Voice Related Quality of Life in Pediatric Patients with a History of Prematurity

Abstract

Objective

To determine incidence of dysphonia in patients with history of prematurity and evaluate the correlation between dysphonia and risk factors unique to premature infants. The aim of this study is to determine parent-perceived vocal quality in patients with history of prematurity and whether duration of intubation, number of intubations, and incidence of patent ductus arteriosus repair were correlated with these perceptions.

Methods

Cohort study of premature patients presenting to outpatient clinics from January 2010 to January 2013 in tertiary care center. Patients gestational age ≤ 37 weeks at birth without history of tracheostomy or known vocal fold pathology were eligible. A volunteer sample was obtained from patients presenting in Otolaryngology clinics from January 2010 through January 2013 whose parents agreed to complete surveys. Outcomes were assessed via parental completion of pediatric voice outcomes score (pVOS) and pediatric voice-related quality of life (pVRQOL) instruments. The primary outcome assessed was the incidence of dysphonia in infants with a history of prematurity without known vocal pathology. Additionally, patient factors associated with dysphonia were evaluated. The hypothesis tested was formulated prior to data collection.

Results

Sixty-nine participants were included. Mean age at follow-up was 28 (3–197) months. Mean gestational age was 29 (23–37) weeks. Mean intubation duration was 3 (0–22) weeks and median number of intubations was 1 (range 0–5). Voice outcome scores varied widely with pVRQOL scores demonstrating a mean of 89.2 ±18.1 (25–100) and pVOS with a mean of 11.4±2.2 (0–13). Univariate analysis utilized Spearman correlation coefficients for continuous variables and Wilcoxon Two-sample test for categorical groups. Significance was set at p<0.05. All significant univariate associations were placed in a multivariate model. Duration of intubation ≥4 weeks was the only factor which correlated with dysphonia on multivariate analysis (p=0.0028, OR=6.4, 95% CI=1.9–21.6).

Conclusions

The data suggest that prolonged intubation is associated with poorer long term parent-perceived voice quality in premature patients. Further study is required to correlate parent perceptions with objective vocal quality data and physical findings of vocal pathology. These data may increase the clinician’s suspicion for and evaluation of dysphonia in this population.

Introduction

In the United States, over half a million children are born prematurely each year¹. The incidence of prematurity has steadily increased by 36% since the early 1980’s¹. With the increase in prematurity, there have been advances in the care of premature infants and improved long-term survival. Due to myriad medical problems including pulmonary disease, premature infants are at risk for respiratory failure requiring prolonged intubation and mechanical ventilation. Additionally, this group may also require multiple surgical interventions or suffer episodic respiratory distress requiring multiple intubations early in life. The risk to the pediatric airway with prolonged and multiple intubations is well known and significant progress in the prevention of airway sequelae has been made in the last 3 decades^2–10.

As these previously premature patients mature and develop, they face significant physical and psychosocial challenges^11–14. Of these challenges, recent reports have also begun to cite voice disturbances in premature patients^{15, 16}. As previous literature has supported a link between voice, quality of life, and even occupation choice¹⁷, the impact of vocal quality, though of less urgency in the setting of multiple complex medical concerns that can be associated with prematurity, does deserve consideration due to its potential long-term impact.

Several instruments have been designed in the past decade specifically to evaluate outcomes in voice related problems and treatments. The most utilized and best validated surveys are the Voice Handicap Index (VHI)¹⁸, the Voice Outcomes Survey (VOS)¹⁹, and the Voice-Related Quality of Life (VRQOL)²⁰. These instruments were designed to be used in the care of laryngology patients and were validated in adult populations. Since their dissemination, each of these instruments have spawned a pediatric analogue, namely the pediatric VHI (pVHI)²¹, pediatric VOS (pVOS)²², and pediatric VRQOL (pVRQOL)²³. Validation of both the pVRQOL²³ and the pVOS ^{22, 23} have been performed in the pediatric population. Normative data from a healthy population has also been published²⁴.

Only one previous investigation has utilized any quality of life instrument to investigate voice outcomes in pediatric patients with a history of prematurity²⁵, identifying a link between gestational age at birth < 25 weeks and moderate to severe voice disturbance in childhood. In this population where patients can spend weeks to months on respiratory support, the larynx is susceptible to injury from prolonged intubation, multiple intubations, or prolonged non-invasive ventilation. French et al identified female gender and increased number of intubations as significant risk factors for moderate to severe dysphonia in the premature patient²⁵.

