y-axis the percentage of occurrence. (b) An example of a speech phonetogram obtained during reading of a standard text. The x-axis shows the fundamental frequency in Hz and in semitones as illustrated by the keyboard above. The y-axis shows the sound intensity of the voice in dB (30 cm). The arrows show where the data for maximum and minimum F0 were chosen. Extreme measures were not included in the analysis.
regarding fundamental frequency and intensity measured over time.18 Data for the lowest F0 (minimum F0) and the highest (maximum F0) during the reading were collected (Fig. 2b). Speech range phonetograms are missing for four patients because of technical problems the day of the recording. The voice analysis programs Soundswell and Phog are used routinely in voice clinics in Sweden. Perceptual analysis Three voice clinicians made perceptual analysis regarding dark voice timbre using visual analogue scales based on the Stockholm Voice Evaluation Approach, SVEA.19 The scales were 100 mm where 0 mm indicated ‘not at all’ and 100 mm ‘high degree of ’. Seven voices were randomly chosen and duplicated for calculation of intrajudge reliability (test–retest). In total, the listeners judged 69 voices (38 patients, 24 controls and seven duplications), which were played in random order to them. The listeners were informed only about the subjects’ ages and that all were female voices. A training session preceded the actual rating session, which took in total about 2 h. The Research Ethics Committee, Karolinska Institutet, Stockholm, Sweden approved the study. All participants gave their written informed consent. Statistics Results are presented as mean ± SD if not otherwise stated. Comparison between the whole group of the 38 patients and the 24 controls were carried out with unpaired t-test when normally distributed values. Otherwise, the Mann–Whitney U-test was used and in these cases median and range are described. Paired t-test or Sign test were performed in analyses for the age-matched pairs. When 2 × 2 frequency table calculations were performed, Fischer’s exact test was used. One-way anova was used to compare the 3 CAH
forms (SW, SV and NC) regarding data from the acoustic analysis and Kruskal–Wallis anova by Ranks regarding data from self-ratings and the perceptual analysis. Comparison between women with CAH with dark voice or normal voice and their age-matched controls regarding the clinical variables were analysed using a two-way repeated measures anova. The between-groups factor in the anova was ‘Voice group’ (dark, normal) and the within-groups factor was ‘Matched pair’ (CAH, control). Spearman Rank Order Correlation was used for analysis of correlation between different data. A P-value < 0·05 was considered statistically significant for all analyses above. Intraclass correlation (ICC) was used for analyses of intra- and interjudge reliability.
Results The patient cohort consisted of 17 women with SW (10 null and seven I2splice mutation), 15 were SV and six had the NC form of CAH. Smoking and profession No significant difference in smoking between the patient and the control group was found. Eight of the patients smoked, all less than 15 cigarettes per day. Among the controls, two did not respond and two smoked less than 15 cigarettes per day. Ten of the patients were regarded as having vocally demanding professions, such as teacher, interpreter, sales person, nurse and medical doctor. Three patients did not respond. Among the control subjects 10 were regarded having vocally demanding professions. Subjective self-ratings of voice function Women with CAH rated significantly higher values on the statement: ‘my voice is a problem in my daily life’, as compared with the controls
© 2009 The Authors Journal compilation © 2009 Blackwell Publishing Ltd, Clinical Endocrinology, 70, 18–25
Voice characteristics in women with CAH 21
Fig. 3 Box plot diagram showing ratings regarding the statement ‘my voice is a problem in my daily life’ for patients (CAH, n = 38) and controls (CTR, n = 24). The ratings were given on 100 mm visual analogue scales (VAS), 0 mm indicated ‘not at all’ and 100 mm ‘high degree of ’ (z = 2·457, P = 0·013, Mann–Whitney U-test).
