Customize your JAMA Network experience by selecting one or more topics from the list below.
Liu YC, Varier I, Ongkasuwan J. Use of Audiometric Measurement for Assessment of Vocal-Fold Function in Postextubation Infants. JAMA Otolaryngol Head Neck Surg. 2017;143(9):908–911. doi:10.1001/jamaoto.2017.0848
Is there a statically significant difference in the cry volume between normal vocal-fold function and vocal-fold movement impairment?
A cry volume of 90 dB or greater has a 90.47% sensitivity of identifying normal vocal-fold mobility, whereas a cry volume of 75 dB or less has a 90.5% specificity of identifying infants with vocal function movement impairment.
Patients with a cry volume of 90 dB or greater should be observed, and those with a cry volume less than 90 dB should undergo flexible nasolaryngoscopy.
Infants with vocal-fold motion impairment (VFMI) have an increased risk of aspiration and pulmonary complications. Flexible nasolaryngoscopy (FNL) is the gold standard for evaluation of vocal-fold mobility. Although safe, FNL causes measurable physiologic changes. Noxious stimuli, especially in neonates in the cardiovascular intensive care unit, may cause imbalance between the pulmonary and systemic circulations and potentially circulatory collapse.
To examine whether bedside measurement of infant cry volume using a smartphone application can be a screening tool for vocal-fold movement in FNL.
Design, Study, and Participants
This case-control study performed from December 1, 2013, through January 31, 2015, included 42 infants in the intensive care unit at Texas Children's Hospital, Houston.
Main Outcomes and Measures
Patient cry volume in decibels was recorded using a smartphone application placed 12 in from their mouth.
Forty-two infants were identified at the intensive care unit (median age, 33 days; 20 [48%] female and 22 [52%] male), 21 with VFMI and 21 without, based on FNL findings. A statistically significant difference was found in the mean cry volume of infants with (76.60 dB) and without (85.72 dB) VFMI. The absolute difference in the mean cry volume was 9.12 dB (95% CI, 2.74-15.50 dB). A cry volume of 90 dB or greater had a sensitivity of 90.4% (95% CI, 71%-97%) for identification of normal vocal-fold mobility. A cry volume of 75 dB or less had a specificity of 90.5% (95% CI, 71%-97%) for the identification of VFMI. The mean (SE) area under the receiver operating characteristic curve was 0.721 (0.080) (95% CI, 0.565-0.877). The cry volume, however, was not a good screen for aspiration.
Conclusions and Relevance
Bedside measurement of the cry volume with a smartphone application can be used by untrained health care professionals to screen patients for further evaluation of vocal-fold mobility using FNL.
Vocal-fold movement impairment (VFMI) can arise from injury to the recurrent laryngeal nerve or mechanical fixation of the cricoarytenoid joint. Neurologic impairment can result from various causes, including iatrogenic injury, trauma, infections, or neoplastic process, or can be idiopathic.1,2 In the neonatal and pediatric cardiovascular intensive care units, iatrogenic injury related to surgery or intubation is the most common cause of VFMI. The incidence of recurrent laryngeal nerve injury during manipulation of the aortic arch is 8.8% to 58.7%.3-9 In neonates and infants, endotracheal intubation can result in granulation, stenosis, and VFMI.10 Intubation is estimated to cause 4% to 7.5% of unilateral VFMI and 9% to 25% of bilateral VFMI.11 Infants with VFMI may have stridor, a weak cry, and aspiration or feeding problems resulting in prolonged length of stays.3,6-9,12-14
Currently, the criterion standard for the evaluation of vocal-fold movement is flexible nasolaryngoscopy (FNL). Although commonly performed, FNL has physiologic effects, including laryngospasm, desaturation, tachycardia, bradycardia, and epistaxis.15-18 These physiologic effects do not appear to be clinically significant in the adult population but can be particularly worrisome in children with congenital heart disease or pulmonary hypertension, for whom there is a fine balance between the systemic and pulmonary circulations.17,19,20 A catastrophic circulatory event has been reported secondary to airway manipulation, such as suctioning the airway in children in this population. This finding prompted the search for a screening tool to narrow the population of patients who undergo FNL to evaluate vocal-fold mobility.19,20 Experienced listeners are often able to determine vocal-fold closure (and by extension infer vocal-fold mobility) based on the volume and quality of an infant’s cry, which may then inform the pre-FNL probability of finding VFMI. This skill is often honed over years and is not always available to bedside nurses and intensivists unfamiliar with voice disorders.
