Impact of earplugs and eye mask on sleep in critically ill patients: a prospective randomized study

Over the past decade, a large body of literature has raised the major issue of sleep disturbances in critically ill patients [1, 2]. Sleep is reduced, fragmented, does not follow the regular circadian rhythm, and contains increased N1-N2 stages, to the detriment of N3 and rapid eye movement (REM) sleep. Importantly, N3 and REM sleep play a critical role in many physiologic functions, including the central nervous, cardiovascular, endocrine, respiratory, and immune systems. The deleterious consequences of poor sleep in patients admitted to the intensive care unit (ICU) are becoming increasingly clear [1]. Poor sleep is a risk factor for delirium [3], noninvasive ventilation failure [4], and an intrinsic ICU stressor for patients [5], which may in turn participate in the mechanisms of posttraumatic stress disorder [6]. The impact of poor sleep on immune function, metabolism, length of mechanical ventilation, and post-ICU quality of life is also suspected but has not been clearly demonstrated [7]. Sleep improvement has therefore become a goal of care in the ICU [8].

The multiple mechanisms responsible for altered sleep include environmental factors such as noise and light, including those related to human interventions [9]. However, the reduction of noise and light during the night, although theoretically feasible, is not easy to achieve in the ICU, because a high level of human activities in the ICU during the night may be required by the condition of admitted patients or the admission of new patients. In addition, for safety reasons, the noise level of alarms cannot always be turned off or even lowered [10]. An alternative strategy would be to protect patients individually against noise and light by means of earplugs and an eye mask. Previous reports have suggested that this strategy could increase sleep quality in healthy subjects submitted to a level of noise and light encountered in an ICU [11] and in patients sleeping in the postanesthesia care unit [12]. However, the impact of earplugs and an eye mask on sleep quality has not previously been evaluated in ICU patients by polysomnography.

We therefore conducted a randomized controlled trial to determine the efficacy of this strategy on the basis of the hypothesis that earplugs and an eye mask would improve sleep quality during the first night following the initiation of the intervention in critically ill patients, as measured by the proportion of N3 sleep, also known as non-REM sleep stages 3 and 4 or slow-wave sleep. We selected this criterion because N3 sleep is not only thought to be the most “restorative” sleep stage [13–15] but also is involved in the pathogenesis of metabolic and cardiovascular diseases [16, 17] and is noticeably reduced in ICU patients [9, 18, 19].

Human research ethics committee approval for the study was provided by the Comité de Protection des Personnes - Ile de France 6. Patients or their next of kin gave informed consent. Data were collected from July 2012 to December 2013.

The main findings of this study can be summarized as follows. Earplugs and an eye mask applied from awaking following interruption of sedation and until ICU discharge (1) failed to significantly increase the proportion of N3 sleep on the first night following inclusion, but significantly decreased the number of prolonged awakenings; (2) were poorly tolerated, but may increase the duration of N3 sleep in patients who tolerate them; and (3) had no impact on outcome.

This study demonstrates that sleep was severely altered in critically ill patients. Sleep alterations involved both sleep duration and architecture and were consistent with previous reports [19, 26, 27], although the time spent in N3 sleep tended to be longer in our study. The severity of sleep alterations observed in this study shows that the study was conducted in patients with poor-quality sleep who were likely to benefit from an intervention designed to improve sleep quality. Of note, one of the strengths of our study was the use of polysomnography as a key outcome measure, because, to our knowledge, this study is one the largest studies including polysomnography recording in critically ill patients [4, 9, 19, 26–32]. However, performing polysomnography to provide analyzable data is a challenge in the ICU, as recently reported [33]. In contrast with recent reports, we did not observe the atypical sleep stages described in ICU patients, namely pathologic wakefulness and atypical sleep [29, 34], because the EEG patterns observed complied with the Rechtschaffen and Kales scoring system [22]. This result could be explained by the exclusion of patients in whom sedation or high-dose opioids (major sources of altered sleep and EEG patterns) [35] had been discontinued for < 24 h. It is also noteworthy that most patients in our study were not intubated at the time of polysomnography, although a substantial proportion of them had previously been intubated (Table 1) and were globally less severely ill than patients in previously published studies. This could explain the absence of atypical sleep recordings, which have been described mostly in critically ill patients [29, 34]. Of note, the intervention was applied from 10:00 p.m. to 8:00 a.m., which can be considered as too long. However, this allowed a little flexibility and guaranteed that patients would receive the intervention for at least 8 h. Sleep recording was performed from 2:00 p.m. to 8:00 a.m. in order to be consistent with most studies on sleep in the ICU. Also, we admit that sleep outside, and particularly before, the intervention is unlikely to be altered by the intervention. For this reason, we report our results for the whole duration of recording and also for nighttime.