(P = 0·013, Fig. 3). No significant differences were found regarding the other three statements compared with controls. The results however, showed tendencies towards higher values in cases regarding the statements ‘my voice is dark’, ‘I get tired in my voice when speaking’ and ‘my voice is hoarse’ in the patient group. No statistical differences were detected between the three different forms of CAH. Voice characteristics The intrajudge reliabilities for judging dark voice timbre were high, ranging from 0·82 to 0·98 and the interjudge reliability was 0·71. Results from the perceptual evaluation showed that the patient group had a significantly darker voice timbre as compared with the controls (P = 0·002, Fig. 4a). Furthermore, when the patients were compared with their matched controls, the perceptual analysis revealed that they had a darker voice timbre than their matched controls (P = 0·025, Fig. 4b). No correlation was found between different CAH-mutations and dark voice timbre. The patients had a significantly lower mean F0 as compared with the controls (P = 0·039, Fig. 4c). The statistical calculation in Hz was confirmed by a statistical calculation of the differences in semitones (P = 0·036). Seven of the 38 patients (ranging in age from 19 to 51 years) had a mean F0 between 116 and 170 Hz. Their voices were also rated high in the perceptual evaluation of dark voice timbre (Fig. 5). Of these seven patients with clearly deviant voices four were SW, two were SV and one NC. Three of these patients were smokers. The patient with CAH with the lowest mean F0 smoked about 10–15 cigarettes per day. Another 6 patients had voices with mean F0 in the lower SD range according to age and with some degree of dark voice timbre. There were no differences between the three clinical forms of CAH regarding mean F0. Women with CAH had significantly lower minimum F0 (P = 0·016) and lower maximum F0 (P = 0·039) obtained from the habitual speech range phonetogram as compared with the controls, i.e. the patients speaking range was lower down in frequency than found for normals. These findings were not correlated to severity of the mutation.
Fig. 4 (a) Box plot diagram over dark voice timbre for the patients (CAH, n = 38) and the control subjects (CTR, n = 24). A 100-mm visual analogue scale (VAS) was used for the perceptual ratings, 0 mm corresponded to ‘not at all’ and 100 mm to ‘high degree of ’ (z = 3·150, P = 0·002, Mann–Whitney U-test). (b) Results from the 24 patients (CAH) with age-matched controls (CTR) regarding dark voice timbre perceptually rated on a 100 mm visual analogue scale (VAS), 0 mm corresponds ‘not at all’ and 100 mm ‘high degree of ’ (z = 2·245, P = 0·025, Sign Test). (c) Box plot diagram of mean fundamental frequency (mean F0) in Hz for the patients (CAH, n = 38) and controls (CTR, n = 24) (t = −2·108, P = 0·039, t-test for independent samples).
Associations between acoustic and perceptual analysis The correlation between the results from the acoustic analysis of mean F0 and the perceptual analysis of dark voice was high for the patients (r = −0·84, P < 0·001). The correlation between the patient’s self-ratings of the statement ‘my voice is dark’ and mean F0 was
© 2009 The Authors Journal compilation © 2009 Blackwell Publishing Ltd, Clinical Endocrinology, 70, 18–25
22 U. Nygren et al.
Fig. 5 Mean fundamental frequency (mean F0, left y-axis) in bars and dark voice timbre (right y-axis) in dotted curve for all patients with CAH in order from the lowest to highest ratings of dark voice timbre. The three different forms of CAH are marked by patterns of the bars (SW = black, SV = grey, NC = striped). Mean F0 for vocally healthy women, 20–40 years of age, is indicated by a horizontal dotted line. Seven cases to the right of the dashed line had clearly deviant voices, of which four were SW, two SV and one NC. Six patients with mean F0 in the lower SD-range and with some degree of dark voice timbre are marked with arrows.
significant with moderate strength (r = −0·51, P < 0·001) as was also the correlation between the statement and the perceptual analysis of dark voice timbre (r = 0·56, P < 0·001). Correlation between voice characteristics and clinical parameters No differences were found between age, mutation, body height, waist/hip ratio, BMD Z score or total fat comparing the group with ‘dark voice’, meaning both low mean F0 and dark voice timbre (n = 13), and the CAH group with normal female voice (n = 25) (Table 1). BMI and lean body mass however, was significantly higher in the ‘dark voice’ group as compared with matched controls (P = 0·01 and P = 0·001, respectively, Table 1). BMI and lean body mass was also significantly higher in women with CAH and dark voice compared with those with normal voice (P = 0·009 and 0·007, respectively). In the group with ‘dark voice’, seven women with CAH had had periods of incomplete compliance for medical treatment especially during puberty, and in 5 women with CAH the diagnosis and treatment had been delayed after start of symptoms (Table 1). In the group with normal voice (n = 25) these factors were found in five and two cases, respectively.
