Voice Change in End-Stage Renal Disease Patients After Hemodialysis: Correlation of Subjective Hoarseness and Objective Acoustic Parameters *Soo Yeon Jung, †Jung-Hwa Ryu, *Hae Sang Park, *Sung Min Chung, †Dong-Ryeol Ryu, and *Han Su Kim, *ySeoul, Korea
Summary: Introduction. Patients with end-stage renal disease (ESRD) who are treated with hemodialysis (HD) frequently complain about hoarseness after completion of each HD session. The HD treatment affects laryngeal volume and muscle function. This study attempted to evaluate the vocal effect of HD by acoustic and aerodynamic analysis and to determine the difference between voice change group (VCG) and nonvoice change group (NVCG). Materials and Methods. A total of 55 patients (34 females and 21 males) diagnosed with ESRD and undergoing outpatient HD were enrolled. The subjects were divided into the VCG (n ¼ 13) and NVCG (n ¼ 42) by the change of the Korean Voice Handicap Index score. Patients underwent weighing and acoustic, aerodynamic analysis before and after the HD. Fundamental frequency (F0), jitter, shimmer, noise-to-harmonics ratio (NHR), pitch range, habitual pitch, voice energy, and maximal phonation time (MPT) were obtained. The pre- and post-HD data were compared using paired t test. The results were compared after dividing the total group into the VCG and NVCG categories. Correlation between the change of the weight and change of the voice analysis result was certified by Pearson correlation coefficient. Results. The F0 and habitual pitch increased in all subjects. The NHR and MPT parameters significantly decreased (P < 0.05). In the NVCG group, all the results were same as the total group. In the VCG group, the NHR result differed from the total group. All acoustic parameters showed no statistically significant differences between the two groups. There was no correlation between the weight change (%) and the change of acoustic parameter results. Conclusions. The NVCG group of patient displayed improvement in NHR, whereas the VCG group showed no change. Weight change did not significantly correlate with the voice analysis results. Key Words: Acoustic analysis–Hemodialysis–Hoarseness. INTRODUCTION Hemodialysis (HD) is the inevitable treatment procedure for the patients with end-stage renal disease (ESRD).1 Accumulation of uremic toxins, acid-base imbalance, and volume overloads are the features of ESRD. The HD treatment aims to remove the excessive fluid and toxins, and improve acid-base balance. Because voice production involves a precise coordination between the central nervous systems and peripheral phonatory organs,2 those features of ESRD can lead to change of voice owing to decreased lung function and edema of the vocal fold.3 If HD is successful, lung function is improved and the volume of the vocal fold is decreased.4,5 Most patients with ESRD do not complain of hoarseness.6 However, after 3–5 hours of HD, many patients experience general weakness, fatigue, and a voice change that lasts for a few hours. Hoarseness has been reported in 24–60% of the patients with ESRD after completion of each HD session.5 Many patients describe that their symptoms worsen in the session in which they lost more weight than usual. Studies performed on voice of the patients with ESRD5–8 compared the pre- and post-HD measurements using various Accepted for publication July 23, 2013. From the *Department of Otorhinolaryngology—Head & Neck Surgery, School of Medicine, Ewha Womans University, Seoul, Korea; and the yDepartment of Internal Medicine, School of Medicine, Ewha Womans University, Seoul, Korea Address correspondence and reprint requests to Han Su Kim, Department of Otorhinolaryngology—Head & Neck Surgery, School of Medicine, Ewha Womans University, 911-1 Mok-Dong, Yang Cheon-Gu, Seoul 158-710, Korea. E-mail: [email protected] Journal of Voice, Vol. 28, No. 2, pp. 226-230 0892-1997/$36.00 Ó 2014 The Voice Foundation http://dx.doi.org/10.1016/j.jvoice.2013.07.009
modalities. However, no study to date has clarified the objective evidence of the patients’ subjective voice symptom, or simultaneously performed a laryngoscopic examination and voice analysis. Correlation between voice and weight changes was considered in only two studies,5,7 and no significant correlation was clarified. The purpose of this study was to evaluate the vocal effects of HD by acoustic and aerodynamic analyses, and determine the correlation between the change of the voice and the body volume change. The goal was to reveal the subjective or objective causes of the post-HD voice symptoms. MATERIALS AND METHODS Subjects A total of 63 patients (40 females and 23 males) diagnosed with ESRD and who underwent outpatient HD at a tertiary training hospital (Ewha Womans University Mok-Dong Hospital, Seoul) from August to November 2011 were included in this study. History of hypertension, diabetes mellitus, cerebrovascular and neurovascular diseases, and duration of HD were recorded. A laryngoscopic examination was performed on all patients at the beginning of the study. The exclusion criteria were neuromuscular and cerebrovascular diseases, compromised general conditions, pathologic lesion on vocal folds including nodules or polyps, and laryngopharyngeal reflux evident as a reflux score exceeding seven.8 Eight patients were excluded because of cerebrovascular disease history (n ¼ 1), low extremity weakness (n ¼ 1), vocal fold lesions (n ¼ 2), and laryngopharyngeal reflux (n ¼ 4). A total of 55
Soo Yeon Jung, et al
227
Voice Change in ESRD Patients After HD
