Effects of Dexmedetomidine on Sleep Quality in Critically Ill Patients: A Pilot Study

Sleep in critically ill patients has been recently recognized as an important aspect of the modern intensive care. It has been shown that the quality of sleep in critically ill patients is low, characterized by severe sleep fragmentation, excessive stage 1 of sleep (light sleep) and substantial reduction in rapid eye movement (REM) and slow-wave sleep (SWS).^1,2  In addition, some patients may exhibit atypical patterns, consisting of absence of typical stage 2 of sleep. Thus, although the total sleep duration may be normal or even increased,³  critically ill patients are considered to be qualitatively sleep disrupted. The impaired sleep quality may on long-term cause cardiorespiratory, neurological, immunological, and metabolic consequences, leading to increased morbidity.^4–8 

A widely held belief is that sedation in critically ill patients promotes sleep and thus reduces the detrimental effect of sleep deprivation or poor sleep quality. Indeed, it has been shown that in approximately 50% of mechanically ventilated patients, the dose of γ-aminobutyric acid receptor (GABA) agonists (propofol and benzodiazepines) is increased during the night (11:00 am to 7:00 pm).⁹  Recent data, however, suggest that this strategy may not be beneficial. In critically ill patients ventilated on assisted modes, propofol administration, to achieve the recommended level of sedation, does not increase sleep efficiency, whereas suppressing the REM sleep stage further worsens the poor sleep quality of these patients.¹⁰  These results do not support the use of GABA agonists to facilitate sleep in critically ill patients ventilated on assisted modes.

Dexmedetomidine is a potent α-2-adrenergic agonist, with widespread action on the brain that includes sedation, anesthetic-sparing, analgesia, and sympatholytic properties.¹¹  It has been shown that compared to GABA agonists, dexmedetomidine more closely resembles natural non-REM sleep.^12–15  However, polysomnographic data on critically ill patients are scanty. The only study that performed polysomnography to assess sleep with dexmedetomidine sedation was that of Oto et al.,¹⁶  which concluded that, in mechanically ventilated patients, night-time infusion of dexmedetomidine induced severely disturbed sleep architecture since there was no evidence of SWS and REM sleep. Nevertheless, the absence of the control group and the fact that sedation depth was not tightly controlled ranging from light to severe (−1 to −4 on Richmond Agitation Sedation Scale [RASS]) complicate the interpretation of these results. The aim of this pilot study was to examine sleep quality in critically ill patients with and without dexmedetomidine sedation, titrated to achieve RASS −1 to −2.

The study, performed between October 2011 and November 2013, was approved by the human studies subcommittee (Ethics Committee, University Hospital of Heraklion, Greece), and informed consent was obtained from patients and surrogates. Critically ill patients who had been receiving mechanical ventilation for at least 48 h before study entry and who were anticipated to be in the intensive care unit (ICU) for at least 5 days after study entry were enrolled. At the time of the study, the patients were hemodynamically stable without vasoactive and sedative drugs and were either ventilated on assisted modes of support (pressure support or proportional assist ventilation) or breathed spontaneously (t-piece) through cuffed endotracheal or tracheostomy tubes. In all patients, before the study, a morning blood sample was obtained to measure plasma levels of propofol and midazolam. Exclusion criteria were (1) Glasgow Coma Scale less than 11, (2) acute physiology score portion of the Acute Physiology and Chronic Health Evaluation II (APACHE II) greater than 15,¹⁷  (3) administration of any sedative drugs during the last 24 h, (4) detectable plasma levels of propofol or midazolam, (5) heart rate less than 50 min⁻¹, (6) second- or third-degree atrioventricular block (unless pacemaker installed), (7) severe liver failure (bilirubin >100 μmol/l), (8) use of centrally acting α-2 agonists or antagonists within 24 h before start of the study, (9) history of epilepsy, (10) history of any other neurological disease that may potentially have significant effect on sleep quality, (11) history of sleep apnea, (12) ongoing sepsis, and (13) presence of delirium at the time of the study, as defined by the confusion assessment method for ICU.¹⁸  The mode of support and ventilator settings were determined by the primary physician, who was not involved in the study. Patients who received opioids for analgesia were not excluded, provided that there was no change in the dose throughout the study. Where necessary, nonsteroidal antiinflammatory drugs were used for additional analgesia. Administration of neuroleptic (antipsychotic) medications during the entire study period was a reason for patient withdrawal.