Discussion Androgen-related voice changes in women are well known effects after hormonal treatments of e.g. menopausal symptoms, gynaecological carcinoma, fibrocystic breast disease and endometriosis.20 The drugs used in those treatments are testosterone or testosterone derivatives (e.g. danazol, gestrinone or androgenic progestogens) and this may cause the adverse side effects deepening of the voice, pitch fluctuations, vocal weakness and hoarseness.9,20 Women who use anabolic steroids, develop a masculine physical appearance and a deep voice. These side effects are both time and dose dependent. Thus, for example in female-to-male transsexuals treatment with high doses of testosterone result in voice changes after about
3 months. Androgens cause virilization by changing the structure of the vocal fold tissues irreversibly so that the muscle mass increases.8–10 A recent publication demonstrates the presence of androgen 21 receptors in the vocal folds. The development of male voice features can be emotionally difficult for women who may need voice therapy to improve the voice function and voice quality.22 Another option would be phonosurgical treatment to raise fundamental frequency, a procedure that has been tried out experimentally23 and used in male to female patients with transsexualism.24 Smoking may lead to lowering of the pitch because of vocal fold oedema. There was, however, no statistical difference regarding smoking between the two groups, thus it is unlikely that smoking affected the group results. Besides the amount of smoking also the duration of smoking seems to be important for the development of vocal fold oedema. However, those relationships are not well studied and not all heavy smokers develop vocal fold oedema. To fully exclude Reincke’s oedema as cause of low pitch in selected cases, a laryngoscopic assessment would have been necessary. Therefore laryngoscopic evaluation is recommended to be included in future studies when fundamental frequency is a variable of interest. No previous studies are published where patients with CAH have rated their voice function. In the present study the patient group rated high values on the statement that ‘my voice is a problem in my daily life’. However, what specific kind of voice problems the patients refer to was not asked for in the questionnaire. One reason for subjective voice problems may be that the voices were deviant and masculine due to the low pitch and dark voice timbre. A female with a low F0 in the male range may be perceived as a man, i.e. on the telephone. Another reason for subjective voice problems may be related to vocally demanding activities in occupations, such as teachers or sales persons.14 Risk factors for developing voice symptoms are i.e. extensive voice use, loud speech to be heard in high background noise, room acoustics which are not optimal for speech production or not enough voice rest.25 Women seem to be more sensitive to vocal overload and get more functional voice problems than men.26 Ten of the 38 of patients with CAH had chosen professions
© 2009 The Authors Journal compilation © 2009 Blackwell Publishing Ltd, Clinical Endocrinology, 70, 18–25
Voice characteristics in women with CAH 23 Table 1. Comparison between women with CAH due to 21-hydroxylase with dark voice, normal voice and age-matched controls
Medium age (year) (range) Mutation SW (n) SV (n) NC (n) Body height (cm) (range) W/H ratio (range) BMD Z score (range) Body fat (%) 2 BMI (kg/m ) (range)
Lean mass (kg/m2) (range) Age at diagnosis Neonatally (n) First year (n) Before puberty (n) As adult (≥ 18 years) (n) Poorly controlled to treatment (n/total n)* Late diagnosis after first symptom (n/total n)†
‘Dark voice’ CAH group (n = 13)
Matched controls (n = 13)
Normal voice CAH group (n = 25)
Matched controls (n = 25)
31·6 (18–63)
32·7 (19–63)
NA
30·0 (18–57)
5 5 3 159·7 (147–172) 0·82 (0·72–0·93) 0·63 (–0·7–1·9) 37·8 (26·3–51·7) 27·9 (20·7–41·8) 16·3 (14·1–19·2)
NA NA NA 168·2 (161–173) 0·8 (0·73–0·87) 0·85 (0–1·9) 35·7 (25·1–47·7) 23·1 (19·9–29·2) 14·2 (12·5–15·6)
NA NA NA P < 0·001
3 3 5 2 7/13 (54%) 5/13 (38%)
NA NA NA NA NA
NA NA NA NA NA
NA
NA
P-value (P vs C)
NS NS NS P = 0·01 P = 0·001
P-value (P vs C)
P-value Dark voice vs normal voice
31·1 (20–57)
NA
NS
12 10 3 161·2 (151–175) 0·80 (0·69–0·96) 0·067 (–1·7–2·0) 32·0 (17·2–45·3) 23·1 (18·4–32·6) 14·7 (12·2–17·5)
NA NA NA 166·2 (153–175) 0·80 (0·68–0·95) 0·77 (–2·0–1·9) 33·3 (19·4–49·6) 23,3 (17·5–33·8) 14·5 (12·5–16·8)
NA NA NA P < 0·001
NA NA NA NS
NS
NS
NS
NS
NS NS
NS (P = 0·052) P = 0·009
NS
P = 0·007
9 10 6 0 5/25 (20%) 2/25 (8%)
NA NA NA NA NA
NA NA NA NA NA
NA NA NA NA NA
NA
NA
NA
P vs C: patients vs controls. W/H ratio: waist circumference/hip circumference ratio; NA: not applicable; NS: not significant. *Define as at least 12 months with high circulating androgen levels, mainly during the late puberty. †Defined as > 2 years of symptoms before treatment.
