- Email: [email protected]
Voice Formants in Individuals With Congenital, Isolated, Lifetime Growth Hormone Deficiency
Voice Formants in Individuals With Congenital, Isolated, Lifetime Growth Hormone Deficiency *Eugenia H. O. Valenc¸a, †Roberto Salvatori, ‡Anita H. O. Souza, §Luiz A. Oliveira-Neto, §Alaı´de H. A. Oliveira, jjMaria I. R. Gonc¸alves, ‡Carla R. P. Oliveira, {Jeferson S. D’Avila, #Valdinaldo A. Melo, *Susana de Carvalho, *Bruna M. R. de Andrade, *Larisse S. Nascimento, *Savinny B. de V. Rocha, *Thais R. Ribeiro, {Valeria M. Prado-Barreto, #Enaldo V. Melo, and ‡Manuel H. Aguiar-Oliveira, *zx{#Aracaju, Sergipe and kS~ao Paulo, Brazil,
and yBaltimore, Maryland
Summary: Objective. To analyze the voice formants (F1, F2, F3, and F4 in Hz) of seven oral vowels, in Brazilian Portuguese, [a, ε, e, i, ɔ, o, and u] in adult individuals with congenital lifetime untreated isolated growth hormone deficiency (IGHD). Study Design. This is a cross-sectional study. Methods. Acoustic analysis of isolated vowels was performed in 33 individuals with IGHD, age 44.5 (17.6) years (16 women), and 29 controls, age 51.1 (17.6) years (15 women). Results. Compared with controls, IGHD men showed higher values of F3 [i, e, and ε], P ¼ 0.006, P ¼ 0.022, and P ¼ 0.006, respectively and F4 [i], P ¼ 0.001 and lower values of F2 [u], P ¼ 0.034; IGHD women presented higher values of F1 [i and e] P ¼ 0.029 and P ¼ 0.036; F2 [ɔ] P ¼ 0.006; F4 [ɔ] P ¼ 0.031 and lower values of F2 [i] P ¼ 0.004. IGHD abolished most of the gender differences in formant frequencies present in controls. Conclusions. Congenital, severe IGHD results in higher values of most formant frequencies, suggesting smaller oral and pharyngeal cavities. In addition, it causes a reduction in the effect of gender on the structure of the formants, maintaining a prepubertal acoustic prediction. Key Words: Growth hormone–Voice–Formant frequencies–Acoustic analysis. INTRODUCTION Growth hormone (GH) has fundamental roles on stature and voice, two important factors for self-confidence and social acceptation. The consequences of lack of GH on voice are difficult to study, as isolated GH deficiency (IGHD) is a rare disease occurring in 1 in 3480 to 1 in 10 000 live births.1 Furthermore, most GH deficient children in the developed world are treated with GH replacement therapy during childhood. We have identified, in Northeastern Brazil (Itabaianinha County, the State of Sergipe), a cohort of individuals with IGHD caused by a homozygous (c.57 + 1 G > A) null mutation in the GH releasing hormone receptor (GHRHR) gene.2 Most of the adults have never been treated with GH replacement. The untreated individuals have very low serum levels of GH, and of its principal mediator insulin-like growth factor-I, but normal secretion of all the other pituitary hormones.3 This results in severe short stature, associated with a typical doll-like face, and high-pitched voice.4 The voice effect seems to be more homogenous than the degree of stature deficit. Although the adult height ranges from 107 to 137 cm in pooled genders,4 the funda-
Accepted for publication March 24, 2015. From the *Department of Speech and Hearing Therapy, Federal University of Sergipe, S~ao Crist ov~ao, Sergipe, Brazil; yDivision of Endocrinology, Department of Medicine, The Johns Hopkins University, School of Medicine, Baltimore, Maryland; zDivision of Endocrinology, Department of Medicine, Federal University of Sergipe, Aracaju, Sergipe, Brazil; xDepartment of Dentistry, Federal University of Sergipe, Aracaju, Sergipe, Brazil; jjDepartment of Speech and Hearing Therapy, Federal University of S~ao Paulo, S~ao Paulo, Brazil; {Division of Otorhinolaryngology, Department of Medicine, Federal University of Sergipe, Aracaju, Sergipe, Brazil; and the #Department of Medicine, Federal University of Sergipe, Aracaju, Sergipe, Brazil. Address correspondence and reprint requests to Eugenia H.O. Valenc¸a, Division of Speech Therapy, Federal University of Sergipe, Rua Claudio Batista, S/N Bairro Sanatorio, 49060-100 Aracaju, Sergipe, Brazil. E-mail: [email protected] Journal of Voice, Vol. -, No. -, pp. 1-6 0892-1997/$36.00 Ó 2015 The Voice Foundation http://dx.doi.org/10.1016/j.jvoice.2015.03.015
mental voice frequency (f0) is high, and very similar in both genders, not only in adulthood but also in childhood and senescence.5,6 We had previously reported that these IGHD individuals also present marked reduction of cephalic perimeter7 and a typical pattern of reduction of cephalometric measurements, with marked reduction in both maxillary and mandibular length.8 In normal subjects, a significant negative association between the formant frequencies and the length of the mandible and maxilla has been reported.9 Therefore, we hypothesized that the IGHD individuals may have higher formant frequencies. In this article, we have studied the pattern of vowel formants (F) frequencies in adult IGHD individuals never treated with GH. SUBJECTS AND METHODS Subjects were recruited by word of mouth and by an ad placed in the local dwarfs’ association building, located in Itabaianinha County. Exclusion criteria were as follows: age <20 years and obvious mental or speech deficiencies. Inclusion criteria were as follows: Portuguese language native speakers, and for IGHD group, being homozygous for the GHRHR c.57 + 1 G > A mutation and having received no previous GH treatment. Inclusion criteria for controls: normal statured individuals living in the same area, with normal genotyped GHRHR alleles. We had previously reported f0 data and the perceptual-auditory analysis (Grade, Roughness, Breathiness, Asthenia, Strain scale, GRBAS scale) and video laryngostroboscopic assessment5,6 in subjects from this cohort. This cross-sectional study included 33 IGHD (16 women, 47.3 (16.6) years, 117.5 (5.6) cm and 17 men, 41.8 (18.6) years, 129.4 (5.2) cm); and 29 controls (15 women, 48.8 (13.7) years, 155.2 (6.3) cm and 14 men, 53.6 (21.5) years, 165.3 (9.0) cm).