patients (34 females and 21 males) met these criteria. The study was approved by the Institutional Review Board, and written informed consent was obtained from all patients. METHODS Measuring weight change Before and after each HD session, the patients were weighed without heavy accessory or shoes, after changing into a hospital gown. The weight change was expressed in percent (%) of the pre-HD body weight. Subjective evaluation of voice change To evaluate the degree of the hoarseness, all patients were asked to fill the Korean Voice Handicap Index (KVHI) questionnaire twice before and after the HD session.10 The subjects were divided into a voice change group (VCG), defined as patients with an increase of KVHI score exceeding two, and nonvoice change group (NVCG). Acoustic analysis All patients underwent voice analysis at a voice laboratory room before and after the HD session. The voice analyses were conducted using the Computerized Speech Lab (CSL) model 4500 (KayPENTAX, Montvale, NJ). All acoustic analyses were performed with subjects in a sitting position with the microphone positioned 10 cm from their mouth. The first examination was performed using the Multidimensional Voice Program (MDVP). While recording, patients were told to voice a sustained /ah/ at a flat tone and comfortable pitch, without breaks at the same pitch for at least 4 seconds. Fundamental frequency (F0), jitter, shimmer, and noise-to-harmonics ratio (NHR) were recorded. The F0, habitual pitch, and pitch range were recorded in hertz. Jitter and shimmer were recorded in percent (%), and NHR was expressed in unique value in MDVP. The Real-Time Pitch program in CSL was used to obtain habitual pitch and voice energy. Patients read the first six sentences of the Korean standardized paragraph ‘‘Gauel (Autumn).’’ Pitch range was obtained by using the Voice Range Profile program in CSL. The patients were instructed to speak in the lowest and highest voices possible, and then to speak with progressive increase from their lowest to highest voice. The habitual pitch and pitch range were recorded in hertz, and voice energy was recorded in decibel. Aerodynamic evaluation The voice efficiency and maximal phonation time (MPT) was measured using the Phonatory Aerodynamic System model 6600 (KayPENTAX). During MPT measurements, patients were instructed to inhale as much air as they could and to pronounce the vowel /i/ at a normal pitch at the lowest intensity for as long as possible. The examination was run three times and the longest record was adapted. Analyses All pre-HD data were compared with post-HD data individually using a paired t test. Wilcoxon signed rank test was run for nonparametric values, F0, and habitual pitch in each male and
TABLE 1. Demography of Voice Change Group (VCG) and Nonvoice Change Group (NVCG) Parameters
VCG (N ¼ 13)
NVCG (N ¼ 42)
P Value
Mean age, y 55.85 ± 12.89 60.76 ± 14.41 0.191 Duration of 8.31 ± 6.05 4.62 ± 4.44 0.025* hemodialysis, y Sex, n (%) Male 4 (31) 17 (40) Female 9 (69) 25 (60) Weight change (%) 3.01 ± 2.19 3.33 ± 1.33 0.312 Abbreviation: ±, indicate standard deviation of the mean. * P < 0.050.
female groups and in the VCG and NVCG groups. Pearson correlation coefficient was used to evaluate the correlation between the weight change and voice analysis data. All statistical tests were carried out using SPSS version 18.0 (SPSS, Inc., Chicago, IL). All comparison of means tests was considered significant at P value lower than 0.05. RESULTS Demography The mean age of the patients was 59.60 years (range: 30–89 years). The mean duration of HD was 5.5 years (range: 1–18 years). Mean weight change percent during HD session was 3.25% (range: 0.1, 6.49%). Subjective voice change There were 13 (24%) and 42 (76%) patients in the VCG and the NVCG groups, respectively. Duration of HD was statistically significantly longer in the VCG group (P ¼ 0.025). There were no significant differences between the two groups in mean age and weight change (Table 1). Acoustic analysis After the HD session, acoustic analysis revealed an increase in both F0 and habitual pitch, and decreased NHR. The F0 was increased in females from 173.0351 to 191.4108 (Z ¼ 3.753a, P ¼ 0.000) and in males from 128.3633 to 148.3926 (Z ¼ 3.285a, P ¼ 0.001). The NHR parameter decreased from 0.1497 to 0.1418 (t ¼ 2.253, degrees of freedom [DF] ¼ 54, P ¼ 0.028). There were no statistically significant changes in jitter, shimmer, pitch range, and voice energy (Table 2). The VCG and NVCG groups displayed similar parameter values as the total subjects. However, the change of NHR in the VCG group was not statistically significant (Z ¼ 2.200a, P ¼ 0.861; Table 3). There was no correlation between the amount of the weight change (%) and the acoustic parameter change, which is expressed in percent of pre-HD results. The Pearson correlation coefficient of F0 and habitual pitch with a
Wilcoxon sign rinked test.
228
Journal of Voice, Vol. 28, No. 2, 2014
TABLE 2. Comparison of the Acoustic and Aerodynamic Analysis Results of Pre- and Posthemodialysis Prehemodialysis
Posthemodialysis
P Value
155.9786 ± 32.2298 173.0351 ± 21.76572 128.3633 ± 26.9751 0.1497 ± 0.0345 1.4281 ± 1.0124 4.9135 ± 2.2990 267.9060 ± 144.3938 163.1583 ± 33.0079 178.9335 ± 24.2477 137.6176 ± 29.3869 59.4776 ± 4.8013
174.9864 ± 35.8402 191.4108 ± 27.8429 148.3946 ± 31.4325 0.1418 ± 0.0377 1.3948 ± 1.2052 4.6593 ± 3.3409 261.2395 ± 188.6102 173.6923 ± 34.7538 190.7582 ± 28.1632 146.0619 ± 25.6064 59.9398 ± 5.3471
0.000* 0.000* 0.001* 0.028* 0.839 0.589 0.780 0.000* 0.001* 0.017* 0.420
13.8820 ± 5.5111
13.1354 ± 35.8402
0.030*
Parameters Acoustic parameters F0 (Hz) Female (n ¼ 34) Male (n ¼ 21) NHR Jitter (%) Shimmer (%) Pitch range (Hz) Habitual pitch (Hz) Female (n ¼ 34) Male (n ¼ 21) Voice energy (dB) Aerodynamic study MPT (s)
Abbreviations: F0, fundamental frequency; NHR, noise-to-harmonics ratio; MPT, maximal phonation time. * P < 0.050.