in which the voice is an important working tool, as compared with 10 of the 24 controls. One may wonder if women with CAH avoid certain professions because of their voice characteristics. To further investigate voice problems in this patient group, questionnaires like the Voice Handicap Index (VHI), or Voice Activity Participation Profile (VAPP) could be used.27,28 Mean F0 was low in the CAH group as a whole, and 7 of the 38 patients had a very low mean F0 between 116 and 170 Hz. For vocally healthy women in the same age as the patients the mean value is 195 Hz (± 20) in Sweden.7 Plausible reasons for a low mean F0 in women with CAH could be severity of mutation, under-treatment or late diagnosis. There was no correlation to mutation as shown in Table 1. In addition, 3 women with CAH with the NC form were among the 13 women with dark voice. The main finding when comparing clinical parameters was that women with CAH with dark voice had significantly higher BMI and lean body mass than those with normal voice. Although most of the patients had suppressed androgens at the time of the investigation,3 the higher muscle mass
might be the result of exposure to elevated androgens in the past. These data are supported by the finding that among the 13 women with CAH with low pitch and dark voice timbre, 7 had been subjected to under-treatment during long periods in life for different reasons and 5 had had debut symptoms for an extended time before start of an adequate treatment. In only one case, with mean F0 177 Hz, there was no obvious underlying reason. In the CAH group with normal voice, an extended period of under-treatment, that may result in higher serum concentration of adrenal androgens, was found in less than one third (7/25) (Table 1). Among the four patients with extremely low mean F0 (< 154 Hz) two had had symptoms during childhood for 5–7 years before treatment and the other two had been extremely resistant to treatment during puberty. This may indicate that the crucial period for voice changes appears during growth. All patients with CAH had been treated orally with glucocorticoids, most of them lifelong. We have shown earlier that the patients with CAH as a group had been over-treated as indicated by depressed androgens and increased effect on bone mineral.3,15 It
© 2009 The Authors Journal compilation © 2009 Blackwell Publishing Ltd, Clinical Endocrinology, 70, 18–25
24 U. Nygren et al. is not known whether these drugs themselves can have an effect on the voice, being drugs that reduce oedema it should be the opposite. However, inhalation of corticosteroids can cause reversible hoarseness but not the voice changes discussed here.10,29 The frequency analysis showed that minimum and maximum F0 as well as the habitual speech range for the patients were lower than for the normal subjects. This is an interesting finding and could be further investigated by performing maximum Voice Range Phonetograms that would reveal the patients total physiological voice range and compare with data for healthy Swedish women.18,30 The finding that the patients had significantly darker voice timbre than the controls was in agreement with the only previously published study.12 Therefore, the parameter dark voice timbre would be important to add to the evaluation protocol when performing perceptual analysis in voices with these characteristics in relation to increased levels of androgens and gender aspects. A previous study, as well as clinical experience, has shown that voice timbre is a crucial voice parameter besides pitch if male-to-female transsexuals should pass as women based on their speech.31 It would be of interest in the future to measure formant frequencies in women with CAH having a dark voice timbre to increase our knowledge whether the vocal tract can be affected by increased androgen production besides virilization of the larynx.