Polysomnography was performed on each patient as previously described.¹⁹  Sleep data were processed automatically (Alice; Respironics, Pittsburgh, PA) according to established rules²⁰  and stored in a computer disc for later analysis. Sleep architecture was scored manually using standard criteria.²⁰  Sleep efficiency was calculated as the ratio between the total sleep time and the total recording time and expressed as percentage.²⁰  Sleep fragmentation index was calculated as the sum of arousals and awakenings per hour of sleep²⁰  (additional details are shown in Supplemental Digital Content 1, https://links.lww.com/ALN/B62). Each patient was studied for a total of 57 h, divided into three night-time (9:00 pm to 6:00 am) and two daytime (6:00 am to 9:00 pm) periods (fig. 1). The entire study started at 9:00 pm and ended 57 h later at 6:00 am. At 9:00 pm on the second night, a loading dose of dexmedetomidine was given (0.5 μg/kg over 20 min), followed by a continuous infusion at a rate adjusted (0.2 to 0.7 μg kg⁻¹ h⁻¹) to maintain a sedation level −1 to −2 on RASS. To prevent bias in dexmedetomidine administration, only two senior intensivists (C.A. and E.K.) were responsible for dexmedetomidine titration and did not have access to electroencephalogram. Continuous infusion was stopped at 6:00 am. Apart from this time interval, the patients did not receive any sedation throughout the study (fig. 1). If there was need for sedation, as judged by the primary physician, the patient was withdrawn from the study. Noise, nursing, and other interventions were minimized during the nights of the study. In addition, during the nights, light was decreased to a minimum level that did not interfere with patients’ assessment. Care was taken to ensure similar environmental conditions among the 3 study nights. In addition, the mode of ventilation and assist level remained unaltered during the study nights. Changes in the mode of ventilation or assist level based on the primary physician’s judgment resulted in patient withdrawal from the study. Patients who required a change in the mode of ventilation during the daytime periods were not excluded from the study but were analyzed only during the night-time.

Arterial blood gasses were measured at the beginning and end of each study night. Heart rate, invasive arterial blood pressure, tidal volume (V_T), and respiratory rate were recorded during the night at half-hour intervals. V_T was recorded only in mechanically ventilated patients as measured by the ventilator.

During the entire study, the patients were attended by dedicated research assistants (ICU nurse). In addition, the proper placement of the polysomnography leads was inspected periodically by one of the researchers.

No a priori power analysis was conducted. The number of studied patients was a compromise between the difficulty in recruitment (due to study design) and meaningful statistical analysis. Data were analyzed using nonparametric tests (SPSS version 17; IBM, Armonk, NY). Continuous variables were expressed as medians (25th to 75th interquartile range). Variables were compared using the Friedman test for repeated measurements followed, when indicated, by a pairwise comparison with the Wilcoxon signed-rank test using Bonferroni post hoc correction. Categorical variables were compared using the Fisher exact test. Statistical tests were two sided, and a P value less than 0.05 was considered statistically significant.

Sixteen patients were studied. All met the definition of difficult or prolonged weaning, having been on mechanical ventilation for several days before the study. None of the patients had detectable levels of midazolam (or any other benzodiazepine) or propofol. Three patients did not complete the study and were excluded from the analysis. The reasons for study interruption were bradycardia (<50 beats/min) during dexmedetomidine infusion in one patient and decision of the primary physician to use sedation with GABA agonists and mode change in the others. Baseline characteristics of the remaining 13 patients are shown in table 1. By study design, all had sleep data during the 3 study nights, and 10 patients had also sleep data during the daytime periods (i.e., no mode change during the daytime periods was performed in these 10 patients). Three out of 13 patients received opioids (transdermal fentanyl 25 μg/h [n = 1], morphine 0.5 mg/h [n = 1], and tramadol hydrochloride 50 mg every morning [n = 1]) for analgesia and/or respiratory distress.

The starting maintenance infusion dose of dexmedetomidine was 0.2 μg kg⁻¹ h⁻¹ (0.2 to 0.2) and increased to 0.6 μg kg⁻¹ h⁻¹ (0.4 to 0.7) to achieve a score of −1 to −2 on RASS. APACHE II score (calculated before study nights) did not differ significantly among nights (14 [13 to 17], 15 [13 to 18], 14 [12 to 17], P = 0.161). Throughout the entire night, arterial blood gasses, V_T, respiratory rate, and arterial blood gasses remained stable and without significant differences among nights (see these results in tables 1 and 2, Supplemental Digital Content 2, https://links.lww.com/ALN/B63). The maximum dose of dexmedetomidine was associated with a significant decrease in heart rate (see table 3, Supplemental Digital Content 2, https://links.lww.com/ALN/B63).

