Heart rate

One of the earlier studies in the field looked at 60 preterm infants (28–32 weeks GA) randomly assigned to an experimental or control group.17 The experimental group was exposed to a recording of their mother's voice playing for 30 min daily throughout their hospitalisation. At 36 weeks corrected gestational age (CGA), all infants were exposed to 10 s of an 85 dB (re 0.0002 dyne/cm²) white noise in a drowsy state and a 30 s recording of a female's voice in an active crying state. In the drowsy state, the white noise stimulus elicited heart rate acceleration in both groups (p<0.01). However, in the crying state, the white noise stimulus elicited heart rate deceleration, which was more pronounced in the experimental group (p<0.01). Results suggest that an infant's response to noise may not only be dependent on the infant's behavioural state but also on prior exposure to sounds. However, the authors did not mention the sound level of the maternal voice recording, and did not account for the fact that all infants were likely exposed to a female voice for more than 30 min per day at baseline.

Another early study exposed two preterm infants (34–35 weeks GA) to sudden loud NICU noise at 70–75 dBA caused by doors closing, diaper pails and staff conversation.18 Although all of these environmental stimuli were associated with a transient increase in heart rate, the small sample size of this study clearly limits its applicability.

Zahr and Balian looked at 55 preterm infants (23–37 weeks GA) and their heart rate responses to nursing interventions and environmental noise.14 Sound levels were not measured; a bedside observer recorded any loud noises and the infant's response. Although a slight increase in heart rate was found in response to all the noise stimuli, these results were not statistically significant. The lack of sound level measurements, and the extreme variability in gestational age and age at testing limit the impact of this study.

Schulman looked at 31 infants (29–40 weeks GA) at high and low risk for neurological damage.19 All infants were exposed to an 80 dB (re 0.02 dynes/cm²) low frequency buzzer for 3 s in either a quiet alert or quiet asleep state. The high-risk infants had heart rate acceleration in both states. However, the low-risk infants had heart rate acceleration only in the quiet sleep state. Thus, both the behavioural state of the infant and predisposing conditions of the CNS affect the cardiac response. However, this study included both preterm and full-term infants, with significant variation in the age at testing.

Williams et al looked at the effects of noise on 11 extremely low birthweight (ELBW) infants (<1000 g) during the first week of life.20 All infants were exposed to routine incubator noise at baseline levels of 50–60 dBA over a 2 h period. The lower birthweight infants (454–694 g) responded with an increase in heart rate to loud noises. However, the higher birthweight infants (766–910 g) experienced a bi-phasic heart rate response, showing an initial decrease in heart rate followed by heart rate acceleration. The authors concluded that the higher birthweight newborns had more mature CNS control over heart rate.

Several researchers have compared the heart rate responses to noise in full-term versus preterm infants. In one study, the responses of 30 full-term and 15 preterm infants (<1750 g) to a 5 s 3000 cps (sound pressure level 100 dB) warbled tone were compared.21 All infants had heart rate acceleration followed by rebound deceleration below baseline levels after the auditory stimulus (p<0.05). However, the preterm infants were found to have less cardiac acceleration and an overall milder response when compared to full-term infants.21

Field et al compared 18 healthy full-term and 18 high-risk preterm (mean 33.6 weeks GA) infants' responses to five shakes of a rattle or buzzer at 90 dB (re 0.0002 dynes/cm2, A scale).22 All infants responded with heart rate acceleration, however, only the full-term infants showed habituation, with a decreased cardiac response after repeated stimulation (p<0.01). Of note, the preterm infants in this study were tested at a mean chronological age of 23.8 days (37.4 weeks CGA), while the term infants were tested at 2–3 days of life; thus all infants were term at the time of testing. Prior NICU noise exposure and medical complications in ex-preterm infants may have affected their heart rate response to the stimuli.

A more recent study compared the heart rate response of full-term (n=22) and preterm (n=20) infants (mean 32 weeks GA) to a 5 s white noise stimulus of 80, 90 and 100 dBA.23 In both groups, the increase in heart rate was proportional to the intensity of the stimulation; however, this was only significant in the preterm infants with the 90 and 100 dBA stimuli (p<0.01). There was also an overall greater rise in heart rate in the full-term infants.

