Nutritional Biomarkers as Predictors of Dysphonia Severity in Patients with Ischemic Stroke

Dysphonia and malnutrition are major problems in patients who have suffered an ischemic stroke. Tools to assess dysphonia severity include the dysphonia severity index (DSI) and maximum phonation time (MPT). This study aimed to investigate whether the nutritional biomarkers transferrin, albumin, and prealbumin could be predictors of dysphonia severity. A retrospective analysis was conducted between January 2018 and October 2022. A total of 180 patients who had suffered an ischemic stroke were included. Serum transferrin, albumin, and prealbumin levels were significantly correlated with DSI and MPT levels. In a multiple regression analysis, prealbumin and transferrin were significant predictors of DSI, whereas only prealbumin was a significant predictor of MPT. Serum transferrin, albumin, and prealbumin levels in patients who have suffered an ischemic stroke may correlate with dysphonia severity as assessed using DSI and MPT. These results may provide objective evidence that nutritional biomarkers affect dysphonia severity.


Introduction
Every year, there are more than 7.6 million new cases of an ischemic stroke reported all over the globe [1]. Ischemic strokes account for more than 62 percent of all strokes [1]. Ischemic strokes account for the deaths of 3.3 million individuals per year [1]. Several motor, sensory, cognitive, and communication disorders occur after ischemic stroke, and among them, dysphonia is an important problem that is easily overlooked in terms of the functional outcome of ischemic rehabilitation and the quality of life of patients. Dysphonia refers to changes in voice, such as hoarseness or pitch quality [2]. Dysphonia is common after an ischemic stroke and occurs in approximately 20% of patients [3]. Dysphonia causes communication difficulties and reduces the patient's quality of life [4]. When listening to speakers who have dysphonia, listeners need more time to process the speech inputs they receive, and speakers also make more intelligibility mistakes [5]. Dysphonia also affects rehabilitation treatment compliance in patients who have suffered an ischemic stroke and may cause poor functional outcomes [6]. Dysphonia has also been reported to be associated with swallowing dysfunction and is known as a predictor of aspiration risk [7,8]. Aspiration pneumonia is a risk factor that increases mortality in patients who have suffered an ischemic stroke [9,10]. Therefore, predicting dysphonia in ischemic stroke survivors is important. The tools used to assess dysphonia severity include the maximum phonation time (MPT) and dysphonia severity index (DSI). The MPT evaluates the function of the vocal folds by holding /a/ vocalizations for as long as possible and measuring the duration [11]. The MPT evaluation reflects the severity of dysphonia by measuring how completely the vocal folds can be kept closed while maintaining a consistent pitch and volume during vocalization [12]. The function of the vocal folds is not only to produce sound but also to protect the airway from foreign objects. DSI was created to study phonetic functions Malnutrition may exacerbate the loss of muscle mass and strength [35]. Skeletal muscle is susceptible to muscle protein dissociation in catabolic conditions, which often occur during malnutrition or severe sickness [36,37]. Malnutrition, especially in conjunction with physical inactivity, may thereby hasten the onset of muscle wasting, which can have devastating consequences [38,39]. Malnutrition is a major cause of sarcopenia which causes deterioration of laryngeal muscle movement [40]. Vocal fold muscles are also skeletal muscles and may be affected by malnutrition. Due to the comparable histological architecture of the tongue and thyroarytenoid muscles, sarcopenia may arise in the vocal fold muscles [41,42]. Weakness and wasting of the vocal fold muscles can cause changes in voice quality [41]. In addition, sarcopenia of the respiratory muscles may cause changes in voice quality and dysphonia due to a decrease in expiratory function [43]. However, there are no studies on the association between serum transferrin, albumin, and prealbumin and dysphonia severity in patients who have suffered an ischemic stroke. Therefore, this study aimed to investigate whether serum transferrin, albumin, and prealbumin could be predictors of dysphonia severity.