The present study seeks to identify the association between airway management and other potentially confounding factors and comorbid conditions on voice outcomes in the pediatric population with a history of prematurity by surveying parents of patients with a history of prematurity using the pVRQOL and the pVOS instruments.

Patients and Methods

This study was a prospective observational study with retrospective chart analysis. Patients were prospectively accrued and surveys were obtained with subsequent retrospective chart review for additional data. Institutional Review Board approval was obtained from the Nationwide Children’s Hospital IRB (IRB10-00161).

Inclusion Criteria

Patients with a history of prematurity, defined as gestational age at birth of 37 weeks or less, who presented for routine care to the Neonatology or Otolaryngology outpatient clinic for routine care between January 2010 and January 2013 were eligible for inclusion. Participation was voluntary, so a consecutive sample of all patients was not able to be obtained. Parents of patients meeting these criteria were informed of the observational nature of the study and given the pVOS and pVRQOL instruments. Patients meeting criteria with returned surveys and consent for inclusion were ultimately included in the analysis.

Exclusion Criteria

Patients without consent for inclusion or who were not born prematurely were excluded from the analysis. Additionally, patients with a history of tracheostomy or known vocal pathology and those presenting with voice complaints were excluded.

Data analysis

After accruing pVOS and pVRQOL scores as well as demographic data for each patient, statistical analysis was performed.

Univariate analysis

The association between pVOS, pVRQOL, and patient demographic and clinical factors was assessed. Spearman correlations were used to identify associations between outcomes and continuous variables, and Wilcoxon Two-sample test was used to compare pVOS and pVRQOL scores between categorical groups. P value < 0.05 was considered significant. All tests were conducted in SAS 9.3 (by SAS Institute Inc., Cary, NC, USA).

Multivariate analysis

After identifying associated factors by univariate analysis, multivariate analysis was then undertaken to identify factors with independent association to pVOS and pVRQOL. The pVOS and pVRQOL scores were not normally distributed, and there was no improvement in the distribution with logarithmic or square root transformation, so linear regression was not able to be applied. As such, the pVOS and pVRQOL outcomes were converted to binary data reflecting a score either greater than or less than the median score (12 for pVOS, 100 for pVRQOL). Two separate models of multivariable logistic regression were then used to model the two separate outcome scores: pVOS and pVRQOL. Independent variables were chosen from previous univariate analysis with p value <0.05. Backward stepwise selection method was then used to select the best model. P value < 0.05 was considered significant. All tests were conducted using SAS 9.3 (by SAS Institute Inc., Cary, NC, USA).

Results

Sixty-nine participants were identified meeting inclusion criteria and were included in the analysis. Mean age at follow up was 28 months (range 3–197 months). Mean gestational age was 29 weeks (range 23–37 weeks). Mean length of intubation was 3 weeks (range 0–22 weeks) and median number of intubations was 1 (range 0–5). Patient demographics are further summarized in Table I. pVRQOL and pVOS scores varied widely, demonstrating dysphonia that ranged from severe to absent based upon parental responses (Table II).

Table 1.

Patient Background Information.

Patient Background Information
Age (months)	27.8±35.8 (3–197)
Gestational Age (weeks)	29.4±4.1 (23–37)
Gender (M/F)	40 (58) / 29 (42)
Intubation Length (weeks)	2.99±5.2 (0–22)
Number of Intubations	1.1±1.1 (0–5)
NICU stay (weeks)	9.16±8.5 (0–36)
Weight (kg)	1.48±0.9 (0.51–4.48)
Bronchopulmonary Dysplasia	25 (37.3)
GERD	28 (40.6)
History of Cardiac Surgery	9 (13.8)
PDAL	8 (11.6)

Values listed are mean ± standard deviation (range) for continuous variables and number (percentage) for categorical variables. Age indicates age of the patient (in months) at time of study enrollment. Gestational Age indicates age (in weeks) at birth. NICU- neonatal intensive care unit; GERD- gastroesophageal reflux disease; PDAL- patent ductus arteriosus ligation

Table II.

Spearman Correlations between voice outcome scores and continuous variables.