the weight change was 0.059 and 0.019, respectively. The P value of the correlation with habitual pitch was 0.889, and it was not statistically significant (Table 4). Aerodynamic evaluation The MPT decreased 0.7 seconds, from 13.8820 seconds to 13.1354 seconds. The decrease was statistically significant (t ¼ 2.230, DF ¼ 54, P ¼ 0.030). DISCUSSION To objectively evaluate voice function after HD, we used different methods established in prior studies. First, all subjects re-
ceived a routine laryngoscopic examination before enrollment to exclude bias from preexisting vocal fold lesions. Second, the subjects were divided into two subgroups according to the change of KVHI score, and data between these groups were compared to find objective evidence of the hoarseness. Finally, we statistically analyzed the correlation between the amount of weight change and voice analysis values. The changes of habitual pitch and F0 were consistent with a previous study.8 These results can be explained by decreased vocal fold thickness and increased subglottic pressure. A variety of factors affect vocal pitch; length, thickness, tension, and elasticity of the vocal folds are the most important factors.
TABLE 3. Comparison of the Acoustic and Aerodynamic Analysis Results of Pre- and Posthemodialysis in the Voice Change Group (VCG) and Nonvoice Change Group (NVCG) VCG (n ¼ 13) Parameters Acoustic parameters F0 (Hz) Female Male NHR Jitter (%) Shimmer (%) Pitch range (Hz) Habitual pitch (Hz) Female Male Voice energy (dB) Aerodynamic study MPT (s)
Prehemodialysis
NVCG (n ¼ 42)
Posthemodialysis P Value
Prehemodialysis
Posthemodialysis
P Value
168.3787 ± 29.0140 178.6428 ± 20.0825 145.2845 ± 35.5611 0.1412 ± 0.0232 1.6291 ± 0.8990 5.2100 ± 3.1491 319.0254 ± 116.5957 176.7708 ± 36.2215 188.9900 ± 25.0742 149.2775 ± 45.9961 60.5392 ± 4.6102
185.7872 ± 33.6569 198.1443 ± 21.6452 157.9838 ± 42.3435 0.1425 ± 0.0238 1.5225 ± 1.1346 4.1612 ± 1.5026 308.1639 ± 166.3293 189.6669 ± 38.7820 205.3144 ± 27.6690 154.4600 ± 39.8424 61.3792 ± 5.4879
0.009* 0.028* 0.144 0.861 0.600 0.600 0.552 0.011* 0.011* 0.465 0.507
152.1405 ± 32.5252 171.0164 ± 22.3814 124.3819 ± 24.1668 0.1524 ± 0.0371 1.3659 ± 1.0472 4.8217 ± 2.0055 252.0833 ± 149.6546 158.9450 ± 31.2097 175.3132 ± 23.3941 134.8741 ± 25.3144 59.1490 ± 4.8655
171.6434 ± 36.2190 188.9868 ± 29.7765 146.1384 ± 29.5029 0.1416 ± 0.0412 1.3553 ± 1.2367 4.8136 ± 3.7332 246.715 3 ± 194.5298 168.7479 ± 32.3208 185.5180 ± 26.9694 144.0859 ± 22.3649 59.4943 ± 5.2898
0.000* 0.002* 0.003* 0.018* 0.798 0.352 0.087 0.001* 0.012* 0.028* 0.569
14.1731 ± 5.1737
13.0923 ± 5.5940
0.028*
13.7919 ± 5.6686
13.1488 ± 5.8036
0.045*
Abbreviations: F0, fundamental frequency; NHR, noise-to-harmonics ratio; MPT, maximal phonation time. * P < 0.050.
Soo Yeon Jung, et al
Voice Change in ESRD Patients After HD
TABLE 4. Correlation Between the Weight Change and the Change of the Acoustic and Aerodynamic Analysis Results of Preand Posthemodialysis Parameters Acoustic parameters F0 (Hz) NHR Jitter (%) Shimmer (%) Pitch range (Hz) Habitual pitch (Hz) Voice energy (dB) Aerodynamic study MPT (s)
Pearson Correlation Coefficient
P Value
0.057 0.142 0.040 0.093 0.214 0.018 0.041
0.677 0.301 0.774 0.501 0.114 0.897 0.764
0.177
0.239
Abbreviations: F0, fundamental frequency; NHR, noise-to-harmonics ratio; MPT, maximal phonation time.
Subglottic pressure and glottal airway flow also play important roles.9 Excessive body fluid accumulates in the interstitial space in ESRD patients before HD. This excessive fluid is removed during HD; therefore, it is expected that the excessive fluid in the superficial lamina propria of vocal fold (‘‘Reinke’s space’’) will also be removed. Therefore, it might be expected that the thickness of the true vocal folds decreases after HD. Decreased vocal fold volume was previously proven by endoscopic vocal fold examination.5 During HD, excessive interstitial fluids and uremic toxins are removed from the body. These effects of HD improve muscle function, which leads to improvement of lung function after HD. Expiratory pressure, which can affect the subglottic pressure and transglottal airflow rate, is increased by the improved lung function.3 The present observation of decreased NHR after HD differs from a prior report of the insignificant change in NHR.8 The NHR reflects the laryngeal efficiency, which is related to complete glottal closure and which was shown to be related to the degree of hoarseness.12,13 This result implies that the glottal gap and hoarseness decreased after HD. This result is at variance with the conclusion of a previous study by Ori et al5 in which volume loss of the vocal folds was the suggested cause of hoarseness. However, the previous study showed only decreasing vocal fold volume, not increasing vocal fold gap.5 Despite the vocal volume loss, increasing function of the intrinsic laryngeal muscles could complete closure of the vocal fold. When patients finish each HD session, they tend to reach their goal weight, which is calculated by nephrologists to improve positive fluid balance of the body. This result also conflicts with the post-HD hoarseness. The NHR parameter is useful at assessing hoarseness.12,13 However, when we analyzed the voice analysis results in the VCG and NVCG groups, decreased NHR was evident in the NVCG group, but not in the VCG group. This result implies that VCG patients do not experience improvement of hoarseness. Difference of weight change between the two groups could explain this result.