Conclusions Women with CAH have lower mean F0 and darker voice timbre and rate higher values regarding the statement ‘my voice is a problem in my daily life’ as compared with controls. There was no correlation to the severity of the disease, i.e. mutations, which suggests that the voice characteristics in this patient group such as low mean F0 and dark voice timbre are caused by late diagnosis or under-treatment with glucocorticosteroids postnatally. Higher lean body mass in the group of CAH women with dark voice may be another result of previous exposure to elevated androgens. The results also suggest a need for information to medical staff as well as women with CAH that voice problems may occur and thus, referral for voice assessment and maybe voice treatment may be needed. Physicians caring for CAH patients especially during childhood and puberty need to acknowledge the risk for altered voice and to avoid long periods of high androgen influence to prevent irreversible voice changes.
Acknowledgements We want to thank all participating women for taking time and interest in this study as well as acknowledge the excellent help from Anette Härström, RN and Ingrid Hansson, RN as well as to Elisabeth Berg at the Department of Learning, Informatics, Management and Ethics (LIME), Karolinska institutet for overall statistics support, Anna Starbäck and Eva Borell for help with the perceptual analyses, Hans Larsson for technical support with the acoustic analyses, Eva B. Holmberg for revising the language and Svante Granqvist for help with the figures. This study was supported by grants from the Swedish Research Council, HRH Crown Princess Lovisa Foundation, ALF-Foundation Stockholm, Foundation Frimurarna and the Samariten Foundation.
References 1 White, P.C. & Speiser, P.W. (2000) Congenital adrenal hyperplasia due to 21-hydroxylase deficiency. Endocrine Reviews, 21, 245 – 291. 2 Merke, D.P. & Bornstein, S.R. (2005) Congenital adrenal hyperplasia. Lancet, 365, 2125 – 2136. 3 Falhammar, H., Filipsson, H., Holmdahl, G., Janson, P.-O., Nordenskjöld, A., Hagenfeldt, K. & Thorén, M. (2007) Metabolic profile and body composition in adult women with congenital adrenal hyperplasia due to 21-hydroxylase deficiency. Journal of Clinical Endocrinology and Metabolism, 92, 110 –116. 4 Wedell, A., Thilen, A., Ritzen, M., Stengler, B. & Luthman, H. (1994) Mutational spectrum of the steroid 21-hydroxylase gene in Sweden: implication for genetic diagnosis and association with disease manifestation. Journal of Clinical Endocrinology and Metabolism, 78, 1145 –1152. 5 Thilén, A., Nordenström, A., Hagenfeldt, L., von Döbeln, U., Guthenberg, C. & Larsson, A. (1998) Benefits of neonatal screening for congenital adrenal hyperplasia (21-hydroxylase deficiency) in Sweden. Pediatrics, 101 (4), E 1 1 . 6 Pinto, G., Tardy, V., Trivin, C., Thalassinos, C., Lortat-Jacob, S., Nihoul-Fekete, C., Morel, Y. & Brauner, R. (2003) Follow-up of 68 children with congenital adrenal hyperplasia due to 21-hydroxylase deficiency: relevance of genotype for management. Journal of Clinical Endocrinology and Metabolism, 88, 2624 –2633. 7 Pegoraro Krook, M. (1988) Speaking fundamental frequency characteristics of normal Swedish subjects obtained by glottal frequency analysis. Folia Phoniatrica, 40, 82 – 90. 8 Sheppard, H. (1966) Androgenic hoarseness. The Journal of Laryngology and Otology, 80, 403– 405. 9 Pattie, M.A., Murdoch, B.E., Theodoros, D. & Forbes, K. (1988) Voice changes in women treated for endometriosis and related conditions: the need for comprehensive vocal assessment. Journal of Voice, 12, 366 – 371. 10 Martin, G. (1988) Drugs and vocal function. Journal of Voice, 2, 338 – 344. 11 Fürst-Recktenwald, S., Dörr, H.G. & Rosanowski, F. (2000) Androglottia in a young female adolescent with congenital adrenal hyperplasia and 21-hydroxylase deficiency. Journal of Pediatric Endocrinology and Metabolism, 13, 959 – 962. 12 Heinemann, M. (1974) Laryngeal and voice findings in congenital syndrome with adrenocortical hyperplasia. Folia Phoniatrica, 26, 450 – 460. 13 Fritzell, B., Sundberg, J. & Strange-Ebbesen, A. (1982) Pitch change after stripping oedematous vocal folds. Folia Phoniatrica, 34, 29– 32. 14 Fritzell, B. (1996) Voice disorders and occupations. Logopedics Phoniatrics Vocology, 21, 7 –12. 15 Falhammar, H., Filipsson, H., Holmdahl, G., Janson, P.-O., Nordenskjöld, A., Hagenfeldt, K. & Thorén, M. (2007) Fractures and bone mineral density in adult women with congenital adrenal hyperplasia due to 21-hydroxylase deficiency. Journal of Clinical Endocrinology and Metabolism, 92, 4643 – 4649. 16 Nordenskjöld, A., Holmdahl, G., Frisén, L., Falhammar, H., Filipsson, H., Thorén, M., Janson, P.-O. & Hagenfeldt, K. (2008) Type of mutation and surgical procedure affect long-term quality of life for women with congenital adrenal hyperplasia. Journal of Clinical Endocrinology and Metabolism, 93, 380 – 386. 17 Hammarberg, B. (1986) Perceptual and acoustic analysis of dysphonia. Doctoral Dissertation. Karolinska Institutet, Huddinge University Hospital, Stockholm, Sweden.