Spearman Correlation Coefficients (ρ)
	pVOS score			pVRQOL score
	ρ	p-value	N	ρ	p-value	N
Age (mo)	−0.0793	0.5172	69	−0.14833	0.2844	54
Gestational Age (wk)	0.34495	0.0037	69	0.2786	0.0414	54
Intubation length (wk)	−0.52023	<0.0001	69	−0.29512	0.0303	54
Number of Intubations	−0.41013	0.0006	66	−0.20994	0.1392	51
NICU stay (mo)	−0.42526	0.0003	69	−0.29915	0.028	54
Weight (kg)	0.32473	0.0065	69	0.2553	0.0624	54
GERD duration (mo)	−0.2081	0.2976	27	−0.08865	0.7182	19

Spearman correlation coefficients (ρ) are listed in the columns headed by ρ. N indicates total number of observations for a given variable. NICU- neonatal intensive care unit; GERD- gastroesophageal reflux disease; pVOS- pediatric voice outcomes scale; pVRQOL- pediatric voice related quality of life

Spearman Correlation between outcomes and continuous variables

Continuous variables were then evaluated for their correlation with the pVOS and pVRQOL outcomes (Table II). pVOS score was significantly positively correlated with gestational age (ρ=0.34, p=0.0037) and birth weight (ρ=0.32, p=0.0065). Patients closer to term and patients with higher birth weights had better pVOS scores. pVOS score was significantly negatively correlated with length of intubation (ρ=−0.52, p<0.0001), number of intubations (ρ=−0.41, p=0.0006) and length of NICU stay (ρ=−0.42, p=0.0003). Patients with longer length of intubation, increased number of intubations, and longer NICU stays have poorer pVOS scores. pVRQOL score was significantly positively correlated with gestational age (ρ=0.28, p=0.0414) and significantly negatively correlated with length of intubation (ρ=−0.30, p=0.0303) and length of NICU stay (ρ=−0.30, p=0.0280).

Analysis of voice outcome scores by categorical groups

Categorical variables were then correlated to voice outcomes as outlined in Table III. pVOS score was significantly higher in patients without BPD (11.93±1.2) than patients with BPD (10.40±3.07) (p=0.0126). pVOS score was significantly higher in patients without PDA (11.79±1.84) than patients with PDA (10.65±2.71) (p=0.0124). pVOS score was significantly higher in patients without intubation(12.24±1.26) than patients with any intubation (10.98±2.40) (p=0.0036). Additionally, pVOS score was significantly higher in patients with <4 week intubation length (12.02±1.16) than patients with ≥4 week intubation length (9.75±3.14) (p<0.0001). Furthermore, pVOS score was significantly higher in patients with <8 week intubation length (11.61±2.29) than patients with ≥8 intubation length (10.31±1.32) (p=0.0005).

Table III.

Univariate analysis of categorical variables relative to pVOS and pVRQOL.

Variable	N	Mean (Std. Dev)	Mean (Std. Dev)	p-value (2- sided)
Gender		F	M
pVOS score	69	11.45 (2.15)	11.30 (2.24)	0.5705
pVRQOL score	54	88.88 (18.80)	89.52 (17.80)	0.5633
BPD		No	Yes
pVOS score	67	11.93 (1.20)	10.40 (3.07)	0.0126
pVRQOL score	52	92.80 (12.61)	83.03 (24.18)	0.2003
GERD		No	Yes
pVOS score	69	11.32 (1.88)	11.43 (2.62)	0.3168
pVRQOL score	54	91.16 (14.85)	85.97 (22.61)	0.5521
History of heart surgery		No	Yes
pVOS score	65	11.64 (1.75)	9.78 (3.93)	0.0556
pVRQOL score	50	92.06 (14.69)	69.45 (31.17)	0.0232
PDA		No	Yes
pVOS score	65	11.79 (1.84)	10.65 (2.71)	0.0124
pVRQOL score	50	92.26 (14.79)	83.16 (24.10)	0.3299
CPAP		No	Yes
pVOS score	62	12.09 (1.20)	11.18 (2.56)	0.0866
pVRQOL score	48	90.97 (16.01)	90.35 (19.21)	0.9534
Treatment of PDA		PDAL	Indocin
pVOS score	21	9.50 (4.11)	11.46 (1.13)	0.2037
pVRQOL score	14	63.84 (31.27)	94.24 (12.81)	0.0323
Intubation (weeks)		>0	0
pVOS score	69	10.98 (2.40)	12.24 (1.26)	0.0036
pVRQOL score	54	87.20 (20.17)	93.00 (13.10)	0.3146
Intubation (weeks)		≥4	<4
pVOS score	69	9.75 (3.14)	12.02 (1.16)	<0.0001
pVRQOL score	54	75.54 (25.06)	94.51 (11.05)	0.0026
Intubation (weeks)		≥8	<8
pVOS score	69	10.31 (1.32)	11.61 (2.29)	0.0005
pVRQOL score	54	77.79 (20.67)	91.23 (17.07)	0.0165