Sleep efficiency differed significantly among the 3 study nights; compared to nights without dexmedetomidine infusion, sleep efficiency was significantly higher during the night with dexmedetomidine (fig. 2). Six patients during the first night and seven patients during the third night had sleep efficiency less than 10%, whereas in none during the second night (dexmedetomidine infusion) sleep efficiency was less than 10% (P = 0.006, Fisher exact test). One patient did not achieve sleep during the second or third night. Overall, three and four patients, during the first and third night, respectively, did not achieve sleep (see table 4, Supplemental Digital Content 2, https://links.lww.com/ALN/B63, for sleep architecture during the three study nights).

Because in patients who achieved sleep both during the first and third night (n = 7), sleep architecture did not differ between these 2 nights, sleep architecture data were analyzed as follows. The sleep stages, expressed as percentage of total sleep time, and sleep fragmentation index were averaged during the 2 nights (first and third) in patients in whom sleep was observed on both nights (n = 7), whereas in patients in whom sleep occurred only on one of these nights (n = 5) the corresponding single value was used for analysis. Thus, each patient (n = 12) had sleep architecture data pertained to nights without dexmedetomidine, which was compared to that with dexmedetomidine. With dexmedetomidine, sleep fragmentation index and stage 1 of sleep (N1) were significantly lower, whereas stage 2 (N2) was higher than the corresponding values observed during the nights without dexmedetomidine (figs. 3 and 4). SWS (N3) and REM sleep did not differ with or without dexmedetomidine.

In 10 of 13 patients polysomnography was performed for two consecutive 24-h periods (first and second periods). Sleep efficiency did not differ during the daytime periods before and after dexmedetomidine (14.3% [6.5 to 25.5] vs. 17.3% [5.8 to 28.0], P = 0.695). All patients achieved sleep during the daytime period before (first day) and nine of them achieved sleep after dexmedetomidine (second day). Compared to the previous night (first night), sleep efficiency during the first day did not differ significantly. Sleep efficiency during the second day was significantly lower than that during the previous night when dexmedetomidine was given (second night) (see these data in table 5, Supplemental Digital Content 2, https://links.lww.com/ALN/B63). Compared to sleep architecture during the first day (before dexmedetomidine infusion), sleep architecture during the second day (after dexmedetomidine infusion) was characterized by less stage 1 and more stage 2 and REM sleep, the difference, however, being nonsignificant (see fig. 1, Supplemental Digital Content 3, https://links.lww.com/ALN/B64). Sleep was equally distributed between day and night in the first period (no dexmedetomidine infusion during the night), while it mainly occurred during the night in the second period (dexmedetomidine infusion during the night); 48% (32 to 71) of total sleep time occurred during the night in the first period and 79% (66 to 87) in the second (P = 0.032).

The main findings of our study were that in critically ill patients, dexmedetomidine infusion during the night titrated to achieve light sedation (RASS −1 to −2) significantly increased sleep efficiency and improved sleep architecture by reducing the sleep fragmentation and stage 1 and increasing the stage 2 of sleep. Furthermore, our study showed that without sedation, sleep was equally distributed between day and night, a pattern that was modified by night-time dexmedetomidine infusion with more than three quarters of sleep occurring during the night.

The RASS was used for titration of dexmedetomidine administration.^21,22  This scale has been validated properly, has excellent interrater reliability, is relatively simple, and is used extensively in ICUs worldwide. Furthermore, in our study, to avoid errors in sedation depth, the titration of dexmedetomidine was performed by two senior intensivists (C.A. and E.K.). In addition, the physicians responsible for titration did not have access to electroencephalogram, which guarantees that the titration of dexmedetomidine followed the usual clinical practice (i.e., use of sedation scale) and was not influenced by electroencephalogram.

Our study reconfirmed previous studies showing that critically ill patients exhibit disorganized and poor quality sleep as evidenced by the lack of sequential progression through sleep stages and low percentages of SWS and REM sleep.^2,3,10,19  Dexmedetomidine infusion at doses titrated to achieve the recommended sedation depth in critically ill patients increased sleep efficiency and improved the sleep architecture by decreasing the stage 1 and increasing stage 2 of sleep. Notwithstanding that few patients exhibited SWS and REM, dexmedetomidine did not affect SWS and REM sleep. In addition, sleep became less fragmented with dexmedetomidine. These results contradict those obtained with the GABA agonist propofol, which worsened sleep quality because the drug virtually eliminated REM sleep with no effect on sleep efficiency, sleep fragmentation, and stage 1 and 2 of sleep.¹⁰  Nevertheless, the fact that dexmedetomidine sedation does not increase the most restorative sleep stages (SWS and REM)²³  indicates that sleep quality of critically ill patients remains low even during sedation with this agent. It is not known, however, if longer or more often (i.e., every night) administration of this sedative might further influence sleep quality in these patients.