2 F1 (affiliated to the posterior cavity, pharynx), F2 (to the anterior oral cavity), F3 (to the cavities above the vocal folds and around the inferior incisors), and F4 (to the laryngeal tube length)10 of the seven oral vowels in Brazilian Portuguese [a, ε, e, i, ɔ, o, and u] were measured (Hz) in a quiet room with <40 dB of noise. After training, 3-second sustained vowel emission was studied, using comfortable conversational pitch and loudness levels (SM58 microphone; Shure Inc., Niles, IL) kept at fixed distance of 5 cm from the subject’s mouth.11 Three trials were performed per each vowel. The criteria for trial acceptance/rejection were frequency stability in a range from 0.1 to 3 seconds.12 Samples were digitally recorded with the by software Sound Forge 10.0 (Professional audio program Sony Creative Software Inc., Madison, WI, USA) in format data (WAV extension), 22 050 Hz sampling frequency, 16-bit and stored for acoustic analysis by software PRAAT (Version 5.3.51)13, with semiautomatic extraction. Portuguese vowel inventory internal symmetry was used to compare F1 for vowel height differences and F2 for frontback vowels. Portuguese vowel inventory has an internal symmetry: the central low vowel [a], three unrounded front vowels [i, e, and ε] and three rounded back vowels [ u, o, and ɔ], and three pairs, one with two high vowels [i-u], one with two higher-mid vowels [e-o], and the last with two lower-mid vowels [ε-ɔ].14 The Institutional Review Board of the Federal University of Sergipe approved the protocol, and written informed consent was obtained from all participants. Statistics Data are expressed as mean (standard deviation) for variables with distribution normal or median (interquartile range) for variables with nonnormal distribution. MANCOVA for Multivariate analysis of covariance (as it is statistical technique) was used to analyze possible influence of confounder variables in the formant frequencies. The comparison between groups was made by t test for variables with normal distribution and Mann-Whitney test for variables with nonnormal distribution. Statistical Analysis Package for Social Sciences (SPSS, Version 18.0, IBM., Chicago, IL, USA)15 was used, and probability values <0.05 were considered. RESULTS Table 1 presents the comparison of F1, F2, F3, and F4 (Hz) for the seven oral vowels in pooled genders in IGHD and controls. MANCOVA using F1, F2, F3, and F4 of the seven vowels [a, ε, e, i, ɔ, o, and u] as dependent variables, group and gender as factors, and age as cofactor revealed that F1 differs between IGHD and controls with an effect (P ¼ 0.026) of 0.258 (partial eta squared) with an observed power of 0.835 and F3 with an effect (P ¼ 0.052) of 0.230 (partial eta squared) with an observed power of 0.763. It was revealed differences between IGHD and controls in F1 [i], with an effect (P ¼ 0.025) of 0.085 (partial eta squared) with an observed power of 0.62; F1 [ε] with an effect (P ¼ 0.012) of 0.105 (partial eta squared) with an observed power of 0.719; F1 [e] with an effect (P ¼ 0.033) of 0.077(partial eta squared) with an observed power of 0.574; F2 [u] with
Journal of Voice, Vol. -, No. -, 2015
TABLE 1. Formants Frequencies (Hz) in 33 Individuals With Isolated GH Deficiency (IGHD) and 29 Controls (Gender Pooled) Formant
Vowel
IGHD
Controls
P
F1
F1 [a] F1 [ɔ] F1 [ε] F1 [o] F1 [e] F1 [u] F1 [i] F2 [a] F2 [ɔ] F2 [ε] F2 [o] F2 [e] F2 [u] F2 [i] F3 [a] F3 [ɔ] F3 [ε] F3 [o] F3 [e] F3 [u] F3 [i] F4 [a] F4 [ɔ] F4 [ε] F4 [o] F4 [e] F4 [u] F1 [i]
886 (155) 667 (99) 664 (93) 483 (77) 499 (114) 475 (117) 446 (144) 1568 (205) 1273 (277) 1884 (495) 1011 (427) 1986 (596) 1129 (315) 2126 (613) 2900 (553) 2979 (577) 2941 (360) 2777 (419) 3002 (337) 2813 (438) 3217 (328) 3913 (428) 3935 (464) 3882 (370) 3830 (406) 3959 (477) 3912 (408) 4093 (380)
819 (155) 660 (123) 612 (101) 500 (121) 447 (79) 439 (107) 379 (97) 1517 (340) 1309 (398) 1983 (314) 1205 (667) 2089 (510) 1338 (553) 2261 (409) 2809 (349) 2765 (377) 2809 (312) 2726 (298) 2882 (267) 2856 (386) 3027 (291) 3826 (399) 3663 (396) 3972 (309) 3667 (231) 3890 (429) 3862 (351) 3913 (392)
0.096 0.787 0.040 0.525 0.045 0.214 0.032 0.480 0.682 0.345 0.636 0.471 0.079 0.309 0.436 0.060 0.130 0.578 0.129 0.690 0.020 0.413 0.014 0.304 0.056 0.557 0.613 0.072
F2
F3
F4
Notes: Data are expressed as mean (standard deviation), except for F2 [o], F3 [ɔ and e], F4 [ɔ] expressed as median (interquartile range). Data with significant differences are indicated in bold.
an effect (P ¼ 0.064) of 0.059 (partial eta squared) with an observed power of 0.459; F3 [i] with an effect (P ¼ 0.032) of 0.078 (partial eta squared) with an observed power of 0.578; F4 [i] with an effect (P ¼ 0.052) of 0.065 (partial eta squared) with an observed power of 0.497; and F4 [ɔ] with an effect (P ¼ 0.041) of 0.071(partial eta squared) with an observed power of 0.539. Tables 2 and 3 present the same comparison in women and men, respectively. Female IGHD had higher F1 [i and e] P ¼ 0.029 and P ¼ 0.036; F2[ɔ] P ¼ 0.006; and F4[ɔ] P ¼ 0.031; and lower values of F4 [ε] P ¼ 0.002 and F2 [i] P ¼ 0.004. Male IGHD showed higher values of F3 [i, e, and ε], P ¼ 0.006, P ¼ 0.022, and P ¼ 0.006, respectively and F4 [i], P ¼ 0.001 and lower value of F2 [u], P ¼ 0.034. Table 4 presents the influence of gender in values of F1, F2, F3, and F4 (Hz) for the seven oral vowels in 31 women and 31 men, separated in IGHD and controls. Male and female IGHD subjects had similar values of formant frequencies, except for F1 [a, ɔ, and ε] P < 0.0001, P ¼ 0.004, and P ¼ 0.001, respectively.
Eugenia H.O. Valenc¸a, et al
TABLE 2. Formants Frequencies (Hz) in 16 Women With Isolated GH Deficiency (IGHD) and 15 Controls Formant
Vowel
IGHD
Controls
P
F1
F1 [a] F1 [ɔ] F1 [ε] F1 [o] F1 [e] F1 [u] F1 [i] F2 [a] F2 [ɔ] F2 [ε] F2 [o] F2 [e] F2 [u] F2 [i] F3 [a] F3 [ɔ] F3 [ε] F3 [o] F3 [e] F3 [u] F3 [i] F4 [a] F4 [ɔ] F4 [ε] F4 [o] F4 [e] F4 [u] F1 [i]
987 (166) 716 (94) 715 (91) 501 (81) 486 (110) 478 (94) 454 (93) 1606 (243) 1260 (152) 1964 (707) 1093 (215) 2254 (1529) 1085 (281) 1949 (642) 2777 (723) 3073 (781) 2862 (467) 2751 (487) 3081 (370) 2870 (551) 3206 (372) 3897 (407) 3924 (364) 3805 (337) 3835 (365) 4031 (557) 4010 (440) 4105 (432)
868 (217) 679 (130) 656 (907) 490 (112) 449 (80) 423 (70) 380 (87) 1556 (153) 1155 (144) 2220 (173) 1153 (413) 2447 (300) 1174 (556) 2524 (310) 2849 (280) 2907 (235) 3079 (214) 2753 (347) 3053 (185) 2876 (300) 3102 (332) 3846 (392) 3680 (215) 4183 (270) 3693 (243) 4053 (505) 3844 (299) 4066 (461)
0.063 0.362 0.090 0.764 0.036 0.075 0.029 0.498 0.006 0.114 0.621 0.406 0.583 0.004 0.716 0.105 0.429 0.991 0.793 0.970 0.422 0.720 0.031 0.002 0.210 0.909 0.232 0.808
F2
F3
F4
3
Voice Formants in Isolated Growth Hormone Deficiency
from the pattern of linear cephalometric growth. Accordingly, we have previously reported that in these individuals, total maxillary length is the most reduced parameter (6.5 SD), followed by a cluster of six measurements: posterior cranial base length, total mandibular length, total posterior facial height, total anterior facial height, mandibular corpus length, and anterior cranial base length (4 SD).8 The less affected measurements are the lower anterior facial height and mandibular ramus height (2.5 SD).8 IGHD women had higher values of F1 [i and e], F2 [ɔ], and F4 [ɔ] and IGHD men had higher values of F3 [i, e, and ε] and F4 [i]. In this descriptive article, we can only speculate about the reasons for these findings. The higher values of F1 [i and e] in IGHD woman suggest pharyngeal constriction, as this pattern is associated with small pharyngeal cavity both in normal individuals20 and in women with severe dwarfism due to Laron syndrome (a model of dwarfism caused by resistance to—rather than lack of—GH), who also present a high-pitched voice.21 On other hand, the higher value of F2 [ɔ] seen in IGHD women suggests reduced anterior oral cavity,22 in agreement with the accentuated reduction of the length of the maxilla and mandible,8 and high f0 5 in IGHD. TABLE 3. Formants Frequencies (Hz) in 17 Men With Isolated GH Deficiency (IGHD) and 14 Controls Formant
Vowel
IGHD
Controls
P
F1
F1 [a] F1 [ɔ] F1 [ε] F1 [o] F1 [e] F1 [u] F1 [i] F2 [a] F2 [ɔ] F2 [ε] F2 [o] F2 [e] F2 [u] F2 [i] F3 [a] F3 [ɔ] F3 [ε] F3 [o] F3 [e] F3 [u] F3 [i] F4 [a] F4 [ɔ] F4 [ε] F4 [o] F4 [e] F4 [u] F4 [i]
806 (93) 621 (81) 616 (69) 467 (71) 430 (148) 442 (84) 392 (110) 1533 (61) 1284 (341) 1894 (383) 1009 (446) 1993 (417) 1171 (347) 2293 (553) 3017 (302) 2896 (262) 2956 (374) 2802 (357) 2927 (294) 2760 (304) 3227 (291) 3928 (460) 3915 (379) 3954 (395) 3749 (680) 3891 (391) 3820 (363) 4053 (286)
739 (129) 639 (117) 566 (85) 511 (134) 465 (121) 469 (207) 346 (188) 1475 (470) 1499 (495) 1803 (220) 1404 (607) 1813 (282) 1515 (511) 1979 (300) 2767 (417) 2660 (404) 2608 (245) 2697 (246) 2697 (215) 2833 (473) 2946 (224) 3806 (420) 3742 (509) 3746 (140) 3587 (216) 3717 (243) 3883 (411) 3738 (227)
0.106 0.619 0.077 0.251 0.500 0.874 0.341 0.655 0.166 0.436 0.321 0.179 0.034 0.055 0.063 0.059 0.006 0.361 0.022 0.604 0.006 0.449 0.288 0.056 0.165 0.157 0.655 0.001
Notes: Data are expressed as mean (standard deviation), except for F1 [a and e], F2 [ɔ, e, and ε], and F3 [ɔ and ε] expressed as median (interquartile range). Data with significant differences are indicated in bold.