Participants
A retrospective analysis was conducted between January 2018 and October 2022 on patients admitted an with ischemic stroke at the Kyung Hee University Hospital in Gangdong. Patients who were evaluated for serum transferrin, albumin, and prealbumin and who had a voice evaluation were recruited. Patients who had suffered an ischemic stroke for the first time and those who completed the evaluation within 1 month of onset were included. Patients with an acute inflammatory illness, a history of thyroid disease, nephrotic syndrome, liver disease, metabolic disease, or inflammatory bowel disease, and a history of steroid medication that might impact transferrin, albumin, and prealbumin levels were excluded. Patients with other brain diseases that could affect dysphonia severity, such as a brain tumor and traumatic brain injury, were also excluded. Patients with a cognitive impairment that made voice assessment difficult, aphasia, or tracheostomy were excluded. Patients with a history of chronic laryngitis, laryngopharyngeal reflux, polyps and nodules on the vocal fold, laryngeal cancer, Parkinson's disease, and motor neuron diseases were excluded as these conditions could also cause dysphonia. Patients with a history of iatrogenic injury to the recurrent laryngeal or vagus nerves by thyroid and parathyroid surgery, carotid endarterectomy, or cardiothoracic procedures were also excluded ( Figure 1). Dysphonia severity was compared between the normal value and low nutritional biomarker groups. A transferrin serum level of ≤200 mg/dL was designated as the low transferrin group [44], an albumin serum level of ≤3.5 g/dL was designated as the low albumin group [45], and a prealbumin serum level of ≤20 mg/dL was designated as the low prealbumin group [46]. The study was approved by the Institutional Review Board (IRB) of Kyung Hee University Hospital in Gangdong, Korea (IRB approval number: 2022-12-003).

Nutritional Biomarkers
The participants' blood samples were collected, and nutritional biomarkers were measured. Serum levels of transferrin, albumin, and prealbumin were measured using AU5800 automatic chemistry analyzers (Beckman Coulter, Brea, CA, USA).

Maximum Phonation Time
Participants in the research vocalized the /a/ sound for as long as feasible after achieving maximal inspiration. The participants were advised to speak at a usual conversational volume level. The duration during which the vocalization was maintained with a steady volume and the audible sound was recorded. The individuals were instructed to sit in a relaxed posture, and the examination was administered three times. The maximum result of three measurements was recorded as MPT and measured in seconds (s) [12].

Nutritional Biomarkers
The participants' blood samples were collected, and nutritional biomarkers were measured. Serum levels of transferrin, albumin, and prealbumin were measured using AU5800 automatic chemistry analyzers (Beckman Coulter, Brea, CA, USA).

Maximum Phonation Time
Participants in the research vocalized the /a/ sound for as long as feasible after achieving maximal inspiration. The participants were advised to speak at a usual conversational volume level. The duration during which the vocalization was maintained with a steady volume and the audible sound was recorded. The individuals were instructed to sit in a relaxed posture, and the examination was administered three times. The maximum result of three measurements was recorded as MPT and measured in seconds (s) [12].

Dysphonia Severity Index
The sustained /a/ was evaluated for at least 3 s using the Model 3950 multidimensional voice software (Kay Pentax, Montvale, NJ, USA). Acoustic voice signals were evaluated for their F0-High, I-Low, jitter, and MPT. Individuals were advised to continue speaking the /a/ sound at a comfortable volume and pitch for 3 s to measure the jitter captured at 5 kHz [47,48]. To determine F0-High, /a/ was spoken at a standard pitch, which was elevated to the maximum achievable pitch [47]. After measuring I-Low at the participants' usual pitch, they were told to gradually reduce the intensity for 5 s until they were whispering [47].
Positive DSI values indicate better vocal function, and negative DSI values indicate more severe dysphonia [48]. Afterward, the DSI was computed using the following formula: [48] (1)

Statistical Analysis
Using the Statistical Package for the Social Sciences version 20.0 for Windows (IBM Corp., Armonk, NY, USA), variables were statistically evaluated (IBM Corp., Armonk, NY, USA). The Kolmogorov-Smirnov and Levene tests were conducted to assess the normality of the data distribution and the homogeneity of the variance, respectively. Continuous variables were analyzed using the independent t-test, and categorical variables were analyzed using the chi-square test. Pearson's correlation coefficient was used to analyze the relationship between serum transferrin, albumin, and prealbumin levels and MPT and DSI. Using multiple linear regression analysis with stepwise selection for sociodemographic characteristics, lifestyle characteristics, comorbidities, mean mini-mental state examination (MMSE) and modified Barthel index (MBI), the influence of serum transferrin,

