Abstract
Patients admitted to the intensive care unit (ICU) after surgery often develop sleep disturbances. The authors tested the hypothesis that low-dose dexmedetomidine infusion could improve sleep architecture in nonmechanically ventilated elderly patients in the ICU after surgery.
This was a pilot, randomized controlled trial. Seventy-six patients age 65 yr or older who were admitted to the ICU after noncardiac surgery and did not require mechanical ventilation were randomized to receive dexmedetomidine (continuous infusion at a rate of 0.1 μg kg−1 h−1; n = 38) or placebo (n = 38) for 15 h, i.e., from 5:00 pm on the day of surgery until 8:00 am on the first day after surgery. Polysomnogram was monitored during the period of study-drug infusion. The primary endpoint was the percentage of stage 2 non–rapid eye movement (stage N2) sleep.
Complete polysomnogram recordings were obtained in 61 patients (30 in the placebo group and 31 in the dexmedetomidine group). Dexmedetomidine infusion increased the percentage of stage N2 sleep from median 15.8% (interquartile range, 1.3 to 62.8) with placebo to 43.5% (16.6 to 80.2) with dexmedetomidine (difference, 14.7%; 95% CI, 0.0 to 31.9; P = 0.048); it also prolonged the total sleep time, decreased the percentage of stage N1 sleep, increased the sleep efficiency, and improved the subjective sleep quality. Dexmedetomidine increased the incidence of hypotension without significant intervention.
In nonmechanically ventilated elderly patients who were admitted to the ICU after noncardiac surgery, the prophylactic low-dose dexmedetomidine infusion may improve overall sleep quality.
Dexmedetomidine sedation can improve sleep architecture in mechanically ventilated patients, but its effectiveness in nonventilated patients is unknown.
In a pilot trial of 76 adults (age, greater than 65 yr) admitted to the intensive care unit after noncardiac surgery, patients were randomized to receive low-dose dexmedetomidine infusion versus placebo for 15 h. Dexmedetomidine increased stage N2 (and decreased N1) sleep, total sleep time, sleep efficiency, and subjective sleep quality.
SLEEP disturbances, including sleep deprivation, disruption, and abnormal architecture, are prevalent in postsurgical patients, particularly in those who are admitted to the intensive care unit (ICU).1,2 Previous studies reported that poor sleep was one of the most frequent complaints among patients who recovered from critical illness.3–8 Studies using polysomnography showed that the sleep pattern of ICU patients was characterized as disorganized circadian rhythm, prolonged sleep latencies, fragmented sleep, decreased sleep efficiency, abnormally increased stages 1 and 2 of non–rapid eye movement (NREM) sleep (also called stage N1 and N2 sleep), and decreased or absent stage 3 of NREM sleep (also called stage N3 sleep or slow-wave sleep) and rapid eye movement (REM) sleep.9–13 The causes of sleep disturbances in the ICU are multifactorial, including critical illness, surgery, mechanical ventilation, medications, and the ICU environment.9–11,14–17 The occurrence of sleep disturbances can produce significant adverse consequences such as immune system compromise, delayed weaning from mechanical ventilation, cardiovascular events, development of delirium, and post-ICU physical and mental health decline.18–25
Given the importance of good sleep for patients’ recovery from critical illness and major surgery, multiple nonpharmacologic interventions have been implemented to improve patients’ sleep quality in the ICU, such as elimination of unnecessary noise and light, consolidation of patient care interactions, use of earplugs and eye masks, relaxation techniques, and addition of white noise.26–30 However, the effects of these strategies are limited, and adjunctive drug therapy is often needed in some circumstances.31 Sedatives and analgesics, such as propofol, benzodiazepines, and opioids, are frequently used in ICU patients in order to increase their comfort and sleep quality. Unfortunately, these drugs may produce sleep architecture disruption, which could potentially contribute to other postsurgical complications, including delirium.32,33
Dexmedetomidine, an α2 adrenoceptor agonist with both sedative and analgesic properties, has increasingly been used in ICU patients.34 Unlike other sedative agents, dexmedetomidine exerts its sedative effects through an endogenous sleep-promoting pathway and preserves sleep architecture to some degree in the preclinical settings.35,36 In a recent study of mechanically ventilated ICU patients, nighttime infusion of sedative dose of dexmedetomidine (median, 0.6 μg kg−1 h−1 [interquartile range {IQR}, 0.4 to 0.7]) that induced a sedation level from −1 to −2 on the Richmond Agitation Sedation Scale (RASS) helped preserve the day–night cycle of sleep and improved the sleep architecture by increasing sleep efficiency and stage N2 sleep.37 A similar phenomenon was also observed previously.38 We hypothesized that nonsedative dose of dexmedetomidine might also improve sleep architecture in nonmechanically ventilated postoperative patients. The purpose of this study was to investigate the effect of low-dose dexmedetomidine infusion on the sleep architecture of nonmechanically ventilated elderly patients who were admitted to the ICU after noncardiac surgery.