Cry characteristics, such as extremely high pitched or variable cries, have been correlated with poor neural control and neurologic disorders, the prototypical example being crit-du-chat syndrome.21 Expiratory and inspiratory cry segments, as well as the long-time spectral characteristics, differ between infants with laryngomalacia and healthy children.22 Gerber et al23 performed analysis of cry data from a single child with vocal-fold paralysis, which revealed relatively flat intensity amplitude compared with healthy children. Although objective measurements for voice quality are available, such as the noise-harmonic ratio, that correlate with glottic incompetence, it is not typically feasible to perform these measurements in the intensive care unit.23 In addition, infants are not able to participate in phonatory tasks, such as maximum phonation time. The purpose of this study was to use a decibel meter application on a smartphone to measure voice volume during crying and determine whether it could be used as a screening tool for VFMI.
This case-control study, performed from December 1, 2013, through January 31, 2015, included 42 infants in the intensive care unit at Texas Children's Hospital, Houston. Audiometric measurements are routinely obtained during FNL and are available as part of the procedure note. The data were obtained during the retrospective medical record review and deidentified. Institutional review board approval was received from the Baylor College of Medicine Institutional Review Board.
The cry volume was recorded using a smartphone (HTC Desire 610) with an audiometric measurement application (Sound Meter, version 1.6, Smart Tools Co) placed 12 in from either side of patient’s mouth while the patient was in the supine position. Three subjective peak measurements (in decibels) were recorded while the patient was crying during FNL. Intensive care unit background noise was not recorded to account for background noise. Additional information recorded included age, sex, duration of intubation, associated operations, and swallow evaluation. Data analysis was performed from December 1, 2013, through January 31, 2015, using the unpaired, 2-tailed t test, Pearson correlation, and receiver operating characteristic (ROC) curves with SPSS statistical software, version 24 (IBM Inc).
Forty-two infants were identified at the intensive care unit (median age, 33 days; 20 [48%] female and 22 [52%] male), 21 with VFMI and 21 without, based on FNL findings. There was no difference in sex, age at nasolaryngoscopy examination, or duration of intubation between the VFMI and normal mobility groups. Sex, age, and duration of intubation did not correlate with any difference in cry volume. The age of the patient at the time of FNL did not correlate with the presence of VFMI. A statistically significant difference was found in the mean (SD) cry volume for infants with normal vocal-fold mobility (85.72 [8.467] dB) vs infants with VFMI (76.60 [11.58] dB) (Figure 1). The absolute difference of the mean cry volume was 9.12 dB (95% CI, 2.74-15.50 dB). Three patients (14%) with normal vocal fold mobility and 14 (67%) with VFMI had aspiration on swallow evaluation.
A cry volume of 90 dB or greater had a sensitivity of 90.4% for identification of normal vocal-fold mobility. A cry volume of 75 dB or less had a specificity of 90.5% for the identification of VFMI (Table). The mean (SE) area under the ROC was 0.721 (0.080) (95% CI, 0.565-0.877) (Figure 2). An area of 0.90 to 1 represents an excellent test result, 0.80 to 0.90 represents a good test result, 0.70 to 0.80 represents a fair test result, 0.60 to 0.70 represents a poor test result, 0.50 to 0.60 represents a failed test, and an area less than 0.50 is not statistically significant.
As a screening tool for aspiration, a cry volume of greater than 90 dB had a sensitivity of 82.4% for normal vocal-fold mobility. A volume of 70 dB had a specificity of 83.3% for VFMI (Table). The mean (SE) area under the ROC curve was 0.583 (0.096) (95% CI, 0.395-0.771) (Figure 3).
