Article Text
Abstract
Background Noise exposure in the neonatal intensive care unit is believed to be a risk factor for hearing loss in preterm neonates. Continuous positive airway pressure (CPAP) devices exceed recommended noise levels. High flow nasal cannulae (HFNC) are an increasingly popular alternative to CPAP for treating preterm infants, but there are no in vivo studies assessing noise production by HFNC.
Objective To study whether HFNC are noisier than bubble CPAP (BCPAP) for preterm infants.
Methods An observational study of preterm infants receiving HFNC or BCPAP. Noise levels within the external auditory meatus (EAM) were measured using a microphone probe tube connected to a calibrated digital dosimeter. Noise was measured across a range of frequencies and reported as decibels A-weighted (dBA).
Results A total of 21 HFNC and 13 BCPAP noise measurements were performed in 21 infants. HFNC gas flows were 2–5 L/min, and BCPAP gas flows were 6–10 L/min with set pressures of 5–7 cm of water. There was no evidence of a difference in average noise levels measured at the EAM: mean difference (95% CI) of −1.6 (−4.0 to 0.9) dBA for HFNC compared to BCPAP. At low frequency (500 Hz), HFNC was mean (95% CI) 3.0 (0.3 to 5.7) dBA quieter than BCPAP. Noise increased with increasing BCPAP gas flow (p=0.007), but not with increasing set pressure. There was a trend to noise increasing with increasing HFNC gas flows.
Conclusions At the gas flows studied, HFNC are not noisier than BCPAP for preterm infants.
- Noise
- Prematurity
- Continuous positive airway pressure
- Noise induced hearing loss
- High flow nasal cannulae
Statistics from Altmetric.com
- Noise
- Prematurity
- Continuous positive airway pressure
- Noise induced hearing loss
- High flow nasal cannulae
What is already known on this topic
-
▸ Noise exposure is believed to be a risk factor for hearing loss in preterm neonates.
-
▸ Continuous positive airway pressure (CPAP) devices exceed recommended noise levels in clinical and in vitro studies
-
▸ High flow nasal cannulae (HFNC) are increasingly popular as an alternative to CPAP for treating preterm infants.
What this study adds
-
▸ HFNC are not noisier than bubble CPAP for preterm infants.
-
▸ Infants receiving HFNC support are exposed to noise levels above those recommended for neonates.
-
▸ Noise exposure increases with bubble CPAP gas flow but not set pressure, and there is a trend to increasing noise with increasing HFNC gas flow.
Background
Permanent hearing loss is a recognised complication of being born extremely preterm. In a population-based study of children born ≤28 weeks’ gestation or <1250 g birth weight,1 Robertson and colleagues found that 3.1% of survivors at 3 years of age had permanent hearing loss, and 1.9% had severe or profound loss. In their study, duration of treatment with supplemental oxygen was the most significant predictor of hearing loss.
Noise exposure during admission in a neonatal intensive care unit (NICU) is also believed to be a risk factor for hearing loss in vulnerable neonates, and cochlear damage secondary to noise exposure has been demonstrated in animal models during the newborn period.2–4 Exposure of preterm infants to noise in the NICU has been associated with negative short-term effects on their cardiovascular and respiratory systems,5 and guidelines exist for the recommended acoustic design of NICUs.6 The American Academy of Paediatrics has noted that prolonged exposure to sound levels above 90 decibels (dB) leads to hearing loss and recommends that ideally, for neonates in the NICU environment, noise levels above 45 dB should be avoided.7 Many studies have reported noise levels in the NICU exceeding this recommended limit, in some cases for as much as 70% of the time.8 ,9
The use of ‘non-invasive’ respiratory support via binasal prongs, such as continuous positive airway pressure (CPAP) and high flow nasal cannulae (HFNC), to treat preterm infants in the NICU has greatly increased in recent years. These devices produce noise audible at the cot side, and there has been concern that they may negatively impact the hearing of infants, particularly given that some preterm infants are exposed to these devices for weeks or even months.