229
However, at present, the weight change in the VCG subjects (3.016%) was lower than the weight change in the NVCG subjects (3.332%), with no statistically significant difference (P ¼ 0.312). The only significant difference between the two groups was the duration of the HD (longer duration in the VCG group). The duration of HD could be related with the fatigue. However, correlation between the duration of HD and fatigue was controversial in an earlier study.14,15 The MPT was similarly decreased after HD in the VCG and NVCG subjects. This result is opposite to the previous report of increased MPT.8 Improvement of lung function and laryngeal muscle power increased the lung capacity, which in turn increased of the amount of exhaled air during phonation. Therefore, increased MPT was expected. Despite the improved muscle power, decreasing MPT can be brought about owing to insufficient effort from patients because of their fatigue after HD.14,15 There was at present no statistical correlation between weight change and the parameters of voice analysis. This can be interpreted in two ways. First, weight change may be an inadequate parameter to reflect the decrease of the vocal fold volume. Second, other factors could affect the result voice analysis. Changes of subglottic pressure, vocal tract, lung function, and intrinsic laryngeal muscle function also have roles in phonation, and these factors can disturb the linear correlation between the weight change and voice analysis results change.11 Owing to several potential limitations, the current results proved not to be an adequate explanation for the subjective hoarseness symptoms after the HD. The first limitation is that we used the short /a/ sound uttered for less than 10 seconds as the test, in patients who could maintain sufficient expiratory pressure. However, long sentences are used in normal life; it could be difficult for the patients to maintain an adequate pressure for a long time. Second, the fatigue that can arise after HD might worsen the perception of hoarseness in patients.16 Finally, as mentioned previously, the percent of weight change might not be representative parameter for vocal fold volume change. In future studies, a more specific method reflecting the extracellular fluid should be used to evaluate the correlation between voice change and body volume change. In summary, this study was designed to reveal the effect of the HD on voice by objective parameters and to determine the reason of post-HD hoarseness in some patients. In the total group, pitch increased and NHR decreased. The NHR parameter decreased in NVCG group, whereas being insignificantly changed in VCG group. The weight change and acoustic parameters had no statistically significant correlation. Contrary to NVCG group, no improvement of voice was evident in VCG. However, except for the duration of HD, there were no definite different causative factors between the two groups. Future studies should involve other means of measuring the volume of the vocal fold and aerodynamic measurements of subglottic pressure and voice efficiency. CONCLUSIONS The presence of hoarseness in ESRD patients after HD was examined prospectively using objective measurements. The F0 and habitual pitch were elevated, and NHR and MPT were decreased.
230 The NVCG group showed improvement in NHR and the VCG group showed no change in NHR. Weight change was not significantly correlated with the results of voice analysis. REFERENCES 1. Longo DL, Fauci AL, Kasper DL, Hauser SL, Jameson JL, Loscalzo J. Harrison’s Principles of Internal Medicine. 18th ed. New York, NY: McGrawHill; 2011:2322. 2. Hirano M. Clinical Examination of Voice. 1st ed. Wien, Austria: SpringerVerlag; 1981:1–2. 3. Prezant DJ. Effect of uremia and its treatment on pulmonary function. Lung. 1990;168:1–14. 4. Rocha CBJ, Araujo S. Evaluation of maximum respiratory pressures in chronic renal patients at the pre and post hemodialysis moment. J Bras Nefrol. 2010;32:105–111 [In Portuguese]. 5. Ori Y, Sabo R, Binder Y, et al. Effect of hemodialysis on the thickness of vocal folds: a possible explanation for postdialysis hoarseness. Nephorn Clin Pract. 2006;103:144–148. 6. Kumar RB, Bhat JS. Voice in chronic renal failure. J Voice. 2010;24:690–693. 7. Fisher KV, Ligon J, Sobecks JL, Roxe DM. Phonatory effects of body fluid removal. J Speech Lang Hear Res. 2001;44:354–367.
Journal of Voice, Vol. 28, No. 2, 2014 8. Hamdan AL, Medawar W, Younes A, Bkhazi H, Fuleihan N. The effect of hemodialysis on voice: an acoustic analysis. J Voice. 2005;19: 290–295. 9. Belafsky PC, Postma GN, Koufman JA. The validity and reliability of the reflux finding score (RFS). Laryngoscope. 2001;111:1313–1317. 10. Choi DY, Choi SM, Lim GC, Nam SY. Usefulness of Voice Handicap Index in patients with hoarseness. Korean J Otolaryngol. 2001;45: 706–710. 11. Boone DR, McFarlane SC, Von Berg SL. The Voice and Voice Therapy. 7th ed. Englewood Cliffs, NJ: Allyn and Bacon; 2005:36–40. 12. Yumoto E, Gould WJ, Baer T. Harmonics-to-noise ratio as an index of the degree of hoarseness. J Acoust Soc Am. 1982;71:1544–1550. 13. Yumoto E. Quantitative evaluation of hoarseness. Arch Otolaryngol. 1983; 109:48–52. 14. Bossola M, Vulpio C, Tazza L. Fatigue in chronic dialysis patients. Semin Dial. 2011;24:550–555. 15. Jhamb M, Weisbord SD, Steel JL, Unruh M. Fatigue in patients receiving maintenance dialysis: a review of definitions, measures, and contributing factors. Am J Kidney Dis. 2008;52:353–365. 16. Nesic M, Veljkovic S, Obrenovic J, Cekic S, Velickovic D, Radenkovic M. Voice frequencies in patients treated with chronic hemodialysis. Srp Arh Celok Lek. 1996;124:99–101 [In Serbian].