F2
DISCUSSION Although fundamental frequency (f0) is inversely correlated to stature, formants depend on size and shape of the vocal tract and its constrictions, tongue, and lip position.16 Age and gender also influence the formant structures, with lower formant in men, followed by women and children.16 Craniofacial dimensions also contribute to vocal tract length and influence f0 and formant frequencies.17 Individuals with short face have a trend toward higher formant frequency values than those with long face.18 Because of the rarity of IGHD and its common treatment with GH therapy during childhood, there is no published formant data in adults with untreated IGHD to compare with our data. Therefore, we will discuss these data by using our cephalometric,8 voice,6 laryngeal,5 and hearing findings.19 As a group, individuals with lifetime untreated IGHD present higher than normal formant frequencies, although a significant effect was observed only in some of the Brazilian Portuguese vowels. The discrepancies in this formant pattern suggest that the different portions of the vocal tract length are not equally affected by the lack of GH, but the abnormality may depend
F3
F4
Notes: Data are expressed as mean (standard deviation), except for F1 [e, u, and i], F2 [o], and F3 [o and i] expressed as median (interquartile range). Data with significant differences are indicated in bold.
4
Journal of Voice, Vol. -, No. -, 2015
TABLE 4. Influence of Gender on F1, F2, F3, and F4 (Hz) in 33 Individuals With Isolated GH Deficiency and 29 Controls IGHD Formant [Vowel] F1 [a] F1 [ɔ] F1 [ε] F1 [o] F1 [e] F1 [u] F1 [i] F2 [a] F2 [ɔ] F2 [ε] F2 [o] F2 [e] F2 [u] F2 [i] F3 [a] F3 [ɔ] F3 [ε] F3 [o] F3 [e] F3 [u] F3 [i] F4 [a] F4 [ɔ] F4 [ε] F4 [o] F4 [e] F4 [u] F4 [i]
Controls
Male (n ¼ 17)
Female (n ¼ 16)
P
Male (n ¼ 14)
Female (n ¼ 15)
P
806 (93) 621 (81) 616 (69) 467 (71) 430 (148) 442 (84) 392 (110) 1533 (61) 1284 (341) 1894 (383) 1009 (446) 1993 (417) 1171 (347) 2293 (553) 3017 (302) 2896 (262) 2956 (374) 2802 (357) 2927 (294) 2760 (304) 3227 (291) 3928 (460) 3915 (379) 3954 (395) 3749 (680) 3891 (391) 3820 (363) 4053 (286)
987 (166) 716 (94) 715 (91) 501 (81) 486 (110) 478 (94) 454 (93) 1606 (243) 1260 (152) 1964 (707) 1093 (215) 2254 (1529) 1085 (281) 1949 (642) 2777 (723) 3073 (781) 2862 (467) 2751 (487) 3081 (370) 2870 (551) 3206 (372) 3897 (407) 3924 (364) 3805 (337) 3835 (365) 4031 (557) 4010 (440) 4105 (432)
<0.00101 0.004 0.001 0.205 0.084 0.614 0.084 0.311 0.235 0.907 0.428 0.948 0.442 0.109 0.234 0.892 0.810 0.734 0.192 0.478 0.858 0.841 0.942 0.253 0.940 0.410 0.185 0.829
739 (129) 639 (117) 566 (85) 511 (134) 465 (121) 469 (207) 346 (188) 1475 (470) 1499 (495) 1803 (220) 1404 (607) 1813 (282) 1515 (511) 1979 (300) 2767 (417) 2660 (404) 2608 (245) 2697 (246) 2697 (215) 2833 (473) 2946 (224) 3806 (420) 3742 (509) 3746 (140) 3587 (216) 3717 (243) 3883 (411) 3738 (227)
868 (217) 679 (130) 656 (907) 490 (112) 449 (80) 423 (70) 380 (87) 1556 (153) 1155 (144) 2220 (173) 1153 (413) 2447 (300) 1174 (556) 2524 (310) 2849 (280) 2907 (235) 3079 (214) 2753 (347) 3053 (185) 2876 (300) 3102 (332) 3846 (392) 3680 (215) 4183 (270) 3693 (243) 4053 (505) 3844 (299) 4066 (461)
0.005 0.403 0.014 0.662 0.930 0.416 0.979 0.548 0.018 0.001 0.177 <0.0001 0.098 <0.0001 0.537 0.127 <0.0001 0.625 <0.0001 0.771 0.152 0.793 0.666 <0.0001 0.458 0.033 0.773 0.008
Notes: Data are expressed as mean (standard deviation), except in IGHD for F1[a, e, u, and i], F2[ɔ and o], and F4[i] and in controls for F1[e], F2[ε and e], F3[ɔ and ε], and F4[o and i] expressed as median (interquartile range). Data with significant differences are indicated in bold.