Dysphonia Severity Index
The sustained /a/ was evaluated for at least 3 s using the Model 3950 multidimensional voice software (Kay Pentax, Montvale, NJ, USA). Acoustic voice signals were evaluated for their F0-High, I-Low, jitter, and MPT. Individuals were advised to continue speaking the /a/ sound at a comfortable volume and pitch for 3 s to measure the jitter captured at 5 kHz [47,48]. To determine F0-High, /a/ was spoken at a standard pitch, which was elevated to the maximum achievable pitch [47]. After measuring I-Low at the participants' usual pitch, they were told to gradually reduce the intensity for 5 s until they were whispering [47].

Statistical Analysis
Using the Statistical Package for the Social Sciences version 20.0 for Windows (IBM Corp., Armonk, NY, USA), variables were statistically evaluated (IBM Corp., Armonk, NY, USA). The Kolmogorov-Smirnov and Levene tests were conducted to assess the normality of the data distribution and the homogeneity of the variance, respectively. Continuous variables were analyzed using the independent t-test, and categorical variables were analyzed using the chi-square test. Pearson's correlation coefficient was used to analyze the relationship between serum transferrin, albumin, and prealbumin levels and MPT and DSI. Using multiple linear regression analysis with stepwise selection for sociodemographic characteristics, lifestyle characteristics, comorbidities, mean mini-mental state examination (MMSE) and modified Barthel index (MBI), the influence of serum transferrin, albumin, and prealbumin levels on the DSI and MPT was determined. In all statistical tests, a p-value less than 0.05 was deemed statistically significant.

Demographic Characteristics of the Study Participants
The sociodemographic characteristics, lifestyle characteristics, brain lesion location, comorbidities, nutritional biomarkers, dysphonia severity are shown in Table 1

Comparison of Characteristics between Low Transferrin Group and Normal Value Group
Compared to the normal group, DSI (−1.66 ± 1.77, p = 0.002) and MPT (9.95 ± 4.21 s, p = 0.043) were statistically significantly lower in the low transferrin group. There were no statistically significant differences in age, sex, BMI, smoking history, regular alcohol use history, brain lesion location, MMSE score, or MBI. (Table 2).

Comparison of Characteristics between the Low Albumin Group and the Normal Value Group
Compared to the normal group, DSI (−1.51 ± 0.93, p = 0.013) and MPT (10.88 ± 3.55 s, p = 0.017) were statistically significantly lower in the low albumin group. The proportion of patients with a smoking history was significantly higher in the low albumin group (p = 0.047). There were no statistically significant differences in age, sex, BMI, alcohol use history, brain lesion location, MMSE score, or MBI. (Table 3).

Comparison of Characteristics between the Low Prealbumin Group and the Normal Value Group
Compared to the normal group, DSI (−1.61 ± 1.57, p < 0.001) and MPT (10.17 ± 3.75 s, p = 0.004) were statistically significantly lower in the low prealbumin group. There were no statistically significant differences in age, sex, BMI, smoking history, regular alcohol use history, brain lesion location, MMSE score, or MBI. (Table 4).