CPAP, the most commonly used mode of non-invasive respiratory support, has been shown to exceed recommended noise levels in in vitro10 ,11 and clinical studies in the NICU.12 ,13 The set gas flow was the main determining factor of noise production in three of these studies.11–13 HFNC are increasingly popular as a mode of non-invasive respiratory support in NICUs and non-tertiary neonatal units, with gas flows as high as 6–8 L/min commonly used.14–16 Given the findings of the CPAP studies, there is concern that HFNC may expose infants to noise levels greater than from CPAP. A recent in vitro study17 is the first to compare noise production from HFNC with that from CPAP. Noise levels generated by two HFNC devices (Fisher & Paykel ‘Optiflow’ and Vapotherm ‘Precision Flow’) were higher than those produced by continuous-flow CPAP, and again well above recommended levels.
There are no published clinical studies of noise levels from HFNC compared with CPAP in preterm infants. We designed a clinical study to test the hypothesis that HFNC is noisier than CPAP for preterm infants.
Methods
This was an observational study of a convenience cohort of preterm infants, conducted in The Royal Women's Hospital (RWH) Neonatal Intensive and Special Care unit in Melbourne, Australia. The study was approved by the RWH Human Research Ethics Committee. Prospective parental consent was obtained before study entry.
Infants were eligible for inclusion in the study if they were inpatients in the RWH NICU, born preterm (<37 weeks’ gestation), receiving non-invasive respiratory support from either HFNC or ‘bubble’ CPAP (BCPAP), and considered medically stable by the treating clinical team. At the time of the study at RWH, heated and humidified HFNC was delivered using the Fisher & Paykel ‘Optiflow’ circuit and binasal ‘BC’ prongs (Fisher & Paykel Healthcare, Auckland, New Zealand), and BCPAP was delivered using the Fisher & Paykel bubbler system (Fisher & Paykel Healthcare, Auckland, New Zealand) and binasal Hudson prongs (Hudson RCI, Research Triangle, North Carolina, USA). Infants were in shared four or six cot rooms, and were nursed in closed incubators or open cots, depending on gestation, weight, and severity of illness. The primary outcome of the study was the average noise level measured at the external auditory meatus (EAM) of infants during treatment with HFNC or BCPAP. Noise level was expressed as decibels A-weighted (dBA) (see Terminology).
At the time of study design there was no data available on noise production with HFNC, and previous clinical studies of CPAP noise production did not provide a sample size calculation. We aimed to recruit a convenience sample of 10 preterm infants in each of the HFNC and BCPAP groups (20 infants total).
Data were collected on infant and maternal demographics. The type and level of respiratory support being received was recorded: flow in Litres per minute for HFNC; flow (L/min) and set pressure in centimetres of water (cm H2O) for BCPAP.
Terminology
The intensity (loudness) of sound is measured in decibels (dB), a logarithmic scale. To be meaningfully measured, a sound must be compared to a reference level. ‘Sound pressure level’ (SPL) is a logarithmic measure of the effective sound pressure of a sound relative to a reference value, measured in dB above a standard reference level (dB SPL). The human ear is less sensitive to sound frequencies that are either very low or very high in frequency. ‘A-weighted’ decibels (dBA) is an accepted method of measuring noise levels, that reduces the contribution to the measured level by these very high or very low frequency sounds.
Noise measurements
Measurements were timed to coincide with routine nursing cares, when respiratory support would routinely be discontinued for a brief period by the nursing staff. Staff were requested to minimise noise in the room at the time of the study, but total silence was not enforced. Noise measurements were discarded and repeated if loud incidental noises occurred (eg, loud talking, doors closing, infant crying).