Besides a negative correlation between f0 and maxillary and mandibular length,23 in normal subjects of both genders, it was also recently described the presence of a negative correlation between F1 [ɔ], F2 [a and ɔ], F3 and F4 for vowels [a, i, ɔ and u] and the maxillary and mandibular length.9 This association indicates a role of the growth of the maxilla and mandible in deepening of the formant frequencies.9 Higher values of F4 [ɔ] in women and F4 [i] in men are often found in GH-sufficient women and men with narrowed laryngeal tube.24,25 We had previously reported in this cohort (by laryngeal videostroboscopic evaluation), signs of laryngeal pharynx reflux and laryngeal constriction.5 Laryngopharyngeal reflux is the most common extraesophageal problem associated to dysphonia.26 The nearness of organs caused by the severe short stature may lead to increased risk of laryngopharyngeal reflux and laryngeal constriction, which may also contribute to the formant pattern. We postulate that the cause of higher prevalence of constriction is related to the presence of laryngopharyngeal reflux rather than to the IGHD status itself.5 Men with IGHD had higher values F3[i, e, and ε] than controls, suggesting a relationship between pharynx height
and oral cavity length,27 similar to what seen normal 8-yearold children,28 whose average stature is roughly equivalent to one of our IGHD adults.4 This voice pattern may reflect the shape of the skull17 (with reduced cephalic perimeter7), the reduced dimensions of the maxilla and mandible,8 and the reduced growth of the posterior region of the vocal tract.29 Therefore, our finding suggests a prepubertal pattern of the voice. As expected, in controls, we observed lower values in F1 [a and ε], F3 [e and ε], F2 and F4 for vowels [i, e and ε] in men than in women, the pattern of gender difference for normal Brazilian Portuguese speakers.30 IGHD abolished most of the gender differences, indicating a paramount role of IGHD on the formant structure as shown with the f0,5 which may mask the sexual dimorphism of the vocal tract.31 This work has some limitations. Our interpretation of the acoustic data is speculative. Future steps are planned to clarify some assumptions that include contrasted deglutition studies, airspace measurements, and MRI of oropharynx and related structures. Nevertheless, this article provides the first description of the formants structure in a genetic model of severe,
Eugenia H.O. Valenc¸a, et al
Voice Formants in Isolated Growth Hormone Deficiency
untreated, lifetime IGHD. It seems that IGHD induces higher formant frequency, reduces the role of gender in the frequent formant, and changes vocal tract and laryngeal configuration, findings not completely assessed in this study. Overall, these findings suggest smaller oral and pharyngeal cavities in untreated IGHD individuals, resulting in waves of faster cycles in less space, high f0, and higher formant structure. CONCLUSIONS On average, individuals with lifetime untreated IGHD present higher than normal formant frequencies, although a significant effect was observed only in some of the Brazilian Portuguese vowels. This is consistent with smaller oral and pharyngeal cavities. In addition, congenital severe IGHD causes an attenuation of the effect of gender on the structure of the formants, maintaining a prepubertal acoustic prediction. Acknowledgments The authors thank the ‘‘Associac~ao do Crescimento Fı´sico e Humano de Itabaianinha’’ for their assistance. The authors declare that there is no conflict of interest that could be perceived as influencing the impartiality of the research reported. Author contributions: Eugenia H.O. Valenc¸a was a speech therapist and field voice coordinator of this study and also first author and corresponding author of this article. Roberto Salvatori contributed toward endocrinologic aspects and was the final reviewer of this article. Maria I.R. Gonc¸alves and Susana de Carvalho were speech therapists of this study and contributed toward acoustic data and interpretation. Carla R.P. Oliveira contributed toward endocrinologic aspects, data analysis, and drafting of this article. Alaı´de H.A. Oliveira and Luiz A. Oliveira-Neto contributed toward dentistry, data analysis, and drafting of this article. Valeria M. Prado-Barreto and Jeferson S. D’Avila contributed toward otorhinolaryngology, data analysis, and drafting of this article. Valdinaldo M. Arag~ao, Anita H.O. Souza, Bruna Mateus Rocha de Andrade, and Larisse Silva Nascimento contributed toward acquisition of data and drafting of this article. Savinny Buarque de Vasconcelos Rocha and Thais Ramos Ribeiro were speech therapists of this study and contributed toward achievement of formant frequencies. Enaldo V. Mello contributed toward statistical analyses of this article. Manuel H. Aguiar-Oliveira was a senior author and local coordinator of the Itabaianinha studies and contributed toward endocrinologic aspects of this article. REFERENCES 1. Vimpani GV, Vimpani AF, Lidgard GP, Cameron EH, Farquhar JW. Prevalence of severe growth hormone deficiency. Br Med J. 1977;2:427–430. 2. Salvatori R, Hayashida CY, Aguiar-Oliveira MH, et al. Familial dwarfism due to a novel mutation of the growth hormone-releasing hormone receptor gene. J Clin Endocrinol Metab. 1999;84:917–923. 3. Aguiar-Oliveira MH, Gill MS, Barreto ES, et al. Effect of severe growth hormone (GH) deficiency due to a mutation in the GH-releasing hormone receptor on insulin-like growth factor (IGFs), IGF-binding proteins, and ternary complex formation throughout life. J Clin Endocrinol Metab. 1999;84:4118–4126. 4. Aguiar-Oliveira MH, Salvatori R. Lifetime growth hormone (GH) deficiency: impact on growth, metabolism, body composition and survival capacity. In: Preedy VR, ed. Handbook of Growth and Growth Monitoring
5.
6. 7.
8.
9.
10.
11.
12.
13.
14.
15. 16. 17.
18.
19.
20.
21.
22.
23.
24. 25.
26.
5
in Health and Disease. New York: Springer Science Business Media, LLC; 2012:2699–2710. Barreto VM, Avila JS, Sales NJ, Gonc¸alves MI, Seabra JD, Salvatori R, Aguiar-Oliveira MH. Laryngeal and vocal evaluation in untreated growth hormone deficient adults. Otolaryngol Head Neck Surg. 2009;140:37–42. Valenc¸a EH, Souza AH, Oliveira AH, et al. Voice quality in short stature with and without GH deficiency. J Voice. 2012;26:673.e13–673.e19. Oliveira HA, Salvatori R, Kraus MP, Oliveira CR, Silva PC, AguiarOliveira MH. Magnetic resonance imaging study of pituitary morphology in subjects homozygous and heterozygous for a null mutation of the GHRH receptor gene. Eur J Endocrinol. 2003;148:427–432. Oliveira-Neto LA, Melo MF, Franco AA, et al. Cephalometric features in isolated growth hormone deficiency. Angle Orthod. 2011;81: 578–583. Macari A, T Karam IA, Tabri D, Sarieddine D, Hamdan AL. Formants frequency and dispersion in relation to the length and projection of the upper and lower jaws. J Voice. 2015;28:83–90. Fant G. In: A Note on Vocal Tract Size Factors and Non-Uniform FPattern Scalings. Speech Transmission. Laboratory Quarterly Progress & Status Reports, 4. Stockholm: Royal Institute of Technology; 1975:22–30. Maryn Y, Roy N, De Bodt M, Van Cauwenberge P, Corthals P. Acoustic measurement of overall voice quality: a meta-analysis. J Acoust Soc Am. 2009;126:2619–2634. Pontes PA, Vieira VP, Goncalves MI, Pontes AL. Caracterı´sticas das vozes roucas, asperas e normais: analise acustica espectrografica comparativa. Rev Bras Otorrinolaringol. 2002;68:182–188. Boersma P, Weenink D. Praat: Doing Phonetics by Computer Version 5.3.51 1992–2013. Available at: http://www.praat.org. Accessed June 8, 2013. Seara IC, Nunez VG, Lazzarotto-Volc~ao C. Segmentos Vocalicos. In: Fonetica e fonologia do portugu^es brasileiro: 2 perı´odo. Florianopolis: LLV/ CCE/UFSC; 2011:25–63. Statistical Analysis Package for Social Inc. PASW Statistics for Windows, Version 18.0. Chicago, Illinois: SPSS Inc. 2009. Titze IR. Voice classification and-life-span changes. In: Principles of Voice Production. Englewood Cliffs: Prentice Hall; 1994:169–190. Vorperian HK, Kent RD, Gentry LR, Yandell BS. Magnetic resonance imaging procedures to study the concurrent anatomic development of vocal tract structures: preliminary results. Int J Pediatr Otorhinolaryngol. 1999; 49:197–206. Oliveira VL, Pinho SM. Qualidade da voz e o trato vocal em indivı´duos de face curta e face longa. In: Pinho SMR (org.) Topicos em Voz. Rio de Janeiro, Brasil: Guanabara-Koogan; 2001:81–88. Prado-Barreto VM, Salvatori R, Santos Junior RC, et al. Hearing status in adult individuals with lifetime, untreated isolated growth hormone deficiency. Otolaryngol Head Neck Surg. 2014;150:464–471. Laver J. Phonetic evaluation of voice quality. In: Kent RD, Ball MJ, eds. Voice Quality Measurement. San Diego: Singular Thomson Learning; 2000:37–48. Laron Z, Kopchick JJ. Sexual development in patients with Laron syndrome. In: Laron syndrome-from Man to Mouse. Berlin-Heidelberg: Springer-Verlag; 2011:104. Pinho SMR, Camargo Z. Introduc¸~ao a analise acustica da voz e da fala. In: Pinho SMR. (org.) Topicos em voz. Rio de Janeiro: Guanabara-Koogan; 2001:19–44. Macari AT, Karam IA, Tabri D, Sarieddine D, Hamdan AL. Correlation between the length and sagittal projection of the upper and lower jaw and the fundamental frequency. J Voice. 2014;28:291–296. Kent R, Read C. The Acoustic Analysis of Speech. Albany, NY: Singular/ Thomson Learning; 2002. Borden GJ, Harris KS, Raphael LJ. Speech Science Primer. In: Physiology Acoustics, and Perception of Speech. 4 ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2003:114–115. Tauber S, Gross M, Issing WJ. Association of laryngopharyngeal symptoms with gastroesophageal reflux disease. Laryngoscope. 2002;112: 879–886.