Summary: Introduction. Patients with end-stage renal disease (ESRD) who are treated with hemodialysis (HD) frequently complain about hoarseness after completion of each HD session. The HD treatment affects laryngeal volume and muscle function. This study attempted to evaluate the vocal effect of HD by acoustic and aerodynamic analysis and to determine the difference between voice change group (VCG) and nonvoice change group (NVCG). Materials and Methods. A total of 55 patients (34 females and 21 males) diagnosed with ESRD and undergoing outpatient HD were enrolled. The subjects were divided into the VCG (n ¼ 13) and NVCG (n ¼ 42) by the change of the Korean Voice Handicap Index score. Patients underwent weighing and acoustic, aerodynamic analysis before and after the HD. Fundamental frequency (F0), jitter, shimmer, noise-to-harmonics ratio (NHR), pitch range, habitual pitch, voice energy, and maximal phonation time (MPT) were obtained. The pre- and post-HD data were compared using paired t test. The results were compared after dividing the total group into the VCG and NVCG categories. Correlation between the change of the weight and change of the voice analysis result was certified by Pearson correlation coefficient. Results. The F0 and habitual pitch increased in all subjects. The NHR and MPT parameters significantly decreased (P < 0.05). In the NVCG group, all the results were same as the total group. In the VCG group, the NHR result differed from the total group. All acoustic parameters showed no statistically significant differences between the two groups. There was no correlation between the weight change (%) and the change of acoustic parameter results. Conclusions. The NVCG group of patient displayed improvement in NHR, whereas the VCG group showed no change. Weight change did not significantly correlate with the voice analysis results. Key Words: Acoustic analysis–Hemodialysis–Hoarseness. INTRODUCTION Hemodialysis (HD) is the inevitable treatment procedure for the patients with end-stage renal disease (ESRD).1 Accumulation of uremic toxins, acid-base imbalance, and volume overloads are the features of ESRD. The HD treatment aims to remove the excessive fluid and toxins, and improve acid-base balance. Because voice production involves a precise coordination between the central nervous systems and peripheral phonatory organs,2 those features of ESRD can lead to change of voice owing to decreased lung function and edema of the vocal fold.3 If HD is successful, lung function is improved and the volume of the vocal fold is decreased.4,5 Most patients with ESRD do not complain of hoarseness.6 However, after 3–5 hours of HD, many patients experience general weakness, fatigue, and a voice change that lasts for a few hours. Hoarseness has been reported in 24–60% of the patients with ESRD after completion of each HD session.5 Many patients describe that their symptoms worsen in the session in which they lost more weight than usual. Studies performed on voice of the patients with ESRD5–8 compared the pre- and post-HD measurements using various Accepted for publication July 23, 2013. From the *Department of Otorhinolaryngology—Head & Neck Surgery, School of Medicine, Ewha Womans University, Seoul, Korea; and the yDepartment of Internal Medicine, School of Medicine, Ewha Womans University, Seoul, Korea Address correspondence and reprint requests to Han Su Kim, Department of Otorhinolaryngology—Head & Neck Surgery, School of Medicine, Ewha Womans University, 911-1 Mok-Dong, Yang Cheon-Gu, Seoul 158-710, Korea. E-mail: [email protected] Journal of Voice, Vol. 28, No. 2, pp. 226-230 0892-1997/$36.00 Ó 2014 The Voice Foundation http://dx.doi.org/10.1016/j.jvoice.2013.07.009
modalities. However, no study to date has clarified the objective evidence of the patients’ subjective voice symptom, or simultaneously performed a laryngoscopic examination and voice analysis. Correlation between voice and weight changes was considered in only two studies,5,7 and no significant correlation was clarified. The purpose of this study was to evaluate the vocal effects of HD by acoustic and aerodynamic analyses, and determine the correlation between the change of the voice and the body volume change. The goal was to reveal the subjective or objective causes of the post-HD voice symptoms. MATERIALS AND METHODS Subjects A total of 63 patients (40 females and 23 males) diagnosed with ESRD and who underwent outpatient HD at a tertiary training hospital (Ewha Womans University Mok-Dong Hospital, Seoul) from August to November 2011 were included in this study. History of hypertension, diabetes mellitus, cerebrovascular and neurovascular diseases, and duration of HD were recorded. A laryngoscopic examination was performed on all patients at the beginning of the study. The exclusion criteria were neuromuscular and cerebrovascular diseases, compromised general conditions, pathologic lesion on vocal folds including nodules or polyps, and laryngopharyngeal reflux evident as a reflux score exceeding seven.8 Eight patients were excluded because of cerebrovascular disease history (n ¼ 1), low extremity weakness (n ¼ 1), vocal fold lesions (n ¼ 2), and laryngopharyngeal reflux (n ¼ 4). A total of 55
Soo Yeon Jung, et al
227
Voice Change in ESRD Patients After HD