The N3 percentage of total sleep time, which was the primary outcome of our study, was not significantly different between the two groups. Total sleep and night sleep times, as well as various descriptors of sleep architecture, were also not significantly different, suggesting that a protective strategy based on the use of earplugs and an eye mask at night does not improve sleep duration or architecture. These results are in contrast with those of previous studies reporting the benefit of this strategy. However, none of these studies were conducted in a general ICU population, and none were based on the use of polysomnography, because the only study that used polysomnography was conducted in healthy subjects subjected to ICU noise and light conditions [36]. In this study, earplugs and an eye mask improved REM sleep time and arousals but failed to improve N3 time, total sleep time, and sleep efficiency. Studies conducted in critically ill patients showed that earplugs and an eye mask were associated with an improvement in score-assessed sleep perception [12, 36, 37] as well as melatonin and cortisol levels [36]. In addition, a recent meta-analysis suggested that the use of earplugs and an eye mask was associated with a significant reduction of the incidence of delirium [38, 39]. None of these studies used polysomnography, and most of them were conducted not in a general ICU but in a postanesthesia care unit [12] or cardiac surgical ICU [36].

A major limitation of our study was that many patients did not wear earplugs and an eye mask all night long. Subsequently, the study was likely to be underpowered to detect a significant difference, as suggested by the increase in N3 stage sleep in the per-protocol analysis. Some patients removed their devices, whereas in others earplugs and the eye mask shifted during the night. The general tolerability of the intervention is a key to its success. Previous studies have shown that many patients found earplugs and even an eye mask uncomfortable or very uncomfortable [40], with compliance averaging 13% [38]. Patients complained about earplugs not staying in place and sore ears or reported feeling anxious when they did not hear any background noise [11, 40, 41]. Patients also complained that eye masks made them feel hot and sweaty and were too tight, causing a feeling of claustrophobia [40]. Consequently, patients may be unwilling to use earplugs or an eye mask [42]. This is all the more true in that patients of the intervention group who wore earplugs only part of night had even poorer sleep quality than patients in the control group. It raises the hypothesis that the poor tolerance of the device altered sleep architecture, possibly because of anxiety and claustrophobia. ICU staff should therefore improve the acceptability of these devices by clearly explaining the potential benefits to the patients, by helping patients to choose the best device in terms of shape and size, and by providing adequate assistance and instructions concerning their use [36]. Future improvement of the quality of these devices may also help to improve their tolerance.

Another limitation of our study that could have influenced the results was the fairly large proportion of N3 sleep in the control group. With a median of 11% of total sleep time, the proportion of N3 sleep observed in our study exceeded that reported in previous studies [9, 19]. This high proportion of N3 sleep could be consistent with the major efforts undertaken in our unit to improve sleep quality, including, but not limited to, lights turned off at 10:00 p.m. in rooms, televisions turned off at 10:00 p.m., and minimal room alarms. In addition, window blinds are raised during the daytime, and patient mobilization is encouraged in order to promote normal circadian rhythms. All patients are also accommodated in single rooms. Despite these measures, median sound levels were 55 (54–58) dB and 56 (54–57) dB in the control group and intervention group, respectively, with peaks > 70 dB. These levels are higher than those recommended by the World Health Organization (>35 dB) but similar to the noise levels reported in previous studies [43].

However, this protective strategy was associated with a reduction of prolonged awakenings, which suggests that earplugs and an eye mask had at least a minor impact on sleep architecture. It is of note that this is a relevant criterion in term of sleep disturbance and that one could argue that it would have been more logical to select it as the primary outcome because the main purpose of wearing earplugs and an eye mask is to prevent awakening by external stimuli. Nevertheless, in light of our results, it would be unreasonable to expect a single intervention, namely wearing earplugs and an eye mask, to dramatically improve sleep quantity and quality, which depend on multiple and complex determinants. Recently, Kamdar et al. reported the impact of a multifaceted intervention designed to improve sleep quality in ICU patients [8]. Earplugs and an eye mask were part of this strategy, which also included nighttime interventions to reduce sleep disruption at night and daytime interventions to promote a normal circadian rhythm. Although this multifaceted intervention did not significantly improve sleep quality, it did improve the perceived nighttime noise and the incidence of delirium [8]. In addition, although a correlation has been observed between noise level and arousal frequency in ICU patients [44], a questionnaire administered to patients after ICU discharge showed that other human interventions, such as phlebotomy and measurement of vital signs, were even more sleep-disruptive than noise [45]. Noise and care activities could account for < 15–30% of arousals and awakenings in ICU patients [9, 19].

The use of earplugs and eye masks at night in ICU patents who have awoken from the effects of sedation did not increase the N3 proportion of sleep, but it did decrease the number of prolonged awakenings in ICU patients and also increased N3 duration when these devices were well tolerated. Earplugs and an eye mask had no impact on outcome. A major limitation of this study is the limited willingness of patients to use these devices, which were the source of numerous complaints. It seems that the benefit of wearing earplugs and an eye mask is at least partially counteracted by the discomfort of wearing the devices. Because the improvement of sleep quality in ICU patients remains a major concern, future studies should evaluate multifaceted programs, possibly including protective devices, rather than focusing on a single intervention.