There was no alteration to the prescribed respiratory support at the time of the study. Prior to measurements being taken on each infant, the ‘ambient’ noise at the centre of each room was measured with a digital sound metre (Model 2100 Sound Level Meter, Quest Technologies, Oconomowoc, Wisconsin, USA). For the primary outcome, noise levels within the EAM were measured using a microphone probe tube, connected to a calibrated digital dosimeter (SV 102 Acoustic Dosimeter, SVANTEK, Poland) (see online supplementary figure S1). The tip of the probe tube was inserted only a few millimetres into the EAM so as not to contact the eardrum. A new probe tube was used for each infant, and all equipment was cleaned thoroughly with an aseptic solution before and after each infant was studied. A similar method has been safely used in previous studies by Karam and Surenthiran.12 ,13
At the EAM, noise was measured with the gas flow turned on and then off: the difference between these measurements serves as an estimate of noise contribution from the respiratory support above ambient cot noise. Readings were taken from the same ear, each for a period of 10 s, to produce an equivalent continuous sound level, a measure that accounts for fluctuations in sound energy over a given timeframe, as recommended previously.18 Three such readings were taken, and the results averaged to produce a final measurement at each level of support.
Data were recorded in a case report form, entered into an electronic database (Epidata, The Epidata Association, Odense, Denmark), and then exported for statistical analysis to an electronic statistical package (Stata, Intercooled 10, Stata Corp, College Station, Texas, USA). Noise levels from HFNC and BCPAP were compared using an unpaired t test, on the basis that despite this being a small sample, the data had an approximately normal distribution. An adjusted regression analysis with a robust SE was then used to account for clustering.
Results
A total of 21 HFNC noise measurements and 13 BCPAP noise measurements were performed in 21 infants (table 1). Some infants were studied more than once, at different set HFNC flows or BCPAP pressures. The demographics of the two groups were similar, including the environmental room noise at the time of study.
The primary outcome is shown in table 2. There was no difference in average noise levels measured at the EAM: a mean difference (95% CI) of −1.6 (−4.0 to 0.9) dBA for HFNC compared to BCPAP. At low frequency (500 Hz), HFNC was significantly quieter than BCPAP, with a mean difference (95% CI) of −3.0 (−5.7, to −0.3) dBA. At higher frequencies, although average noise measurements were lower for HFNC by between 0.4 and 1.8 dBA, these differences were not statistically significant.
Table 3 shows the differences in noise measurements for each group with the gas flow on and off. There were no statistical differences between HFNC and BCPAP. The devices contributed noise between 0.2 and 4.8 dBA above ambient cot noise, depending on the frequency measured.
Figure 1 shows the relationships between gas flow and EAM noise for HFNC and BCPAP, and between set pressure and noise for BCPAP, illustrated by a line of best fit, and p value resulting from linear regression analysis. There was a direct relationship between greater BCPAP gas flow and increased noise measurement at the EAM (p=0.007). A similar trend was observed for HFNC, but this did not reach statistical significance. There was no relationship between BCPAP set pressure and noise.
Discussion
This is the first clinical study measuring the noise levels to which preterm infants are exposed during HFNC therapy. This is important due to the rapidly increasing popularity of HFNC as non-invasive respiratory support in the neonatal population,19 and ongoing concerns regarding the contribution of noise exposure to the risk of long-term hearing impairment in preterm infants.
We have demonstrated that HFNC therapy delivered at the flows used in this study is not noisier than bubble CPAP. However, it should be noted that the median noise levels measured during both types of therapy were above the American Academy of Paediatrics’ recommended limits for prolonged noise exposure,7 and we must continue to strive to minimise noise from these devices.
This study found that noise from BCPAP increased with increasing gas flows, and there was a trend towards a similar finding with HFNC, which was consistent with previous studies of noise associated with CPAP.11–13 The statistically lower noise level generated by HFNC than BCPAP at 500 Hz, which we postulate may be due to the low frequency noise generated by the bubbling of the BCPAP system, is interesting, but probably not clinically important.