patients (34 females and 21 males) met these criteria. The study was approved by the Institutional Review Board, and written informed consent was obtained from all patients. METHODS Measuring weight change Before and after each HD session, the patients were weighed without heavy accessory or shoes, after changing into a hospital gown. The weight change was expressed in percent (%) of the pre-HD body weight. Subjective evaluation of voice change To evaluate the degree of the hoarseness, all patients were asked to fill the Korean Voice Handicap Index (KVHI) questionnaire twice before and after the HD session.10 The subjects were divided into a voice change group (VCG), defined as patients with an increase of KVHI score exceeding two, and nonvoice change group (NVCG). Acoustic analysis All patients underwent voice analysis at a voice laboratory room before and after the HD session. The voice analyses were conducted using the Computerized Speech Lab (CSL) model 4500 (KayPENTAX, Montvale, NJ). All acoustic analyses were performed with subjects in a sitting position with the microphone positioned 10 cm from their mouth. The first examination was performed using the Multidimensional Voice Program (MDVP). While recording, patients were told to voice a sustained /ah/ at a flat tone and comfortable pitch, without breaks at the same pitch for at least 4 seconds. Fundamental frequency (F0), jitter, shimmer, and noise-to-harmonics ratio (NHR) were recorded. The F0, habitual pitch, and pitch range were recorded in hertz. Jitter and shimmer were recorded in percent (%), and NHR was expressed in unique value in MDVP. The Real-Time Pitch program in CSL was used to obtain habitual pitch and voice energy. Patients read the first six sentences of the Korean standardized paragraph ‘‘Gauel (Autumn).’’ Pitch range was obtained by using the Voice Range Profile program in CSL. The patients were instructed to speak in the lowest and highest voices possible, and then to speak with progressive increase from their lowest to highest voice. The habitual pitch and pitch range were recorded in hertz, and voice energy was recorded in decibel. Aerodynamic evaluation The voice efficiency and maximal phonation time (MPT) was measured using the Phonatory Aerodynamic System model 6600 (KayPENTAX). During MPT measurements, patients were instructed to inhale as much air as they could and to pronounce the vowel /i/ at a normal pitch at the lowest intensity for as long as possible. The examination was run three times and the longest record was adapted. Analyses All pre-HD data were compared with post-HD data individually using a paired t test. Wilcoxon signed rank test was run for nonparametric values, F0, and habitual pitch in each male and
TABLE 1. Demography of Voice Change Group (VCG) and Nonvoice Change Group (NVCG) Parameters
VCG (N ¼ 13)
NVCG (N ¼ 42)
P Value
Mean age, y 55.85 ± 12.89 60.76 ± 14.41 0.191 Duration of 8.31 ± 6.05 4.62 ± 4.44 0.025* hemodialysis, y Sex, n (%) Male 4 (31) 17 (40) Female 9 (69) 25 (60) Weight change (%) 3.01 ± 2.19 3.33 ± 1.33 0.312 Abbreviation: ±, indicate standard deviation of the mean. * P < 0.050.
female groups and in the VCG and NVCG groups. Pearson correlation coefficient was used to evaluate the correlation between the weight change and voice analysis data. All statistical tests were carried out using SPSS version 18.0 (SPSS, Inc., Chicago, IL). All comparison of means tests was considered significant at P value lower than 0.05. RESULTS Demography The mean age of the patients was 59.60 years (range: 30–89 years). The mean duration of HD was 5.5 years (range: 1–18 years). Mean weight change percent during HD session was 3.25% (range: 0.1, 6.49%). Subjective voice change There were 13 (24%) and 42 (76%) patients in the VCG and the NVCG groups, respectively. Duration of HD was statistically significantly longer in the VCG group (P ¼ 0.025). There were no significant differences between the two groups in mean age and weight change (Table 1). Acoustic analysis After the HD session, acoustic analysis revealed an increase in both F0 and habitual pitch, and decreased NHR. The F0 was increased in females from 173.0351 to 191.4108 (Z ¼ 3.753a, P ¼ 0.000) and in males from 128.3633 to 148.3926 (Z ¼ 3.285a, P ¼ 0.001). The NHR parameter decreased from 0.1497 to 0.1418 (t ¼ 2.253, degrees of freedom [DF] ¼ 54, P ¼ 0.028). There were no statistically significant changes in jitter, shimmer, pitch range, and voice energy (Table 2). The VCG and NVCG groups displayed similar parameter values as the total subjects. However, the change of NHR in the VCG group was not statistically significant (Z ¼ 2.200a, P ¼ 0.861; Table 3). There was no correlation between the amount of the weight change (%) and the acoustic parameter change, which is expressed in percent of pre-HD results. The Pearson correlation coefficient of F0 and habitual pitch with a
Wilcoxon sign rinked test.
228
Journal of Voice, Vol. 28, No. 2, 2014
TABLE 2. Comparison of the Acoustic and Aerodynamic Analysis Results of Pre- and Posthemodialysis Prehemodialysis
Posthemodialysis
P Value
155.9786 ± 32.2298 173.0351 ± 21.76572 128.3633 ± 26.9751 0.1497 ± 0.0345 1.4281 ± 1.0124 4.9135 ± 2.2990 267.9060 ± 144.3938 163.1583 ± 33.0079 178.9335 ± 24.2477 137.6176 ± 29.3869 59.4776 ± 4.8013
174.9864 ± 35.8402 191.4108 ± 27.8429 148.3946 ± 31.4325 0.1418 ± 0.0377 1.3948 ± 1.2052 4.6593 ± 3.3409 261.2395 ± 188.6102 173.6923 ± 34.7538 190.7582 ± 28.1632 146.0619 ± 25.6064 59.9398 ± 5.3471
0.000* 0.000* 0.001* 0.028* 0.839 0.589 0.780 0.000* 0.001* 0.017* 0.420
13.8820 ± 5.5111
13.1354 ± 35.8402
0.030*
Parameters Acoustic parameters F0 (Hz) Female (n ¼ 34) Male (n ¼ 21) NHR Jitter (%) Shimmer (%) Pitch range (Hz) Habitual pitch (Hz) Female (n ¼ 34) Male (n ¼ 21) Voice energy (dB) Aerodynamic study MPT (s)
Abbreviations: F0, fundamental frequency; NHR, noise-to-harmonics ratio; MPT, maximal phonation time. * P < 0.050.
the weight change was 0.059 and 0.019, respectively. The P value of the correlation with habitual pitch was 0.889, and it was not statistically significant (Table 4). Aerodynamic evaluation The MPT decreased 0.7 seconds, from 13.8820 seconds to 13.1354 seconds. The decrease was statistically significant (t ¼ 2.230, DF ¼ 54, P ¼ 0.030). DISCUSSION To objectively evaluate voice function after HD, we used different methods established in prior studies. First, all subjects re-
ceived a routine laryngoscopic examination before enrollment to exclude bias from preexisting vocal fold lesions. Second, the subjects were divided into two subgroups according to the change of KVHI score, and data between these groups were compared to find objective evidence of the hoarseness. Finally, we statistically analyzed the correlation between the amount of weight change and voice analysis values. The changes of habitual pitch and F0 were consistent with a previous study.8 These results can be explained by decreased vocal fold thickness and increased subglottic pressure. A variety of factors affect vocal pitch; length, thickness, tension, and elasticity of the vocal folds are the most important factors.