6 27. Vorperian HK, Wang S, Schimek EM, Durtschi RB, Kent RD, Gentry LR, Chungbet MK. Developmental sexual dimorphism of the oral and pharyngeal portions of the vocal tract: an imaging study. J Speech Lang Hear Res. 2011;54:995–1010. 28. Lieberman DE, McCarthy RC, Hiiemae KM, Palmer JB. Ontogeny of postnatal hyoid and larynx descent in humans. Arch Oral Biol. 2001;46:117–128. 29. Vorperian HK, Kent RD, Lindstrom MJ, Kalina CM, Gentry LR, Yandell BS. Development of vocal tract length during childhood: a
Journal of Voice, Vol. -, No. -, 2015 magnetic resonance imaging study. J Acoust Soc Am. 2005;117: 338–350. 30. Escudero P, Boersma P, Rauber AS, Bion RA. A cross-dialect acoustic description of vowels: Brazilian and European Portuguese. J Acoust Soc Am. 2009;126:1370–1379. 31. Vorperian HK, Kent RD. Vowel acoustic space development. In children: a synthesis of acoustic and anatomic data. J Speech Lang Hear Res. 2007;50: 1510–1545.
and yBaltimore, Maryland
Summary: Objective. To analyze the voice formants (F1, F2, F3, and F4 in Hz) of seven oral vowels, in Brazilian Portuguese, [a, ε, e, i, ɔ, o, and u] in adult individuals with congenital lifetime untreated isolated growth hormone deficiency (IGHD). Study Design. This is a cross-sectional study. Methods. Acoustic analysis of isolated vowels was performed in 33 individuals with IGHD, age 44.5 (17.6) years (16 women), and 29 controls, age 51.1 (17.6) years (15 women). Results. Compared with controls, IGHD men showed higher values of F3 [i, e, and ε], P ¼ 0.006, P ¼ 0.022, and P ¼ 0.006, respectively and F4 [i], P ¼ 0.001 and lower values of F2 [u], P ¼ 0.034; IGHD women presented higher values of F1 [i and e] P ¼ 0.029 and P ¼ 0.036; F2 [ɔ] P ¼ 0.006; F4 [ɔ] P ¼ 0.031 and lower values of F2 [i] P ¼ 0.004. IGHD abolished most of the gender differences in formant frequencies present in controls. Conclusions. Congenital, severe IGHD results in higher values of most formant frequencies, suggesting smaller oral and pharyngeal cavities. In addition, it causes a reduction in the effect of gender on the structure of the formants, maintaining a prepubertal acoustic prediction. Key Words: Growth hormone–Voice–Formant frequencies–Acoustic analysis. INTRODUCTION Growth hormone (GH) has fundamental roles on stature and voice, two important factors for self-confidence and social acceptation. The consequences of lack of GH on voice are difficult to study, as isolated GH deficiency (IGHD) is a rare disease occurring in 1 in 3480 to 1 in 10 000 live births.1 Furthermore, most GH deficient children in the developed world are treated with GH replacement therapy during childhood. We have identified, in Northeastern Brazil (Itabaianinha County, the State of Sergipe), a cohort of individuals with IGHD caused by a homozygous (c.57 + 1 G > A) null mutation in the GH releasing hormone receptor (GHRHR) gene.2 Most of the adults have never been treated with GH replacement. The untreated individuals have very low serum levels of GH, and of its principal mediator insulin-like growth factor-I, but normal secretion of all the other pituitary hormones.3 This results in severe short stature, associated with a typical doll-like face, and high-pitched voice.4 The voice effect seems to be more homogenous than the degree of stature deficit. Although the adult height ranges from 107 to 137 cm in pooled genders,4 the funda-
Accepted for publication March 24, 2015. From the *Department of Speech and Hearing Therapy, Federal University of Sergipe, S~ao Crist ov~ao, Sergipe, Brazil; yDivision of Endocrinology, Department of Medicine, The Johns Hopkins University, School of Medicine, Baltimore, Maryland; zDivision of Endocrinology, Department of Medicine, Federal University of Sergipe, Aracaju, Sergipe, Brazil; xDepartment of Dentistry, Federal University of Sergipe, Aracaju, Sergipe, Brazil; jjDepartment of Speech and Hearing Therapy, Federal University of S~ao Paulo, S~ao Paulo, Brazil; {Division of Otorhinolaryngology, Department of Medicine, Federal University of Sergipe, Aracaju, Sergipe, Brazil; and the #Department of Medicine, Federal University of Sergipe, Aracaju, Sergipe, Brazil. Address correspondence and reprint requests to Eugenia H.O. Valenc¸a, Division of Speech Therapy, Federal University of Sergipe, Rua Claudio Batista, S/N Bairro Sanatorio, 49060-100 Aracaju, Sergipe, Brazil. E-mail: [email protected] Journal of Voice, Vol. -, No. -, pp. 1-6 0892-1997/$36.00 Ó 2015 The Voice Foundation http://dx.doi.org/10.1016/j.jvoice.2015.03.015
mental voice frequency (f0) is high, and very similar in both genders, not only in adulthood but also in childhood and senescence.5,6 We had previously reported that these IGHD individuals also present marked reduction of cephalic perimeter7 and a typical pattern of reduction of cephalometric measurements, with marked reduction in both maxillary and mandibular length.8 In normal subjects, a significant negative association between the formant frequencies and the length of the mandible and maxilla has been reported.9 Therefore, we hypothesized that the IGHD individuals may have higher formant frequencies. In this article, we have studied the pattern of vowel formants (F) frequencies in adult IGHD individuals never treated with GH. SUBJECTS AND METHODS Subjects were recruited by word of mouth and by an ad placed in the local dwarfs’ association building, located in Itabaianinha County. Exclusion criteria were as follows: age <20 years and obvious mental or speech deficiencies. Inclusion criteria were as follows: Portuguese language native speakers, and for IGHD group, being homozygous for the GHRHR c.57 + 1 G > A mutation and having received no previous GH treatment. Inclusion criteria for controls: normal statured individuals living in the same area, with normal genotyped GHRHR alleles. We had previously reported f0 data and the perceptual-auditory analysis (Grade, Roughness, Breathiness, Asthenia, Strain scale, GRBAS scale) and video laryngostroboscopic assessment5,6 in subjects from this cohort. This cross-sectional study included 33 IGHD (16 women, 47.3 (16.6) years, 117.5 (5.6) cm and 17 men, 41.8 (18.6) years, 129.4 (5.2) cm); and 29 controls (15 women, 48.8 (13.7) years, 155.2 (6.3) cm and 14 men, 53.6 (21.5) years, 165.3 (9.0) cm).
2 F1 (affiliated to the posterior cavity, pharynx), F2 (to the anterior oral cavity), F3 (to the cavities above the vocal folds and around the inferior incisors), and F4 (to the laryngeal tube length)10 of the seven oral vowels in Brazilian Portuguese [a, ε, e, i, ɔ, o, and u] were measured (Hz) in a quiet room with <40 dB of noise. After training, 3-second sustained vowel emission was studied, using comfortable conversational pitch and loudness levels (SM58 microphone; Shure Inc., Niles, IL) kept at fixed distance of 5 cm from the subject’s mouth.11 Three trials were performed per each vowel. The criteria for trial acceptance/rejection were frequency stability in a range from 0.1 to 3 seconds.12 Samples were digitally recorded with the by software Sound Forge 10.0 (Professional audio program Sony Creative Software Inc., Madison, WI, USA) in format data (WAV extension), 22 050 Hz sampling frequency, 16-bit and stored for acoustic analysis by software PRAAT (Version 5.3.51)13, with semiautomatic extraction. Portuguese vowel inventory internal symmetry was used to compare F1 for vowel height differences and F2 for frontback vowels. Portuguese vowel inventory has an internal symmetry: the central low vowel [a], three unrounded front vowels [i, e, and ε] and three rounded back vowels [ u, o, and ɔ], and three pairs, one with two high vowels [i-u], one with two higher-mid vowels [e-o], and the last with two lower-mid vowels [ε-ɔ].14 The Institutional Review Board of the Federal University of Sergipe approved