N- number included in analysis. pVOS- pediatric voice outcome study, pVRQOL- pediatric voice-related quality of life, BPD- bronchopulmonary dysplasia, GERD- gastroesophageal reflux disease, PDA- patent ductus arteriosus, CPAP- continuous positive airway pressure, PDAL- PDA ligation.

pVRQOL score was significantly higher in patients without cardiac surgery (92.06±14.69) than patients with any history of cardiac surgery (69.45±31.17) (p=0.0232). pVRQOL score was significantly lower for patients with patent ductus arteriosus (PDA) who were managed with PDA ligation than those managed medically (p=0.0323). pVRQOL score was significantly higher in patients with <4 week intubation length (94.51±11.05) than patients with ≥4 week intubation length (75.54±25.06) (p=0.0026). pVRQOL score was significantly higher in patients with <8 week intubation length (91.23±17.07) than patients with ≥8 week intubation length (77.79±20.67) (p=0.0165).

Multivariate logistic regression for outcome pVOS and pVRQOL

Variables included in multivariate analysis of pVOS score were gestational age, number of intubations, length of NICU stay, weight, presence of bronchopulmonary dysplasia (BPD), presence of PDA, and length of intubation (<4 vs. ≥4 weeks). When assessing the impact of multiple variables on pVRQOL score, gestational age, length of NICU stay, history of cardiac surgery, and length of intubation (<4 vs. ≥4 weeks) were included. On initial evaluation, under the control of other covariances, each factor did not play a significant role in either model. By using the Backward stepwise selection method the best model for both pVOS and pVRQOL were chosen and one significant factor was identified: intubation length (<4 weeks vs. ≥ 4 weeks). Patients intubated for <4 weeks had a significantly higher probability of having a higher than median pVOS score (p=0.0028, OR=6.4, 95% CI=1.9–21.6). There was no significant difference in comparison of the pVRQOL analysis.

Discussion

Premature infants are subjected to myriad interventions in order to assist with vital functions as they continue to grow and develop. The long-term effects –both intended and unintended- of the measures undertaken to ensure the long term health of these patients are beginning to be understood. When considering airway interventions, the primary focus in long-term studies has long been airway patency, but in recent years, the impact of prolonged intubation and multiple intubations on voice outcomes has been considered. Our data support the commonly held notion that the duration of intubation is negatively correlated with parent-perceived voice outcomes in premature patients independent of other potential confounding variables.

While previous authors have identified the frequency of intubations as a risk factor for decreased vocal quality and voice-related quality of life, their investigation focused on ex-25 week premature infants, effectively excluding a portion of the premature population. This difference in inclusion criteria may have contributed to the disparate findings between their investigation and the current study, with the Australian cohort selecting a more severely compromised cohort.

The strengths of this preliminary investigation include prospective collection of quality of life data, diverse population including both extremely premature infants and near term infants, and assessment of potentially confounding variables. Weaknesses include the retrospective nature of comorbidity identification, relatively small sample size, and lack of correlation of voice quality of life measures with standardized objective vocal quality scores or laryngeal findings. Another weakness of the study was the voluntary nature of recruitment, potentially leading to a reporting bias. Another interesting finding leading to challenges with data analysis was the relatively good voice quality reported overall, with median pVOS and pVRQOL scores at or near normal.

Areas for future research include the identification of objective vocal quality measures to validate the perceptual findings and correlation of vocal quality and voice-related quality of life measures with pathologic changes. While the current study aids in selecting patients at increased risk for any dysphonia, further investigations will ideally correlate physical findings with perceived dysphonia, enabling providers to identify at-risk patients sooner with the hope for early intervention and improved outcomes.

Conclusion

The data suggest that, independent of comorbidities commonly associated with poor vocal quality of life, duration of intubation in the neonatal period is associated with poorer parent-perceived voice quality in premature patients. As a result, the clinician should maintain an increased suspicion for dysphonia in this population and a low threshold for evaluating the larynx to assess for vocal pathology.