TABLE 3. Comparison of the Acoustic and Aerodynamic Analysis Results of Pre- and Posthemodialysis in the Voice Change Group (VCG) and Nonvoice Change Group (NVCG) VCG (n ¼ 13) Parameters Acoustic parameters F0 (Hz) Female Male NHR Jitter (%) Shimmer (%) Pitch range (Hz) Habitual pitch (Hz) Female Male Voice energy (dB) Aerodynamic study MPT (s)
Prehemodialysis
NVCG (n ¼ 42)
Posthemodialysis P Value
Prehemodialysis
Posthemodialysis
P Value
168.3787 ± 29.0140 178.6428 ± 20.0825 145.2845 ± 35.5611 0.1412 ± 0.0232 1.6291 ± 0.8990 5.2100 ± 3.1491 319.0254 ± 116.5957 176.7708 ± 36.2215 188.9900 ± 25.0742 149.2775 ± 45.9961 60.5392 ± 4.6102
185.7872 ± 33.6569 198.1443 ± 21.6452 157.9838 ± 42.3435 0.1425 ± 0.0238 1.5225 ± 1.1346 4.1612 ± 1.5026 308.1639 ± 166.3293 189.6669 ± 38.7820 205.3144 ± 27.6690 154.4600 ± 39.8424 61.3792 ± 5.4879
0.009* 0.028* 0.144 0.861 0.600 0.600 0.552 0.011* 0.011* 0.465 0.507
152.1405 ± 32.5252 171.0164 ± 22.3814 124.3819 ± 24.1668 0.1524 ± 0.0371 1.3659 ± 1.0472 4.8217 ± 2.0055 252.0833 ± 149.6546 158.9450 ± 31.2097 175.3132 ± 23.3941 134.8741 ± 25.3144 59.1490 ± 4.8655
171.6434 ± 36.2190 188.9868 ± 29.7765 146.1384 ± 29.5029 0.1416 ± 0.0412 1.3553 ± 1.2367 4.8136 ± 3.7332 246.715 3 ± 194.5298 168.7479 ± 32.3208 185.5180 ± 26.9694 144.0859 ± 22.3649 59.4943 ± 5.2898
0.000* 0.002* 0.003* 0.018* 0.798 0.352 0.087 0.001* 0.012* 0.028* 0.569
14.1731 ± 5.1737
13.0923 ± 5.5940
0.028*
13.7919 ± 5.6686
13.1488 ± 5.8036
0.045*
Abbreviations: F0, fundamental frequency; NHR, noise-to-harmonics ratio; MPT, maximal phonation time. * P < 0.050.
Soo Yeon Jung, et al
Voice Change in ESRD Patients After HD
TABLE 4. Correlation Between the Weight Change and the Change of the Acoustic and Aerodynamic Analysis Results of Preand Posthemodialysis Parameters Acoustic parameters F0 (Hz) NHR Jitter (%) Shimmer (%) Pitch range (Hz) Habitual pitch (Hz) Voice energy (dB) Aerodynamic study MPT (s)
Pearson Correlation Coefficient
P Value
0.057 0.142 0.040 0.093 0.214 0.018 0.041
0.677 0.301 0.774 0.501 0.114 0.897 0.764
0.177
0.239
Abbreviations: F0, fundamental frequency; NHR, noise-to-harmonics ratio; MPT, maximal phonation time.
Subglottic pressure and glottal airway flow also play important roles.9 Excessive body fluid accumulates in the interstitial space in ESRD patients before HD. This excessive fluid is removed during HD; therefore, it is expected that the excessive fluid in the superficial lamina propria of vocal fold (‘‘Reinke’s space’’) will also be removed. Therefore, it might be expected that the thickness of the true vocal folds decreases after HD. Decreased vocal fold volume was previously proven by endoscopic vocal fold examination.5 During HD, excessive interstitial fluids and uremic toxins are removed from the body. These effects of HD improve muscle function, which leads to improvement of lung function after HD. Expiratory pressure, which can affect the subglottic pressure and transglottal airflow rate, is increased by the improved lung function.3 The present observation of decreased NHR after HD differs from a prior report of the insignificant change in NHR.8 The NHR reflects the laryngeal efficiency, which is related to complete glottal closure and which was shown to be related to the degree of hoarseness.12,13 This result implies that the glottal gap and hoarseness decreased after HD. This result is at variance with the conclusion of a previous study by Ori et al5 in which volume loss of the vocal folds was the suggested cause of hoarseness. However, the previous study showed only decreasing vocal fold volume, not increasing vocal fold gap.5 Despite the vocal volume loss, increasing function of the intrinsic laryngeal muscles could complete closure of the vocal fold. When patients finish each HD session, they tend to reach their goal weight, which is calculated by nephrologists to improve positive fluid balance of the body. This result also conflicts with the post-HD hoarseness. The NHR parameter is useful at assessing hoarseness.12,13 However, when we analyzed the voice analysis results in the VCG and NVCG groups, decreased NHR was evident in the NVCG group, but not in the VCG group. This result implies that VCG patients do not experience improvement of hoarseness. Difference of weight change between the two groups could explain this result.