the protocol, and written informed consent was obtained from all participants. Statistics Data are expressed as mean (standard deviation) for variables with distribution normal or median (interquartile range) for variables with nonnormal distribution. MANCOVA for Multivariate analysis of covariance (as it is statistical technique) was used to analyze possible influence of confounder variables in the formant frequencies. The comparison between groups was made by t test for variables with normal distribution and Mann-Whitney test for variables with nonnormal distribution. Statistical Analysis Package for Social Sciences (SPSS, Version 18.0, IBM., Chicago, IL, USA)15 was used, and probability values <0.05 were considered. RESULTS Table 1 presents the comparison of F1, F2, F3, and F4 (Hz) for the seven oral vowels in pooled genders in IGHD and controls. MANCOVA using F1, F2, F3, and F4 of the seven vowels [a, ε, e, i, ɔ, o, and u] as dependent variables, group and gender as factors, and age as cofactor revealed that F1 differs between IGHD and controls with an effect (P ¼ 0.026) of 0.258 (partial eta squared) with an observed power of 0.835 and F3 with an effect (P ¼ 0.052) of 0.230 (partial eta squared) with an observed power of 0.763. It was revealed differences between IGHD and controls in F1 [i], with an effect (P ¼ 0.025) of 0.085 (partial eta squared) with an observed power of 0.62; F1 [ε] with an effect (P ¼ 0.012) of 0.105 (partial eta squared) with an observed power of 0.719; F1 [e] with an effect (P ¼ 0.033) of 0.077(partial eta squared) with an observed power of 0.574; F2 [u] with
Journal of Voice, Vol. -, No. -, 2015
TABLE 1. Formants Frequencies (Hz) in 33 Individuals With Isolated GH Deficiency (IGHD) and 29 Controls (Gender Pooled) Formant
Vowel
IGHD
Controls
P
F1
F1 [a] F1 [ɔ] F1 [ε] F1 [o] F1 [e] F1 [u] F1 [i] F2 [a] F2 [ɔ] F2 [ε] F2 [o] F2 [e] F2 [u] F2 [i] F3 [a] F3 [ɔ] F3 [ε] F3 [o] F3 [e] F3 [u] F3 [i] F4 [a] F4 [ɔ] F4 [ε] F4 [o] F4 [e] F4 [u] F1 [i]
886 (155) 667 (99) 664 (93) 483 (77) 499 (114) 475 (117) 446 (144) 1568 (205) 1273 (277) 1884 (495) 1011 (427) 1986 (596) 1129 (315) 2126 (613) 2900 (553) 2979 (577) 2941 (360) 2777 (419) 3002 (337) 2813 (438) 3217 (328) 3913 (428) 3935 (464) 3882 (370) 3830 (406) 3959 (477) 3912 (408) 4093 (380)
819 (155) 660 (123) 612 (101) 500 (121) 447 (79) 439 (107) 379 (97) 1517 (340) 1309 (398) 1983 (314) 1205 (667) 2089 (510) 1338 (553) 2261 (409) 2809 (349) 2765 (377) 2809 (312) 2726 (298) 2882 (267) 2856 (386) 3027 (291) 3826 (399) 3663 (396) 3972 (309) 3667 (231) 3890 (429) 3862 (351) 3913 (392)
0.096 0.787 0.040 0.525 0.045 0.214 0.032 0.480 0.682 0.345 0.636 0.471 0.079 0.309 0.436 0.060 0.130 0.578 0.129 0.690 0.020 0.413 0.014 0.304 0.056 0.557 0.613 0.072
F2
F3
F4
Notes: Data are expressed as mean (standard deviation), except for F2 [o], F3 [ɔ and e], F4 [ɔ] expressed as median (interquartile range). Data with significant differences are indicated in bold.
an effect (P ¼ 0.064) of 0.059 (partial eta squared) with an observed power of 0.459; F3 [i] with an effect (P ¼ 0.032) of 0.078 (partial eta squared) with an observed power of 0.578; F4 [i] with an effect (P ¼ 0.052) of 0.065 (partial eta squared) with an observed power of 0.497; and F4 [ɔ] with an effect (P ¼ 0.041) of 0.071(partial eta squared) with an observed power of 0.539. Tables 2 and 3 present the same comparison in women and men, respectively. Female IGHD had higher F1 [i and e] P ¼ 0.029 and P ¼ 0.036; F2[ɔ] P ¼ 0.006; and F4[ɔ] P ¼ 0.031; and lower values of F4 [ε] P ¼ 0.002 and F2 [i] P ¼ 0.004. Male IGHD showed higher values of F3 [i, e, and ε], P ¼ 0.006, P ¼ 0.022, and P ¼ 0.006, respectively and F4 [i], P ¼ 0.001 and lower value of F2 [u], P ¼ 0.034. Table 4 presents the influence of gender in values of F1, F2, F3, and F4 (Hz) for the seven oral vowels in 31 women and 31 men, separated in IGHD and controls. Male and female IGHD subjects had similar values of formant frequencies, except for F1 [a, ɔ, and ε] P < 0.0001, P ¼ 0.004, and P ¼ 0.001, respectively.
Eugenia H.O. Valenc¸a, et al
TABLE 2. Formants Frequencies (Hz) in 16 Women With Isolated GH Deficiency (IGHD) and 15 Controls Formant
Vowel
IGHD
Controls
P
F1
F1 [a] F1 [ɔ] F1 [ε] F1 [o] F1 [e] F1 [u] F1 [i] F2 [a] F2 [ɔ] F2 [ε] F2 [o] F2 [e] F2 [u] F2 [i] F3 [a] F3 [ɔ] F3 [ε] F3 [o] F3 [e] F3 [u] F3 [i] F4 [a] F4 [ɔ] F4 [ε] F4 [o] F4 [e] F4 [u] F1 [i]
987 (166) 716 (94) 715 (91) 501 (81) 486 (110) 478 (94) 454 (93) 1606 (243) 1260 (152) 1964 (707) 1093 (215) 2254 (1529) 1085 (281) 1949 (642) 2777 (723) 3073 (781) 2862 (467) 2751 (487) 3081 (370) 2870 (551) 3206 (372) 3897 (407) 3924 (364) 3805 (337) 3835 (365) 4031 (557) 4010 (440) 4105 (432)
868 (217) 679 (130) 656 (907) 490 (112) 449 (80) 423 (70) 380 (87) 1556 (153) 1155 (144) 2220 (173) 1153 (413) 2447 (300) 1174 (556) 2524 (310) 2849 (280) 2907 (235) 3079 (214) 2753 (347) 3053 (185) 2876 (300) 3102 (332) 3846 (392) 3680 (215) 4183 (270) 3693 (243) 4053 (505) 3844 (299) 4066 (461)
0.063 0.362 0.090 0.764 0.036 0.075 0.029 0.498 0.006 0.114 0.621 0.406 0.583 0.004 0.716 0.105 0.429 0.991 0.793 0.970 0.422 0.720 0.031 0.002 0.210 0.909 0.232 0.808
F2
F3
F4
3
Voice Formants in Isolated Growth Hormone Deficiency
from the pattern of linear cephalometric growth. Accordingly, we have previously reported that in these individuals, total maxillary length is the most reduced parameter (6.5 SD), followed by a cluster of six measurements: posterior cranial base length, total mandibular length, total posterior facial height, total anterior facial height, mandibular corpus length, and anterior cranial base length (4 SD).8 The less affected measurements are the lower anterior facial height and mandibular ramus height (2.5 SD).8 IGHD women had higher values of F1 [i and e], F2 [ɔ], and F4 [ɔ] and IGHD men had higher values of F3 [i, e, and ε] and F4 [i]. In this descriptive article, we can only speculate about the reasons for these findings. The higher values of F1 [i and e] in IGHD woman suggest pharyngeal constriction, as this pattern is associated with small pharyngeal cavity both in normal individuals20 and in women with severe dwarfism due to Laron syndrome (a model of dwarfism caused by resistance to—rather than lack of—GH), who also present a high-pitched voice.21 On other hand, the higher value of F2 [ɔ] seen in IGHD women suggests reduced anterior oral cavity,22 in agreement with the accentuated reduction of the length of the maxilla and mandible,8 and high f0 5 in IGHD. TABLE 3. Formants Frequencies (Hz) in 17 Men With Isolated GH Deficiency (IGHD) and 14 Controls Formant
Vowel
IGHD
Controls
P
F1
F1 [a] F1 [ɔ] F1 [ε] F1 [o] F1 [e] F1 [u] F1 [i] F2 [a] F2 [ɔ] F2 [ε] F2 [o] F2 [e] F2 [u] F2 [i] F3 [a] F3 [ɔ] F3 [ε] F3 [o] F3 [e] F3 [u] F3 [i] F4 [a] F4 [ɔ] F4 [ε] F4 [o] F4 [e] F4 [u] F4 [i]
806 (93) 621 (81) 616 (69) 467 (71) 430 (148) 442 (84) 392 (110) 1533 (61) 1284 (341) 1894 (383) 1009 (446) 1993 (417) 1171 (347) 2293 (553) 3017 (302) 2896 (262) 2956 (374) 2802 (357) 2927 (294) 2760 (304) 3227 (291) 3928 (460) 3915 (379) 3954 (395) 3749 (680) 3891 (391) 3820 (363) 4053 (286)
739 (129) 639 (117) 566 (85) 511 (134) 465 (121) 469 (207) 346 (188) 1475 (470) 1499 (495) 1803 (220) 1404 (607) 1813 (282) 1515 (511) 1979 (300) 2767 (417) 2660 (404) 2608 (245) 2697 (246) 2697 (215) 2833 (473) 2946 (224) 3806 (420) 3742 (509) 3746 (140) 3587 (216) 3717 (243) 3883 (411) 3738 (227)
0.106 0.619 0.077 0.251 0.500 0.874 0.341 0.655 0.166 0.436 0.321 0.179 0.034 0.055 0.063 0.059 0.006 0.361 0.022 0.604 0.006 0.449 0.288 0.056 0.165 0.157 0.655 0.001
Notes: Data are expressed as mean (standard deviation), except for F1 [a and e], F2 [ɔ, e, and ε], and F3 [ɔ and ε] expressed as median (interquartile range). Data with significant differences are indicated in bold.