The number needed to treat (NNT) was also calculated using 75 and 90 dB as the cutoffs. Normal vocal-fold movement was considered to be the control, whereas VFMI was the event. The first group of patients was divided into those with a cry volume of 90 dB or greater and a cry volume less than 90 dB. Among patients with a cry volume of 90 dB or greater, 7 had normal mobility and 2 had VFMI. In patients with a cry volume of less than 90 dB, 14 had normal mobility and 19 had VFMI. The experimental event rate was 0.905 (19/[19 + 2]), and the control event rate was 0.667 (14/[14 + 7]). The NNT was 4.2 for the patients using 90 dB as the cutoff. The second group of patients was divided into those with a cry volume of 75 dB or greater and a cry volume less than 75 dB. Among patients with a cry volume of 75 dB or greater, 19 had normal mobility and 11 had VFMI. Among patients with a cry volume of less than 75 dB, 2 had normal mobility and 10 had VFMI. The experimental event rate was 0.476 (10/[10 + 11]), and the control event rate was 0.095 (2/[2 + 19]). The NNT was 2.62 for the patients using 75 dB as the cutoff.
Glottic incompetence related to VFMI places individuals at increased risk for aspiration and pulmonary complications. Although FNL is the criterion standard for the evaluation of VFMI, it can have significant physiologic effects and potentially adverse effects on infants with cardiopulmonary disease. In addition, the view of the infant larynx on FNL may be limited by secretions, movement, or floppy supraglottic structures. Clinicians may use other information, such as the quality of the cry and the clinical history, to aid in the assessment of vocal-fold mobility. Although formal objective audiometric measurements correlate with VFMI, these measurements are difficult to bring to the bedside in the intensive care unit. Loudness, of course, is not the only vocal parameter affected by VFMI; however, it is the most accessible to untrained listeners using a decibel meter smartphone application as a screening tool.
Our data indicate that if a patient’s peak cry decibel is 75 dB or less, the likelihood of the patient having VFMI is significant, with a specificity of 90.5%. Furthermore, if a patient’s peak cry decibel is 90 dB or greater, it is unlikely the patient has VFMI and FNL can be avoided with the high sensitivity of 90.5%. Our data are not able to stratify patients with a peak cry decibel of 75 to 90 dB. Therefore, we recommend observation for patients with a cry volume of 90 dB or greater, whereas those with a cry volume of less than 90 dB should undergo FNL after extubation to confirm or rule out VFMI.
The NNT for the 90-dB cutoff groups indicated that 4.2 patients with VFMI are needed to undergo FNL to identify 1 patient who does not need the procedure. The NNT for the 75-dB cutoff group indicated that 2.62 patients with VFMI are needed to undergo FNL to identify 1 patient who does not need the procedure. The NNT data are not significant compared with the data of sensitivity, specificity, and the area under ROC because of the small patient sample in our study.
To date, no publications in the literature have used the strength of a patient's voice measured in decibels to determine the likelihood of VFMI. The criterion standard for evaluation of VFMI is FNL, which although low risk, still necessitates a procedure and its associated risks. By using a noninvasive smartphone application as a screening tool, a number of patients with normal mobility may be able to avoid undergoing this intervention.
Unfortunately, cry volume was not a good determinant of aspiration on swallow evaluation. The 3 children with normal vocal-fold mobility and aspiration also had cry volumes greater than 80 dB. Swallowing is more complex than vocal-fold closure; thus, it is not surprising that cry volume alone cannot determine dysphagia.
The limitations of our study include a small sample size and the use of only 1 audiometric measurement application (Sound Meter). The data obtained using the Sound Meter application were not validated against other audiometric measurement applications to determine whether there are interapplication consistencies. The future goals of the study include increasing the number of patients to further refine the decibel range of patients who will need to undergo FNL for vocal-fold mobility evaluation. In addition, various applications will be tested to study the consistency of the decibel value reported.
Bedside measurement of the cry volume using a smartphone application is a new measure that can be used by untrained health care professionals to screen patients for further evaluation of vocal-fold movement using FNL. This technique could decrease the number of FNLs performed in infants in intensive care units and help allocate resources in facilities with limited access to otolaryngologists.
Corresponding Author: Julina Ongkasuwan, MD, Department of Otolaryngology–Head and Neck Surgery, Baylor College of Medicine/Texas Children’s Hospital, 6701 Fannin, Ste 640, Houston, TX 77030 (email@example.com).
Accepted for Publication: April 17, 2017.
Published Online: June 29, 2017. doi:10.1001/jamaoto.2017.0848
Author Contributions: Drs Liu and Ongkasuwan had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Ongkasuwan.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Liu.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Liu, Varier.
Administrative, technical, or material support: Liu.
Study supervision: Ongkasuwan.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.
Create a personal account or sign in to: