Article Text

Download PDFPDF

Noise levels during nCPAP are flow-dependent but not device-dependent
  1. O Karam1,
  2. C Donatiello1,
  3. E Van Lancker2,
  4. V Chritin2,
  5. R E Pfister1,
  6. P C Rimensberger1
  1. 1
    Pediatric and Neonatology Critical Care Unit, Geneva’s University Hospital, Geneva, Switzerland
  2. 2
    IAV Engineering, Ecole Polytechnique Fédérale de Lausanne, Switzerland
  1. Dr Oliver Karam, Soins Intensifs de Pédiatrie, Hôpital des Enfants, 6, rue Willy Donzé, 1211 Genève, Switzerland; karam{at}


Objective: Nasal continuous positive airway pressure (nCPAP) has been shown to improve the outcome of infants with respiratory distress syndrome. However, noise generation could be of concern. Therefore, our study was designed to measure the noise levels of various CPAP drivers.

Design: For infants admitted to our neonatal intensive care unit and paediatric critical care unit, we measured the noise level in the oral cavity, using a microphonic probe with a flexible capillary tube. Various CPAP drivers and interfaces have been tested.

Results: 27 measurements were made in eight infants. Mean noise level was 88.6 (SD 18.8) dB and was correlated with flow (p<0.01) but not with pressure. A noise level above 90 dB was detected in 67% of the measurements.

Conclusions: nCPAP drivers are valuable devices for neonatal care that may prevent primary mechanical ventilation or re-intubation, but generate a large amount of noise, often higher than occupational limits accepted for adult workers. Therefore, new devices must be designed to minimise this possible noxious exposure of premature infants to unacceptably high noise levels.

Statistics from

In 1971, Gregory et al demonstrated that continuous positive airway pressure (CPAP) significantly improved hypoxia in respiratory distress syndrome (RDS) in the neonate.1 Since then, numerous studies have shown that nasal CPAP (nCPAP) can prevent primary mechanical ventilation for neonates with respiratory distress syndrome, as well as re-intubation after a period of mechanical ventilation.2 Recently, many different nCPAP devices for use in neonates have been made available, generating pressure in different ways. Whereas variable-flow nCPAP devices are thought to be better3 than continuous-flow nCPAP in improving lung volume and diminishing the work of breathing, the method of generating this variable-flow pressure seems not to be clinically relevant.4

All nCPAP devices produce a certain level of noise that, in some cases, has been measured to be at very high levels.5 It is unknown if this noise impacts the hearing development in term or preterm infants, but there are national and international recommendations to limit the noise level in neonatal intensive care units (NICU) and within the incubator below 45 dB and 60 dB, respectively.6 7

As very little is known about noise generation by different devices in the NICU, the purpose of this study was to examine the noise levels to which our patients are exposed, while using various nCPAP devices at various flow rates.


The measurements were made in a paediatric and neonatal critical care unit (PICU and NICU) setting, in a tertiary paediatric hospital. The protocol was approved by the local ethics committee, and we obtained a signed consent from all patients’ parents.

In order to determine the best method of measuring the noise in the oral cavity, we first measured the noise in a manikin equipped with a nCPAP.

A microphonic probe (Brüel & Kjaer 4182, Nærum, Denmark), inside a sterile flexible capillary, was slowly inserted through the closed mouth. For each patient, noise level was measured as the probe was inserted and showed initially important fluctuations (fig 1A). While progressing in the oral cavity a zone of less turbulent airflow was encountered by the microphone and the noise level stabilised (fig 1B), before becoming fluctuant again as the probe was pushed further into turbulent air from the nasal cavity. Measurements were taken in the position where the fluctuations were minimal. The mean noise level was calculated over a 5-second period.

Figure 1 Example of the measured noise level in a patient. (A) When the probe was inserted, the noise levels fluctuated until the probe was out of turbulent airflow, then became turbulent again as the probe reached the flow coming down from the nose. (B) The noise level was measured during the period with the fewest fluctuations.

We used a dBA weighting, in which the low frequencies are attenuated, just as the human auditory system does. For each patient, and within the therapeutic range, we measured the noise level at different flow and pressure levels. All measurements were performed by the same investigator.

We tested three types of nCPAP devices: Arabella (Hamilton Medical, Reno, USA), F15 Medijet (Medical Innovations International, Puckheim, Germany) and Infant Flow (EME Ltd, Brighton, United Kingdom). They had three different types of prongs: nasal prongs 33201 (Hamilton Medical, Reno, USA), Medijet Metalic nasal prongs (Medical Innovations International, Puckheim, Germany) and EME nasal prongs (EME Ltd, Brighton, UK).

Statistical analysis was carried out where applicable after logn transformation using the independent-samples t test procedure (when comparing two groups), analysis of variance (ANOVA) with Tukey HSD post-hoc analysis (for comparison of three or more groups) and Spearman’s test (for non-parametric correlations). The analyses were performed on an IBM compatible personal computer, using SPSS 15.0 (SPSS Inc, Chicago, IL, USA). Results are expressed as mean (SD). Differences were considered significant for p<0.05.


Twenty-seven measurements were performed on eight premature infants, with a mean weight of 1436 g (range 590–3245 g). Each patient had three measurements in the therapeutic range, at different flows and pressures. One patient had two sets of measurements, on two different days, with two different nCPAP devices. The mean noise level measured in the oral cavity during nCPAP was 90.5 (18.7) dB. The noise level was correlated with the CPAP flow (R2 = 0.30, p = 0.008), but was not correlated with the CPAP pressure level (p = 0.97) (figs 2 and 3). Sixty-seven per cent (18/27) of the measurements were above 90 dB, and 96% (26/27) above 60 dB.

Figure 2 Scatterplot of the noise (dB) generated by airflow in infants on nCPAP, showing a significant linear correlation (R2 = 0.30, p<0.01). Eighteen measurements out of 27 were above 90 dB (in the shaded zone).
Figure 3 Scatterplot of the noise, compared to the pressure generated by the nCPAP (p = 0.97).

There was no statistically significant difference in terms of noise level between the different nCPAP devices. The noise was not correlated with the infant’s weight (p = 0.499).


For nearly 40 years1 CPAP has been used to support recently extubated preterm infants, those experiencing significant apnoea of prematurity and those with respiratory distress after birth, as an alternative to intubation and mechanical ventilation. In physiological terms, nCPAP has been shown to increase functional residual capacity and improve oxygenation; CPAP dilates the larynx, reduces supraglottic airway resistance and lessens the incidence of obstructive apnoea, and improves the synchrony of respiratory thoracoabdominal movements.2 Numerous studies indicate that nCPAP reduces the need for mechanical ventilation in neonates with respiratory distress syndrome, as well as the secondary re-intubation rate after a period of mechanical ventilation.2 However, nCPAP generates a significant noise, most probably the result of the air turbulence in the upper airways.

The human cochlea and peripheral sensory end organs complete their normal development by 24 weeks of gestation.8 The hearing threshold at 27–29 weeks of gestation is approximately 40 dB and decreases to nearly adult levels (13.5 dB) by 40 weeks of gestation.8 Occupational standards for the workplace limit exposure to 90 dB for no more than 8 hours.6 The US Environmental Protection Agency has proposed an average sound level of 45 dB in hospitals and in neonatal intensive care units (NICU).7 The International Standard IEC 60601-2-19 (1990) states that “the sound level within the baby compartment of an incubator should not exceed a sound pressure level of 60 dB.”9 Several studies show that noise in a NICU may be responsible for hypoxaemic episodes10 or behavioural changes.11 12 In the animal model, 1-week-old guinea pigs exposed to an incubator noise of 70–75 dB developed internal ear damage.13 These noise levels were measured near the animal, but not directly in the internal ear. Freeman et al showed that rats exposed to 90 dB broadband noise 12 hours a day for 15 days had no long-term change in auditory function, whereas those exposed to 102 dB had reduced auditory brainstem response thresholds.14 Therefore, it is still not known how much noise may be tolerated by the developing ear of a preterm infant, nor if the duration of exposure plays a role. Furthermore, this is complicated by the fact that little is known on sound transmission in the newborn. Relative importance of the different pathways of conduction to the inner ear, bone conduction, internal air conduction via the eustachian tube and air conduction via the external ear and eardrum are at the very best speculative.

We measured high noise levels in the mouth, in neonates on nCPAP. The noise level was directly related to the flow through the circuits, but not to the pressure generated. There was no difference among the various types of nCPAP devices.

Our data confirm those of Surenthiran et al, who had explored the noise generated by nCPAP.5 However, whereas they only measured the noise levels in patients on the Infant Flow Driver nCPAP device, and were therefore unable to generalise their findings, we could show that all of the most commonly available nCPAP devices generate noise levels that are much higher than the recommended maximum noise levels in a NICU (45 dB),7 despite various circuit designs.

It is known that conventional mechanical ventilation and high-frequency oscillatory (HFO) ventilation5 15 generate noise levels of around 50–60 dB. We measured a mean noise level of 90 dB in infants on nCPAP.

Since the reported concept of audio trauma in infants,13 a number of strategies have attempted to diminish noise levels, such as noise-sensor alarms, sound-absorbing panels and blankets as well as staff education. Recently, manufacturers started to offer earmuffs, hoping to reduce noise levels during nCPAP. However, while measuring high noise levels in the mouth, noise transduction through bone conduction and through the eustachian tube may not sufficiently be attenuated by such precautions.

Two-thirds of our infants are subject, 24 hours a day, to noise levels higher than those tolerated for adults working 8 hours a day. This is worrying and, therefore, we believe that increased knowledge on noise generation and suppression of equipment in the intensive care setting and its impact on the developing ear are urgently required. New designs for nCPAP generators and/or circuits are needed and must plan for noise reduction.

What is already known on this topic

  • Nasal continuous positive airway pressure (nCPAP) has been shown to reduce the need for mechanical ventilation in infants.

  • Continuously elevated noise levels above 90 dBA are possibly dangerous for the hearing development in infants and actual safety recommendations limit maximal acceptable noise levels to 60 dB for the babies’ environment. So far, only the infant flow nCPAP device has been shown to generate high noise levels.

What this study adds

  • Noise levels generated by nCPAP devices are correlated to airflow, but not to generated pressure. Generated noise levels seem to be largely independent of the nCPAP device.

  • nCPAP-induced noise exposure to more than 90 dB could be measured for two-thirds of our infants.



  • Competing interests: None.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Fantoms
    Martin Ward Platt, Deputy Editor