229
However, at present, the weight change in the VCG subjects (3.016%) was lower than the weight change in the NVCG subjects (3.332%), with no statistically significant difference (P ¼ 0.312). The only significant difference between the two groups was the duration of the HD (longer duration in the VCG group). The duration of HD could be related with the fatigue. However, correlation between the duration of HD and fatigue was controversial in an earlier study.14,15 The MPT was similarly decreased after HD in the VCG and NVCG subjects. This result is opposite to the previous report of increased MPT.8 Improvement of lung function and laryngeal muscle power increased the lung capacity, which in turn increased of the amount of exhaled air during phonation. Therefore, increased MPT was expected. Despite the improved muscle power, decreasing MPT can be brought about owing to insufficient effort from patients because of their fatigue after HD.14,15 There was at present no statistical correlation between weight change and the parameters of voice analysis. This can be interpreted in two ways. First, weight change may be an inadequate parameter to reflect the decrease of the vocal fold volume. Second, other factors could affect the result voice analysis. Changes of subglottic pressure, vocal tract, lung function, and intrinsic laryngeal muscle function also have roles in phonation, and these factors can disturb the linear correlation between the weight change and voice analysis results change.11 Owing to several potential limitations, the current results proved not to be an adequate explanation for the subjective hoarseness symptoms after the HD. The first limitation is that we used the short /a/ sound uttered for less than 10 seconds as the test, in patients who could maintain sufficient expiratory pressure. However, long sentences are used in normal life; it could be difficult for the patients to maintain an adequate pressure for a long time. Second, the fatigue that can arise after HD might worsen the perception of hoarseness in patients.16 Finally, as mentioned previously, the percent of weight change might not be representative parameter for vocal fold volume change. In future studies, a more specific method reflecting the extracellular fluid should be used to evaluate the correlation between voice change and body volume change. In summary, this study was designed to reveal the effect of the HD on voice by objective parameters and to determine the reason of post-HD hoarseness in some patients. In the total group, pitch increased and NHR decreased. The NHR parameter decreased in NVCG group, whereas being insignificantly changed in VCG group. The weight change and acoustic parameters had no statistically significant correlation. Contrary to NVCG group, no improvement of voice was evident in VCG. However, except for the duration of HD, there were no definite different causative factors between the two groups. Future studies should involve other means of measuring the volume of the vocal fold and aerodynamic measurements of subglottic pressure and voice efficiency. CONCLUSIONS The presence of hoarseness in ESRD patients after HD was examined prospectively using objective measurements. The F0 and habitual pitch were elevated, and NHR and MPT were decreased.
230 The NVCG group showed improvement in NHR and the VCG group showed no change in NHR. Weight change was not significantly correlated with the results of voice analysis. REFERENCES 1. Longo DL, Fauci AL, Kasper DL, Hauser SL, Jameson JL, Loscalzo J. Harrison’s Principles of Internal Medicine. 18th ed. New York, NY: McGrawHill; 2011:2322. 2. Hirano M. Clinical Examination of Voice. 1st ed. Wien, Austria: SpringerVerlag; 1981:1–2. 3. Prezant DJ. Effect of uremia and its treatment on pulmonary function. Lung. 1990;168:1–14. 4. Rocha CBJ, Araujo S. Evaluation of maximum respiratory pressures in chronic renal patients at the pre and post hemodialysis moment. J Bras Nefrol. 2010;32:105–111 [In Portuguese]. 5. Ori Y, Sabo R, Binder Y, et al. Effect of hemodialysis on the thickness of vocal folds: a possible explanation for postdialysis hoarseness. Nephorn Clin Pract. 2006;103:144–148. 6. Kumar RB, Bhat JS. Voice in chronic renal failure. J Voice. 2010;24:690–693. 7. Fisher KV, Ligon J, Sobecks JL, Roxe DM. Phonatory effects of body fluid removal. J Speech Lang Hear Res. 2001;44:354–367.
Journal of Voice, Vol. 28, No. 2, 2014 8. Hamdan AL, Medawar W, Younes A, Bkhazi H, Fuleihan N. The effect of hemodialysis on voice: an acoustic analysis. J Voice. 2005;19: 290–295. 9. Belafsky PC, Postma GN, Koufman JA. The validity and reliability of the reflux finding score (RFS). Laryngoscope. 2001;111:1313–1317. 10. Choi DY, Choi SM, Lim GC, Nam SY. Usefulness of Voice Handicap Index in patients with hoarseness. Korean J Otolaryngol. 2001;45: 706–710. 11. Boone DR, McFarlane SC, Von Berg SL. The Voice and Voice Therapy. 7th ed. Englewood Cliffs, NJ: Allyn and Bacon; 2005:36–40. 12. Yumoto E, Gould WJ, Baer T. Harmonics-to-noise ratio as an index of the degree of hoarseness. J Acoust Soc Am. 1982;71:1544–1550. 13. Yumoto E. Quantitative evaluation of hoarseness. Arch Otolaryngol. 1983; 109:48–52. 14. Bossola M, Vulpio C, Tazza L. Fatigue in chronic dialysis patients. Semin Dial. 2011;24:550–555. 15. Jhamb M, Weisbord SD, Steel JL, Unruh M. Fatigue in patients receiving maintenance dialysis: a review of definitions, measures, and contributing factors. Am J Kidney Dis. 2008;52:353–365. 16. Nesic M, Veljkovic S, Obrenovic J, Cekic S, Velickovic D, Radenkovic M. Voice frequencies in patients treated with chronic hemodialysis. Srp Arh Celok Lek. 1996;124:99–101 [In Serbian].
