The effect of earplugs during the night on the onset of delirium and sleep perception: a randomized controlled trial in intensive care patients

Delirium is a common complication in the ICU caused by a malfunction of the cognitive processes in the brain. The syndrome is characterized by a fluctuating course, shifting attention, disorganized thinking and a changed level of consciousness [1]. Incidences from 20% to more than 80% are reported in different patient groups using different assessment tools. Predisposing and precipitating risk factors related to patient characteristics, chronic pathology, acute illness and the environment have been studied [2, 3]. A patient encountering three or more of these factors has a 60% increased risk for the development of delirium [2, 4]. Ely et al. stated that a patient in the ICU even accumulates ten or more of these factors [5].

Delirium often presents early after admittance to the ICU. The early onset is probably caused by an acute change in the physical situation of the patient stressed by a sensory overload. A few days after admission a cognitive healthy patient may shift to a delirium due to underlying biomedical changes or worsening illness. In this context, delirium may be called the sixth vital sign [6, 7]. Next to delirium, confusion is mentioned as a symptom in different psychiatric and cognitive disorders and is described as 'a state of disturbed orientation in regard to time, place or person, affecting the clarity and the coherence of one's thinking' [8]. Consequently, patients classified as confused have an altered perception or thought but may score negative for delirium.

The ICU is a rapidly changing ward designed to admit severely ill patients. The typical character and the health care process in this unit induce heavier care sustained by high technological equipment. This equipment and the higher intensity of care also produce augmented sound levels [9]. Sound in the ICU has been a subject of research for years. Peak noise is not the main determinant disturbing the patient in the ICU. Phones ringing and people talking are reported as more annoying [10]. Although often suggested, there is ample evidence that sound influences the patient's outcome. Most studies on noise report on a possible relation with sleep or on results of architectural improvements [11‐14]. The quality of sleep in the ICU, however, has been related to environmental sound [10‐12, 15, 16]. Moreover, the impact of disturbed sleep on the onset of delirium in the ICU has been proposed. Several studies showed severe fragmentation, arousals and awakenings in the sleep of ICU patients and pointed at the absence of slow wave and REM sleep. Researchers hypothesized that this disturbance of sleep could be an important role player in the onset of the delirious syndrome [17, 18].

Although the impact of sleep on the onset of delirium has often been suggested, sound influencing sleep has not been identified as a risk factor for delirium yet. We hypothesized that a reduction of sound during the night using earplugs could be beneficial in the prevention of the early onset of intensive care delirium. Two primary research questions were formulated. First, does the use of earplugs during the night reduce the onset of delirium in the ICU? Second, does the use of earplugs during the night improve the quality of sleep in the ICU?

This study was a randomized clinical trial in which adult intensive care patients were assigned in a 1:1 ratio to an intervention group, patients sleeping with earplugs during the night, or a control group, patients sleeping without earplugs during the night. The researchers were blinded during data collection. Assignment to the study or control group was done by an independent nurse researcher using a computer program. Since the focus of this study was the early onset of delirium, patients were to be observed during a maximum of five nights. Earlier research in the same setting showed that most delirium cases presented in the first 72 hours after admission to the ICU [3, 19, 20]. Patients scoring positive for delirium were censored for further observation and analysis.

During the study period 221 patients were found to be eligible in four ICU-subdivisions. After being informed, 46 patients, 36 women and 10 men, refused further cooperation. An early drop out before the randomization was caused by an unexpected length of stay of less than 24 hours (n = 13) or severe worsening of the patient's condition or illness (n = 24). Consequently, these patients, not meeting the inclusion criteria, were not randomized. Additionally, two patients died before the first assessment of delirium and sleep perception. The study population comprised 136 patients, 69 in the study group and 67 in the control group (Figure 1). The mean age of the population was 59 years (range18 to 84), 66% were men. The mean SAPS 3 was 42.3 (0 to 78), the mean SOFA score during the first 24 hours was 7.1 (1 to 14) and the mean TISS 28 score was 24.5 (9 to 43). The patients using earplugs had a significantly longer observation period than the control group (43 hours versus. 33 hours, P = 0.02). During the maximum observation of five days, 20% of the patients were delirious and 27% showed mild confusion on at least one observation moment. Additionally, the NEECHAM assessments showed 23% of the patients were at risk for delirium and 30% were classified as normal. Most included patients stayed only one night in the ICU. Both study groups were comparable at baseline as few statistical differences were found between the study and the control groups (Table 1).