Discussion
To the best of our knowledge, this study is the first to evaluate the association between DSI and MPT and nutritional biomarkers measured in patients who had suffered an ischemic stroke. The serum levels of the nutritional biomarkers were divided into a low nutritional biomarker group and a normal value group as categorical variables, and a comparison was made between the two groups. As a result, MPT and DSI were significantly lower in the group with low nutritional biomarkers in all nutritional biomarkers. Serum transferrin, albumin, and prealbumin levels, which are used to determine nutritional status, were linked with MPT and DSI in this study. In addition, in a multiple linear regression analysis with stepwise selection, transferrin and prealbumin appeared to be significant predictors of DSI, and prealbumin appeared to be a significant predictor of MPT. These results are consistent with previous findings that prealbumin has a shorter half-life than albumin and is a sensitive indicator that reflects short-term impairment of energy intake and protein status in a more timely manner [49,50].
After an ischemic stroke, many patients develop dysphonia [3,4]. Patients with dysphonia complain of hoarseness, poor voice clarity, and discomfort when speaking. The strength and function of the laryngeal muscles that control the vocal folds are important in the vocalization process. The laryngeal muscle produces phonation, the sound energy created by the vocal fold vibration. However, the laryngeal muscle is usually paralyzed in patients who have suffered a stroke, resulting in reduced vocal fold movement and, therefore, vocalization problems [51,52]. Weakness of the laryngeal muscles is strongly associated with vocal fold instability [53,54]. Instability of the vocal folds may result in decreased phonetic function, as well as decreased coughing function, respiratory function, and swallowing function [55]. A study using laryngopharyngeal neuromuscular electrical stimulation showed improvement in dysphonia by increasing laryngeal elevation and improving excessive quaver through the strengthening of mylohyoid and thyrohyoid muscles [56]. Dysphonia is caused by weakness of the muscles controlling the vocal folds, and it is important to predict and evaluate the severity of dysphonia.
Since malnutrition is a major cause of sarcopenia and frailty syndrome, it is important to properly evaluate nutritional status. Transferrin, albumin, and prealbumin are widely used biomarkers that indicate nutritional status [23,57,58]. Malnutrition is an important causative factor for muscle loss [56]. Other studies have reported the role of transferrin in regulating neural regeneration or muscular atrophy [59]. There is also a report which found that low albumin levels are independently associated with a decrease in muscle strength [60]. Positive correlations exist between serum albumin levels, gait speed, and handgrip strength [61]. It has been found that higher serum albumin concentrations are associated with protective effects against skeletal muscle atrophy, decreased gait speed, and the occurrence of muscle wasting [62]. A study reported that prealbumin is more related to and better reflected in muscle mass loss than albumin [62]. The relationship between prealbumin levels and sarcopenia prevalence suggests that higher prealbumin levels may prevent older individuals from developing sarcopenia [62]. Sarcopenia may affect the muscles throughout the whole body, and it can also manifest itself in the form of respiratory sarcopenia and sarcopenic dysphagia [63,64]. The sarcopenia that might develop in these pharyngolaryngeal muscles can have an effect on the function of the voice. Since vocal fold muscles are skeletal muscles, malnutrition effects on the muscles may contribute to dysphonia severity.
MPT and DSI are objective tools for assessing dysphonia severity [48]. The results of this study found an association between MPT and DSI and nutritional biomarkers measured by prealbumin, albumin, and transferrin. This study suggests that nutritional biomarkers may be helpful as predictors of dysphonia severity in patients who had suffered an ischemic stroke. In addition, prealbumin was a significant predictor for both MPT and DSI. Although evaluating nutritional status as a single biomarker is not acceptable, it may be evidence that measuring prealbumin rather than measuring albumin alone may be helpful. It is required to review nutritional biomarkers that fit the demands of the evaluation since each nutritional biomarker has a varied half-life and a different index to reflect.
There were several limitations to this study. First, it was a retrospective cross-sectional study. Second, the proportion of patients with a smoking history was higher in the low albumin group than in the normal value group. Smoking history may have an impact on dysphonia and may have influenced the outcome [65]. Thus, for more reliable results, additional studies where smoking history variables are controlled are needed. Third, there may have been a selection bias by including only patients whose nutritional biomarkers were measured and who were evaluated for dysphonia. Fourth, there was a lack of information on whether malnutrition existed before an ischemic stroke occurred. Finally, although statistically significant results were obtained in the association analysis, the correlation coefficient value was low. Since it is somewhat difficult to conclude that there is a high correlation using only the results of this study, a large-scale prospective longitudinal study is needed in future.

Conclusions
In conclusion, serum transferrin, albumin, and prealbumin levels in patients with ischemic stroke may correlate with dysphonia severity as assessed using DSI and MPT. These results may provide objective evidence that nutritional biomarkers affect dysphonia severity.

Informed Consent Statement:
The need for written informed consent was waived due to the retrospective study design.

Data Availability Statement:
The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.