Our study has limitations. This was an unblinded study, with a pragmatic sample of recordings in a relatively small group of 21 infants. Some infants had more than one measurement, or were included in both groups. While our data appeared to have an approximately normal distribution, this is a relatively small sample size. However, non-parametric analysis of our data (not shown) produced very similar results and did not alter any of our conclusions. There were not sufficient data on noise levels produced by HFNC to allow a formal sample size calculation. The HFNC flows studied in our trial (2–5 L/min) are somewhat lower than the maximum rates reported in clinical practice (up to 8 L/min in preterm infants).15 ,16 This reflects our decision not to alter the prescribed respiratory support of any of the infants participating in this study and that infants receiving higher gas flows were not considered stable for inclusion, as many had been recently extubated from the mechanical ventilator. A previous in vitro study17 demonstrated increasing noise levels with higher HFNC gas flows. Our data showed a relationship between increasing gas flow and noise levels for CPAP, but this relationship did not reach significance for HFNC. It is possible that noise levels at higher HFNC gas flows (6–8 L/min) are greater.
Infants included in this study were treated with the Fisher & Paykel HFNC and BCPAP systems, and noise production from other devices such as the Vapotherm HFNC device (Vapotherm Inc, Stevensville, Maryland, USA), or other CPAP devices, may be different.
While the environmental noise levels measured in the NICU rooms did not differ between groups, background noise may have influenced the readings, despite our efforts to minimise this effect. While we only measured noise levels at the EAM of infants, previous authors have found that in babies receiving CPAP, noise intensity within the upper airway can be significantly greater, potentially due to the turbulent airflow generated.12 ,13 This noise may be transmitted to the inner ear via bone conduction or the Eustachian tube, and if very high might be clinically more relevant than EAM levels. However, at present we have little knowledge of whether noise levels in the upper airway, the EAM or the two in combination are the most important for determining risk of hearing loss.
It is important to remember that while animal models have clearly shown that the cochlea of the newborn is more susceptible to noise-induced damage than that of adults,2–4 our understanding of the contribution of noise to the process of hearing loss in the human neonate is poor. It is likely that hearing loss as a complication of premature birth is multifactorial in nature. Previous studies have postulated that noise may have a synergistic effect with other factors such as hypoxia, and exposure to aminoglycoside antibiotics.20 ,21 The list of identified risk factors for hearing loss is extensive, including family history, craniofacial abnormalities, prenatal infection, culture-positive sepsis and hyperbilirubinaemia requiring exchange transfusion.7
Zahr demonstrated that the use of noise-reducing earmuffs was associated with higher oxygen saturations, less behavioural disturbance and longer periods of sleep in a group of preterm infants.22 It may be that simple measures such as the use of earmuffs in infants receiving non-invasive respiratory support could have an effect on longer-term hearing outcomes, although earmuffs will not reduce conducted upper airway noise secondary to turbulent gas flow. As HFNC has become rapidly accepted into clinical practice, it is important for clinicians to consider not just its effects on the respiratory system, but its wider implications for other important neonatal outcomes.
Conclusion
This study showed that at the gas flows studied, HFNC are not noisier than BCPAP for preterm infants. Both therapies produced noise levels >45 dBA. There was a trend towards noise increasing with higher HFNC gas flow, and this relationship was significant for BCPAP. There was no relationship between BCPAP set pressure and noise.
Acknowledgments
We would like to thank the Department of Biomedical Engineering, The Royal Women's Hospital, Melbourne, for providing some of the equipment used in this trial.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online supplement
Footnotes
-
Contributors CTR analysed the data, drafted the initial manuscript, and revised and approved the final manuscript. He is guarantor. BJM, JAD, PGD, PJC and SD conceptualised and designed the study, interpreted the data and revised and approved the final manuscript. EA conceptualised and designed the study, and analysed and interpreted the data. Data was collected by BJM, JAD and EA.
-
Competing interests None.
-
Ethics approval The Royal Women’s Hospital Human Research Ethics Committee.
-
Provenance and peer review Not commissioned; externally peer reviewed.