The use of earplugs during the night proved to be beneficial in our study group. Fewer patients showed delirium or confusion. A vast improvement was shown in the NEECHAM category 'mild confusion'. Moreover, the onset of cognitive disturbances was delayed compared to the patients sleeping without earplugs. Additionally, patients sleeping with earplugs who developed delirium or confusion did not suffer from the syndrome as soon as patients sleeping without earplugs. More patients reported a better sleep perception after sleeping with earplugs in the ICU.

The imbalance of sample size after day one is mainly due to the higher number with 'first signs of delirium' in the control group resulting in the end of observation for these patients. The patients in the control group showed an earlier onset of delirium resulting in a smaller population to study. This resulted in a difference of observation time as presented in Table 1. These findings, however, are not to be considered as a reflection of the actual length of stay in the ICU.

Delirium induced by environmental and sensorial factors appears early during admission to the ICU. A later onset of the syndrome is probably caused by changes in illness, hemodynamic or biomedical situation or ongoing treatment [6, 7]. Most researchers describe delirium in the ICU as a multifactorial syndrome [1, 3]. Although being the same syndrome, early stage delirium may be different from a 'second period' delirium induced by severity of illness. The focus of our research was on the early onset of the syndrome by observing only the first days after admission of the patient. Known risk factors for delirium were not different for the study and the control groups (Table 1). No differences were observed between the two groups for the known risk factors of delirium.

Since the initial development of the NEECHAM scale, patients are divided into four categories (delirium, mild confusion, at risk and normal). We studied the NEECHAM scale in relation to the CAM-ICU in earlier research. Both scales proved to be comparable in the detection of delirium [19]. Delirium is a syndrome well-defined by the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV-TR) criteria [1]. The category mild confusion, however, has been mentioned by several authors without a clear definition [27, 28]. The clinical relevance of the NEECHAM category 'confusion' has not been studied thoroughly yet, but may be considered as a prelude to delirium. Hard outcomes for 'mild confusion' have not been studied yet in an intensive care population. More patients remained normal or at risk using the earplugs at night. The category mild confusion was more present in the control group. Consequently, this study pointed to a possible relevance of this category for the first time. Further research needs to focus on the clinical relevance of this category. Meanwhile, it is advisable to observe patients scoring 'confusion' in order to receive focused care attempting to prevent delirium.

The incidence of delirium, however, was not different for both groups. Since it is hard to believe that the exclusion of a single risk factor resolves delirium, the use of earplugs is no magical solution in the prevention of the syndrome. Patients in the study group were triggered by other factors to evoke the confusion or delirium. The multivariate model showed age, smoking and severity of disease to be important role players in this population also. This confirms earlier research [3, 5]. The clearly beneficial effect of the use of earplugs, however, is strong enough to advise their use during the night in the ICU. Moreover, patients using earplugs developed delirium or confusion later during their stay in the ICU. A protection in the early stage of the admission to the unit was therefore demonstrated. The effect of mild confusion on the patient's transition into delirium has not been studied yet. Therefore, our interventions must be situated in the prevention of the early onset delirium. Although the harmful outcome of delirium has only been proven for the worst stage, it can also be advised to consider the use of earplugs in the prevention of the early stage of confusion. Mistraletti et al. pointed already at the possible profit in the prevention of delirium in improving patient's sleep [18]. This was not proven in our research yet. The beneficial outcome may not be completely studied yet, but it seems obvious that applying earplugs to all patients favors sleep. A larger scale use may be recommended while the outcome of this improved sleep perception can be studied in a larger design.