F2
DISCUSSION Although fundamental frequency (f0) is inversely correlated to stature, formants depend on size and shape of the vocal tract and its constrictions, tongue, and lip position.16 Age and gender also influence the formant structures, with lower formant in men, followed by women and children.16 Craniofacial dimensions also contribute to vocal tract length and influence f0 and formant frequencies.17 Individuals with short face have a trend toward higher formant frequency values than those with long face.18 Because of the rarity of IGHD and its common treatment with GH therapy during childhood, there is no published formant data in adults with untreated IGHD to compare with our data. Therefore, we will discuss these data by using our cephalometric,8 voice,6 laryngeal,5 and hearing findings.19 As a group, individuals with lifetime untreated IGHD present higher than normal formant frequencies, although a significant effect was observed only in some of the Brazilian Portuguese vowels. The discrepancies in this formant pattern suggest that the different portions of the vocal tract length are not equally affected by the lack of GH, but the abnormality may depend
F3
F4
Notes: Data are expressed as mean (standard deviation), except for F1 [e, u, and i], F2 [o], and F3 [o and i] expressed as median (interquartile range). Data with significant differences are indicated in bold.
4
Journal of Voice, Vol. -, No. -, 2015
TABLE 4. Influence of Gender on F1, F2, F3, and F4 (Hz) in 33 Individuals With Isolated GH Deficiency and 29 Controls IGHD Formant [Vowel] F1 [a] F1 [ɔ] F1 [ε] F1 [o] F1 [e] F1 [u] F1 [i] F2 [a] F2 [ɔ] F2 [ε] F2 [o] F2 [e] F2 [u] F2 [i] F3 [a] F3 [ɔ] F3 [ε] F3 [o] F3 [e] F3 [u] F3 [i] F4 [a] F4 [ɔ] F4 [ε] F4 [o] F4 [e] F4 [u] F4 [i]
Controls
Male (n ¼ 17)
Female (n ¼ 16)
P
Male (n ¼ 14)
Female (n ¼ 15)
P
806 (93) 621 (81) 616 (69) 467 (71) 430 (148) 442 (84) 392 (110) 1533 (61) 1284 (341) 1894 (383) 1009 (446) 1993 (417) 1171 (347) 2293 (553) 3017 (302) 2896 (262) 2956 (374) 2802 (357) 2927 (294) 2760 (304) 3227 (291) 3928 (460) 3915 (379) 3954 (395) 3749 (680) 3891 (391) 3820 (363) 4053 (286)
987 (166) 716 (94) 715 (91) 501 (81) 486 (110) 478 (94) 454 (93) 1606 (243) 1260 (152) 1964 (707) 1093 (215) 2254 (1529) 1085 (281) 1949 (642) 2777 (723) 3073 (781) 2862 (467) 2751 (487) 3081 (370) 2870 (551) 3206 (372) 3897 (407) 3924 (364) 3805 (337) 3835 (365) 4031 (557) 4010 (440) 4105 (432)
<0.00101 0.004 0.001 0.205 0.084 0.614 0.084 0.311 0.235 0.907 0.428 0.948 0.442 0.109 0.234 0.892 0.810 0.734 0.192 0.478 0.858 0.841 0.942 0.253 0.940 0.410 0.185 0.829
739 (129) 639 (117) 566 (85) 511 (134) 465 (121) 469 (207) 346 (188) 1475 (470) 1499 (495) 1803 (220) 1404 (607) 1813 (282) 1515 (511) 1979 (300) 2767 (417) 2660 (404) 2608 (245) 2697 (246) 2697 (215) 2833 (473) 2946 (224) 3806 (420) 3742 (509) 3746 (140) 3587 (216) 3717 (243) 3883 (411) 3738 (227)
868 (217) 679 (130) 656 (907) 490 (112) 449 (80) 423 (70) 380 (87) 1556 (153) 1155 (144) 2220 (173) 1153 (413) 2447 (300) 1174 (556) 2524 (310) 2849 (280) 2907 (235) 3079 (214) 2753 (347) 3053 (185) 2876 (300) 3102 (332) 3846 (392) 3680 (215) 4183 (270) 3693 (243) 4053 (505) 3844 (299) 4066 (461)
0.005 0.403 0.014 0.662 0.930 0.416 0.979 0.548 0.018 0.001 0.177 <0.0001 0.098 <0.0001 0.537 0.127 <0.0001 0.625 <0.0001 0.771 0.152 0.793 0.666 <0.0001 0.458 0.033 0.773 0.008
Notes: Data are expressed as mean (standard deviation), except in IGHD for F1[a, e, u, and i], F2[ɔ and o], and F4[i] and in controls for F1[e], F2[ε and e], F3[ɔ and ε], and F4[o and i] expressed as median (interquartile range). Data with significant differences are indicated in bold.
Besides a negative correlation between f0 and maxillary and mandibular length,23 in normal subjects of both genders, it was also recently described the presence of a negative correlation between F1 [ɔ], F2 [a and ɔ], F3 and F4 for vowels [a, i, ɔ and u] and the maxillary and mandibular length.9 This association indicates a role of the growth of the maxilla and mandible in deepening of the formant frequencies.9 Higher values of F4 [ɔ] in women and F4 [i] in men are often found in GH-sufficient women and men with narrowed laryngeal tube.24,25 We had previously reported in this cohort (by laryngeal videostroboscopic evaluation), signs of laryngeal pharynx reflux and laryngeal constriction.5 Laryngopharyngeal reflux is the most common extraesophageal problem associated to dysphonia.26 The nearness of organs caused by the severe short stature may lead to increased risk of laryngopharyngeal reflux and laryngeal constriction, which may also contribute to the formant pattern. We postulate that the cause of higher prevalence of constriction is related to the presence of laryngopharyngeal reflux rather than to the IGHD status itself.5 Men with IGHD had higher values F3[i, e, and ε] than controls, suggesting a relationship between pharynx height
and oral cavity length,27 similar to what seen normal 8-yearold children,28 whose average stature is roughly equivalent to one of our IGHD adults.4 This voice pattern may reflect the shape of the skull17 (with reduced cephalic perimeter7), the reduced dimensions of the maxilla and mandible,8 and the reduced growth of the posterior region of the vocal tract.29 Therefore, our finding suggests a prepubertal pattern of the voice. As expected, in controls, we observed lower values in F1 [a and ε], F3 [e and ε], F2 and F4 for vowels [i, e and ε] in men than in women, the pattern of gender difference for normal Brazilian Portuguese speakers.30 IGHD abolished most of the gender differences, indicating a paramount role of IGHD on the formant structure as shown with the f0,5 which may mask the sexual dimorphism of the vocal tract.31 This work has some limitations. Our interpretation of the acoustic data is speculative. Future steps are planned to clarify some assumptions that include contrasted deglutition studies, airspace measurements, and MRI of oropharynx and related structures. Nevertheless, this article provides the first description of the formants structure in a genetic model of severe,
Eugenia H.O. Valenc¸a, et al
Voice Formants in Isolated Growth Hormone Deficiency