© 2009 The Authors Journal compilation © 2009 Blackwell Publishing Ltd, Clinical Endocrinology, 70, 18–25
Voice characteristics in women with CAH 25 18 Sulter, A., Wit, H., Schutte, H. & Miller, D. (1994) A structured approach to voice range profile (phonetogram) analysis. Journal of Speech and Hearing Research, 37, 1076 –1085. 19 Hammarberg, B. (2000) Voice research and clinical needs. Folia Phoniatrica et Logopaedica, 52, 93 –102. 20 Greene, M. & Mathieson, L. (2001) The Voice and Its Disorders, 6th edn. Whurr Publishers, London and Philadelphia. 21 Voelter, Ch, Kleinsasser, N., Joa, P., Nowack, I., Martínez, R., Hagen, R. & Voelker, H.U. (2008) Detection of hormone receptors in the human vocal fold. European Archives of Otorhinolaryngology, 265, 1239–1244. 22 Gelder, L. (1974) Psychosomatic aspects of endocrine disorders of the voice. Journal of Communication Disorders, 7, 257 – 262. 23 Tanabe, M., Haji, T., Honjo, I. & Isshiki, N. (1985) Surgical treatment of androphonia. Folia Phoniatrica, 37, 15 – 21. 24 Kanagalingam, J., Georgalas, C., Wood, G.R., Ahluwalia, S., Sandru, G. & Cheesman, D. (2005) Cricothyroid approximation and subluxation in 21 male-to-female transsexuals. The Laryngoscope, 115, 611– 618. 25 Vilkman, E. (2004) Occupational safety and health aspects of voice and speech professions. Folia Phoniatrica et Logopaedica, 56, 220 – 253.
26 Södersten, M., Ternström, S. & Bohman, M. (2005) Loud speech in realistic environmental noise: phonetogram data, perceptual voice quality, subjective ratings and gender differences in healthy speakers. Journal of Voice, 19, 29 – 46. 27 Jacobson, B.H., Johnson, A., Grywalski, C., Silbergleit, A., Jacobson, G., Benninger, M.S. & Newman, C.W. (1997) The voice handicap index (VHI): development and validation. American Journal of Speech-Language Pathology, 6, 66 – 70. 28 Ma, E.P. & Yiu, E.M. (2001) Voice activity and participation profile: assessing the impact of voice disorders on daily activities. Journal of Speech, Language, and Hearing Research, 44, 511– 524. 29 Ihre, E., Zetterström, O., Ihre, E. & Hammarberg, B. (2004) Voice problems as side effects of inhaled corticosteroids in asthma patients – a prevalence study. Journal of Voice, 18, 403 – 414. 30 Holmberg, E.B., Ihre, E. & Södersten, M. (2007) Phonetograms as a tool in the voice clinic: changes across voice therapy for patients with vocal fatigue. Logopedics Phoniatrics Vocology, 32, 113 –127. 31 Söderpalm, E., Larsson, A. & Almquist, S.A. (2004) Evaluation of a consecutive group of transsexual individuals referred for vocal intervention in the west of Sweden. Logopedics Phoniatrics Vocology, 29, 18 – 30.
© 2009 The Authors Journal compilation © 2009 Blackwell Publishing Ltd, Clinical Endocrinology, 70, 18–25