Poor sleep has been shown in intensive care patients [11, 18]. Polysomnography is the golden standard to assess sleep objectively. Since this tool is expensive and very labor intensive, large scale studies are rare and the implementation of polysomnography in a major study to assess sleep seems hard to manage. An objective assessment of sleep, however, is needed in the search for poor sleep as a risk factor for delirium. Self-reported sleep perception as the subjective self-reported assessment of sleep quality is easier to study. Therefore, sleep perception is easier to study as a risk factor but shows some important limitations. Validated scales to assess sleep perception were tested but seemed to create a burden on the intensive care patient because they were too long and required a lot of attention. Therefore, questions were simplified to have an easy response from the patients. Not being validated, the results of the questions must be considered as indicative.

Earlier research showed a development of sound modification programs based on architectonical, structural or staff behavioral interventions. Conversation between staff seems one of the major sources of noise. Therefore, a staff education program could already affect 14% of the peak sound sources [9, 29]. Most likely, the architectural structure based on closed rooms reduces most effectively the sound at the bedside as described by Gabor et al. [15].

Study patients in this trial reported a better sleep perception due to the noise reduction by earplugs. Hardly any studies have been performed with earplugs in the ICU. On the other hand, some potential beneficial effects providing a reasonable basis for testing the effects of earplugs in critically ill subjects were reported. These cheap devices are capable of reducing the incidence of intraoperative awareness with recall during elective orthopedic surgery [30]. One small randomized study in the neonatal ICU showed a significant effect of silicone earplugs on weight gain in 'very low birth weight' (< 1,500 g) and even better outcomes in 'extremely low birth weight' (< 1,000 g) newborns [31]. Earplugs worn by healthy volunteers during exposure to noise levels as observed in the ICU produce a significant decrease in REM sleep latency and an increase in the percentage of REM sleep [32]. The use of earplugs and eye masks together resulted in improvement of polysomnographic variables such as more REM time, shorter REM latency, less arousal and elevated melatonin levels in a limited group of healthy subjects exposed to recorded ICU noise and light together [33]. The latter study in volunteers, however, explored the effects of both ICU noise and light. We are aware of only one previous study exploring the use of earplugs alone in a real ICU environment. Scotto et al. were able to improve the subjective total sleep satisfaction score in non-ventilated, non-sedated adults after the use of earplugs in the ICU but did not explore the effect on delirium [34]. The use of earplugs, however, is cheap, easy and has apparently the same effect on all patients without the necessity to introduce more extensive structural or organizational changes on the ward.

Our study has some limitations. This randomized controlled trial included a specific population in our ICU. Therefore, results may not be applicable to all settings and all patients. Moreover, including this specific population, the findings seem to focus on the first 24 hours of admittance. Larger research may focus on the total length of stay in the ICU of all patients. No accidental or intentional removal of the earplugs was reported. All included patients agreed to sleep with earplugs. Patients who did not like to use earplugs could not give their consent to the study. Including the patients for the trial, the larger group of refusals were women. They indicated that they prefer remaining in direct contact with their environment. Further research could focus on the reasons for this refusal. Also, many patients stayed only one night in the ICU. Consequently, a short term effect of the use of earplugs was studied. A study on the longer term outcome must be included in a larger scale project. Also, the fact of being delirious makes it impossible for patients to report on sleep perception. Other tools must be searched for to study this perception in delirious patients specifically. At that time a validated easy-to-use scale for sleep perception in the ICU may become available.

Despite the fact that the relationship between sleep perception and delirium has not been clearly established, this study pointed at a relation between environmental sound, sleep perception and delirium. The NEECHAM Confusion Scale showed a significantly lower proportion of patients with mild confusion or delirium in the study group sleeping with earplugs during the night in the ICU. Also, patients reported a better sleep perception using earplugs. Earplugs may be a useful instrument in the prevention of confusion or delirium. The beneficial effects seem to be strongest within 48 hours after admission. The relationship between sleep, sound and delirium, however, needs further research.