Blood pressure

Jurkovicova and Aghova exposed 30 low birthweight infants (mean 1613 g) to a 63–250 Hz (79–85 dBA) low frequency noise (generated by the closing of incubator doors) five times over a 30 s interval.24 The infants were also exposed to 30 s of high frequency 4000 Hz continuous noise from the incubator alarm system. In 85% of infants, all noise stimuli resulted in a 10 mm Hg increase in systolic blood pressure by and a 9 mm Hg increase in diastolic blood pressure, both of which returned to baseline after 5 min. However, all of the blood pressure measurements were within the normal range.

Williams et al looked at the effect of noise on mean arterial blood pressure (MABP) in 11 ELBW infants (401–1000 g) during the first week of life.20 Infants were exposed to baseline levels of 50–60 dBA incubator noise over a 2 h period. The higher birthweight infants demonstrated a bi-phasic blood pressure response to noise; they first responded with a decrease in blood pressure followed by an increase. The lower birthweight infants also experienced some decreases in blood pressure after the noise stimulus. The correlations between heart rate and MABP were stronger in higher birthweight newborns.20 However, none of these blood pressure changes were statistically significant.

Language and development

In one of the only randomised controlled trials in the field, Turk et al randomised 34 VLBW (<1500 g) infants to a control group and an experimental group which received silicone ear plugs from 1 week of life to 35 weeks CGA.31 On evaluation of 14 of the surviving ELBW (<1000 g) infants at 18–22 months, those who wore earplugs scored 15.5 points (95% CI 3.03 to 28.02) higher than controls on the Bayley Mental Development Index (p=0.02).

Stromswold and Sheffield evaluated the effects of NICU noise on language development. Parents of 382 ex-preterm children (mean 31.8 weeks GA) retrospectively rated the sound level of their child's NICU and assessed their linguistic development.32 NICU sound level was a better independent predictor of language outcome than gestational age, birth weight or socioeconomic status, with children in louder NICUs having better language outcomes (table 4). This study is limited by the subjectivity of parental retrospect ratings of NICU sound levels, and the potentially biased parental assessments of their child's language abilities. There have been no further studies to challenge or confirm these findings.

Hearing

Whether or not the extra auditory stimulation preterm infants receive in the NICU affects auditory development and hearing has also been questioned. Established predictors of hearing loss in NICU infants are prolonged mechanical ventilation, patent ductus arteriosus ligation, length of stay in the NICU, aminoglycosides, intraventricular haemorrhage, asphyxia and elevated bilirubin levels.2 33 Thus, paradoxically, it appears that the infants needing the most intensive care, and therefore spending the longest amounts of time exposed to high levels of noise, are actually those who have the highest rates of hearing loss. Noise-induced hearing loss is a well established phenomena, with the duration of noise exposure correlating with the severity of hearing impairment in a dose–response relationship.34 However, there are no studies looking directly at the effects of NICU noise on hearing loss in preterm infants. Most of the data in this area of research come from animal models, which suggest delayed development of a normal primary auditory cortex and damage to auditory hair cells with early noise exposure.35 36

Several studies have investigated the possible synergistic effects of aminoglycosides and noise on hearing loss in preterm infants with conflicting results (table 4).35 36 Winkel et al looked retrospectively at 91 ex-preterm infants (<2000 g) when 4–6 years old.39 Infants were divided into three groups: a group treated with both kanamycin and incubators (group 1), a group treated with incubators only (group 2), and a group exposed to neither (group 3). Incubator noise levels were estimated to be 58–75 dB based on prior studies. All five cases of moderate to severe sensorineural hearing loss were infants treated with kanamycin and kept in incubators (range 8–18 days). Of note, among the 56 incubator-treated children without hearing loss, 52% had an audiogram pattern suggesting minimal cochlear lesions similar to noise-induced lesions; this was unrelated to the amount of time spent in the incubator. In the 18 non-incubator treated children without hearing loss, only one child had this audiogram pattern (p<0.002), suggesting a possible negative effect of incubator noise on the cochlea. However, medical complications known to be associated with hearing loss, which are more likely to occur in the infants kept in incubators, likely contributed to the cochlear lesions found in this population.

D'Souza also looked at 26 survivors of severe perinatal asphyxia, all of whom had been kept in incubators.40 Fifteen of the 26 subjects received gentamicin, and only one was diagnosed with deafness. The results of this study should be treated with caution due to the wide range of gestational ages of the subjects and the varied duration of their incubator stays (range 2–27 days).