Recently, Oto et al.¹⁶  infused dexmedetomidine during the night in mechanically ventilated patients and concluded that the drug induced severely disturbed sleep architecture without evidence of SWS and REM. However, in this study, night-time sleep without dexmedetomidine was not studied. In addition, daytime sleep before dexmedetomidine infusion was also characterized by the absence of SWS and REM. Our study did not show any evidence of dexmedetomidine-induced sleep disruption in critically ill patients; compared to no sedation, dexmedetomidine did not affect SWS and REM sleep, and on the other hand, sleep became deeper (increased stage 2, decreased stage 1) and less fragmented. Also, daytime and night-time data after dexmedetomidine infusion did not show any hard evidence of REM rebound phenomenon, although there was a trend for increased REM after sedation. Finally, it is of interest to note that sleep architecture with dexmedetomidine is quite similar to that reported by Oto et al. (relatively high sleep efficiency and stage 2 of sleep and low stage 1 and SWS, see table 2 in Oto et al.¹⁶  study).

It is well documented that circadian rhythm is abolished in critically ill patients due to pronounced temporal disorganization or disturbances in melatonin secretion.²⁴  Our study showed that dexmedetomidine may preserve the day–night cycle of sleep, partly restoring the normal circadian rhythm. Compared to a 24-h period with no sedation in which sleep was equally distributed between day and night, dexmedetomidine infusion during the night resulted in a night-time shift of sleep; when dexmedetomidine was infused during the night, less than one-quarter of the total 24 h of sleep occurred during the day. However, we believe that these findings are largely due to study design (i.e., administration of dexmedetomidine during the night) since there is no known reason to expect different effect of dexmedetomidine on sleep if the medication would be given during daytime (although this issue should be formally studied).

Our results may have clinical implications. Studies have shown that delirium and posttraumatic stress disorder, common morbidities in critically ill patients and ICU survivors, may be linked to sleep abnormalities during the ICU stay.²³  In addition, GABA agonists, which worsen sleep in these patients,¹⁰  are an independent risk factor for both delirium and posttraumatic stress disorder.^9,25,26  On the other hand, some studies have demonstrated that compared to GABA agonists, dexmedotimidine may be associated with a decreased incidence of delirium,^27–29  and this might be partly explained by the achievement of better sleep, as our study demonstrated. Nevertheless, further studies are needed to clarify these issues.

This study has several limitations. First, in order to facilitate sleep classification using the standard criteria,²⁰  a highly selected group of critically ill patients were studied. These criteria have not been developed for critically ill patients in whom atypical sleep pattern, such as electroencephalogram features of SWS and absence of typical stage 2 sleep (i.e., absence of K complexes and sleep spindles), may be observed.³⁰  In our study, patients with conditions known to be associated with this pattern (i.e., ongoing sepsis) were excluded. Thus, the results of this study should be applied with caution to a general population of critically ill patients in whom acute critical illness may cause encephalopathy and atypical sleep patterns.³⁰  Second, because of the study design, patients not requiring sedation were studied. In patients needing sedatives for various reasons (i.e., agitation, patient–ventilator dyssynchrony), the results might be different, and this is also certainly a limitation. Third, because of the strict selection criteria we applied, the recruitment rate was very low (annual rate of ICU admission approximately 500 patients), and thus, the number of patients studied over the 2-yr period was small. Nevertheless, the effect of dexmedetomidine on sleep was rather consistent between patients, partly overcoming this limitation.

In conclusion, this pilot study shows that night-time dexmedetomidine administration to achieve the recommended level of sedation in highly selected critically ill patients increases sleep efficiency and improves sleep quality by reducing the sleep fragmentation and shifting the sleep from stage 1 to stage 2. In addition, dexmedetomidine modifies the 24-h sleep pattern by shifting sleep mainly to the night, partly restoring normal circadian rhythm.

This study was supported by a grant from Orion Pharma (Espoo, Finland) through research contract with the University of Crete, Heraklion, Crete, Greece. Dexmedetomidine was supplied by Orion Pharma. Orion Pharma was not involved in any aspect of the design or conduct of the study, the data analysis, or the article preparation and presentation.

Dr. Georgopoulos received lecture fee (honoraria) from Orion Pharma (Espoo, Finland). The other authors declare no competing interests.