untreated, lifetime IGHD. It seems that IGHD induces higher formant frequency, reduces the role of gender in the frequent formant, and changes vocal tract and laryngeal configuration, findings not completely assessed in this study. Overall, these findings suggest smaller oral and pharyngeal cavities in untreated IGHD individuals, resulting in waves of faster cycles in less space, high f0, and higher formant structure. CONCLUSIONS On average, individuals with lifetime untreated IGHD present higher than normal formant frequencies, although a significant effect was observed only in some of the Brazilian Portuguese vowels. This is consistent with smaller oral and pharyngeal cavities. In addition, congenital severe IGHD causes an attenuation of the effect of gender on the structure of the formants, maintaining a prepubertal acoustic prediction. Acknowledgments The authors thank the ‘‘Associac~ao do Crescimento Fı´sico e Humano de Itabaianinha’’ for their assistance. The authors declare that there is no conflict of interest that could be perceived as influencing the impartiality of the research reported. Author contributions: Eugenia H.O. Valenc¸a was a speech therapist and field voice coordinator of this study and also first author and corresponding author of this article. Roberto Salvatori contributed toward endocrinologic aspects and was the final reviewer of this article. Maria I.R. Gonc¸alves and Susana de Carvalho were speech therapists of this study and contributed toward acoustic data and interpretation. Carla R.P. Oliveira contributed toward endocrinologic aspects, data analysis, and drafting of this article. Alaı´de H.A. Oliveira and Luiz A. Oliveira-Neto contributed toward dentistry, data analysis, and drafting of this article. Valeria M. Prado-Barreto and Jeferson S. D’Avila contributed toward otorhinolaryngology, data analysis, and drafting of this article. Valdinaldo M. Arag~ao, Anita H.O. Souza, Bruna Mateus Rocha de Andrade, and Larisse Silva Nascimento contributed toward acquisition of data and drafting of this article. Savinny Buarque de Vasconcelos Rocha and Thais Ramos Ribeiro were speech therapists of this study and contributed toward achievement of formant frequencies. Enaldo V. Mello contributed toward statistical analyses of this article. Manuel H. Aguiar-Oliveira was a senior author and local coordinator of the Itabaianinha studies and contributed toward endocrinologic aspects of this article. REFERENCES 1. Vimpani GV, Vimpani AF, Lidgard GP, Cameron EH, Farquhar JW. Prevalence of severe growth hormone deficiency. Br Med J. 1977;2:427–430. 2. Salvatori R, Hayashida CY, Aguiar-Oliveira MH, et al. Familial dwarfism due to a novel mutation of the growth hormone-releasing hormone receptor gene. J Clin Endocrinol Metab. 1999;84:917–923. 3. Aguiar-Oliveira MH, Gill MS, Barreto ES, et al. Effect of severe growth hormone (GH) deficiency due to a mutation in the GH-releasing hormone receptor on insulin-like growth factor (IGFs), IGF-binding proteins, and ternary complex formation throughout life. J Clin Endocrinol Metab. 1999;84:4118–4126. 4. Aguiar-Oliveira MH, Salvatori R. Lifetime growth hormone (GH) deficiency: impact on growth, metabolism, body composition and survival capacity. In: Preedy VR, ed. Handbook of Growth and Growth Monitoring
5.
6. 7.
8.
9.
10.
11.
12.
13.
14.
15. 16. 17.
18.
19.
20.
21.
22.
23.
24. 25.
26.
5
in Health and Disease. New York: Springer Science Business Media, LLC; 2012:2699–2710. Barreto VM, Avila JS, Sales NJ, Gonc¸alves MI, Seabra JD, Salvatori R, Aguiar-Oliveira MH. Laryngeal and vocal evaluation in untreated growth hormone deficient adults. Otolaryngol Head Neck Surg. 2009;140:37–42. Valenc¸a EH, Souza AH, Oliveira AH, et al. Voice quality in short stature with and without GH deficiency. J Voice. 2012;26:673.e13–673.e19. Oliveira HA, Salvatori R, Kraus MP, Oliveira CR, Silva PC, AguiarOliveira MH. Magnetic resonance imaging study of pituitary morphology in subjects homozygous and heterozygous for a null mutation of the GHRH receptor gene. Eur J Endocrinol. 2003;148:427–432. Oliveira-Neto LA, Melo MF, Franco AA, et al. Cephalometric features in isolated growth hormone deficiency. Angle Orthod. 2011;81: 578–583. Macari A, T Karam IA, Tabri D, Sarieddine D, Hamdan AL. Formants frequency and dispersion in relation to the length and projection of the upper and lower jaws. J Voice. 2015;28:83–90. Fant G. In: A Note on Vocal Tract Size Factors and Non-Uniform FPattern Scalings. Speech Transmission. Laboratory Quarterly Progress & Status Reports, 4. Stockholm: Royal Institute of Technology; 1975:22–30. Maryn Y, Roy N, De Bodt M, Van Cauwenberge P, Corthals P. Acoustic measurement of overall voice quality: a meta-analysis. J Acoust Soc Am. 2009;126:2619–2634. Pontes PA, Vieira VP, Goncalves MI, Pontes AL. Caracterı´sticas das vozes roucas, asperas e normais: analise acustica espectrografica comparativa. Rev Bras Otorrinolaringol. 2002;68:182–188. Boersma P, Weenink D. Praat: Doing Phonetics by Computer Version 5.3.51 1992–2013. Available at: http://www.praat.org. Accessed June 8, 2013. Seara IC, Nunez VG, Lazzarotto-Volc~ao C. Segmentos Vocalicos. In: Fonetica e fonologia do portugu^es brasileiro: 2 perı´odo. Florianopolis: LLV/ CCE/UFSC; 2011:25–63. Statistical Analysis Package for Social Inc. PASW Statistics for Windows, Version 18.0. Chicago, Illinois: SPSS Inc. 2009. Titze IR. Voice classification and-life-span changes. In: Principles of Voice Production. Englewood Cliffs: Prentice Hall; 1994:169–190. Vorperian HK, Kent RD, Gentry LR, Yandell BS. Magnetic resonance imaging procedures to study the concurrent anatomic development of vocal tract structures: preliminary results. Int J Pediatr Otorhinolaryngol. 1999; 49:197–206. Oliveira VL, Pinho SM. Qualidade da voz e o trato vocal em indivı´duos de face curta e face longa. In: Pinho SMR (org.) Topicos em Voz. Rio de Janeiro, Brasil: Guanabara-Koogan; 2001:81–88. Prado-Barreto VM, Salvatori R, Santos Junior RC, et al. Hearing status in adult individuals with lifetime, untreated isolated growth hormone deficiency. Otolaryngol Head Neck Surg. 2014;150:464–471. Laver J. Phonetic evaluation of voice quality. In: Kent RD, Ball MJ, eds. Voice Quality Measurement. San Diego: Singular Thomson Learning; 2000:37–48. Laron Z, Kopchick JJ. Sexual development in patients with Laron syndrome. In: Laron syndrome-from Man to Mouse. Berlin-Heidelberg: Springer-Verlag; 2011:104. Pinho SMR, Camargo Z. Introduc¸~ao a analise acustica da voz e da fala. In: Pinho SMR. (org.) Topicos em voz. Rio de Janeiro: Guanabara-Koogan; 2001:19–44. Macari AT, Karam IA, Tabri D, Sarieddine D, Hamdan AL. Correlation between the length and sagittal projection of the upper and lower jaw and the fundamental frequency. J Voice. 2014;28:291–296. Kent R, Read C. The Acoustic Analysis of Speech. Albany, NY: Singular/ Thomson Learning; 2002. Borden GJ, Harris KS, Raphael LJ. Speech Science Primer. In: Physiology Acoustics, and Perception of Speech. 4 ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2003:114–115. Tauber S, Gross M, Issing WJ. Association of laryngopharyngeal symptoms with gastroesophageal reflux disease. Laryngoscope. 2002;112: 879–886.
6 27. Vorperian HK, Wang S, Schimek EM, Durtschi RB, Kent RD, Gentry LR, Chungbet MK. Developmental sexual dimorphism of the oral and pharyngeal portions of the vocal tract: an imaging study. J Speech Lang Hear Res. 2011;54:995–1010. 28. Lieberman DE, McCarthy RC, Hiiemae KM, Palmer JB. Ontogeny of postnatal hyoid and larynx descent in humans. Arch Oral Biol. 2001;46:117–128. 29. Vorperian HK, Kent RD, Lindstrom MJ, Kalina CM, Gentry LR, Yandell BS. Development of vocal tract length during childhood: a
Journal of Voice, Vol. -, No. -, 2015 magnetic resonance imaging study. J Acoust Soc Am. 2005;117: 338–350. 30. Escudero P, Boersma P, Rauber AS, Bion RA. A cross-dialect acoustic description of vowels: Brazilian and European Portuguese. J Acoust Soc Am. 2009;126:1370–1379. 31. Vorperian HK, Kent RD. Vowel acoustic space development. In children: a synthesis of acoustic and anatomic data. J Speech Lang Hear Res. 2007;50: 1510–1545.