Materials and Methods
Study Design
This was a randomized, double-blind, placebo-controlled, parallel-arm pilot trial designed to assess the superiority of the intervention, i.e., a prophylactic, low-dose dexmedetomidine infusion. The study protocol was approved by the Clinical Research Ethics Committee of Peking University First Hospital (Beijing, China; 2012-515) and registered on Chinese Clinical Trial Registry (ChiCTR-TRC-12002567). Written informed consent was obtained from each patient or, if the patient could not provide informed consent, from the surrogate of the patient. The study was conducted in the surgical ICU of Peking University First Hospital, Beijing, China.
Patient Recruitment
Potential participants were screened on ICU admission by two qualified investigators (X.-H.W. and F.C.). Elderly patients (age 65 yr or over) who underwent noncardiac surgery during general anesthesia and were admitted to the surgical ICU before 5:00 pm were included in the study. Patients were excluded if they met any of the following criteria: (1) history of schizophrenia, epilepsy, or parkinsonism; (2) history of sleep disorders (requirement of hypnotics/sedatives during the last month); (3) history of obstructive sleep apnea syndrome; (4) preoperative sick sinus syndrome, severe sinus bradycardia (heart rate less than 50 beats/min), or atrioventricular block of second degree or above without pacemaker; (5) preoperative coma; (6) brain injury or neurosurgery; (7) serious hepatic dysfunction (Child-Pugh class C); (8) serious renal dysfunction (undergoing dialysis before surgery); or (9) requirement of mechanical ventilation. Detailed information, including baseline demographic data, preoperative medical history, admission diagnosis, severity of illness, and perioperative variables, was obtained after recruitment.
Randomization, Study-drug Administration, and Procedures
Simple randomization was performed. Random numbers were generated in a 1:1 ratio using the SAS 9.2 software (SAS Institute, USA). Study drugs (either 200 μg/2 ml dexmedetomidine hydrochloride or 2 ml normal saline) were provided as clear aqueous solution in the same type of 3-ml volume bottles (manufactured by Jiangsu Hengrui Medicine Co, Ltd, China) and encoded according to the randomization results before the study by a pharmacist who did not participate in the rest of the study. The results of randomization were sealed in sequentially numbered envelopes until the end of the study.
During the study period, consecutively enrolled patients were randomly assigned to receive either dexmedetomidine or placebo by a study coordinator (Z.-T.M.) who distributed the study drugs accordingly. The study drugs were diluted with normal saline to 50 ml and administered intravenously at an infusion rate of 0.025 ml kg−1 h−1 (0.1 μg kg−1 h−1 of dexmedetomidine in the dexmedetomidine group) for 15 h, i.e., from 5:00 pm on the day of surgery to 8:00 am on the first day after surgery.
All study personnel, healthcare teams, and patients were unaware of treatment group assignment. In case of an emergency (e.g., unexpected, rapid deterioration of the patient’s clinical condition), attending intensivists who were on duty could request unmasking of blinding and/or stopping of the study-drug infusion. In such a case, a statement had to be made in the case report form. These patients were included in the final sleep architecture analyses.
Routine postoperative analgesia was administered including patient-controlled epidural analgesia (established with 250 ml ropivacaine, 0.12%, plus 0.5 μg/ml sufentanil, programmed to deliver a 2-ml bolus with a lockout interval of 20 min and a background infusion of 4 ml/h) or patient-controlled intravenous analgesia (established with 100 ml morphine, 0.5 mg/ml, programmed to deliver a 2-ml bolus with a lockout interval of 6 to 10 min and a background infusion of 1 ml/h). Supplemental morphine (2 mg injected intravenously with a 10-min interval for up to five times per hour) was administered when necessary. For patients without contraindication, flurbiprofen axetil (50 mg injected intravenously for up to two times per day) could be administered.
Polysomnographic Monitoring
Polysomnography was performed with a Compumedics E Series EEG/PSG Recording System (Compumedics Pty Ltd, Australia) from 5:00 pm to 8:00 am, i.e., throughout the duration of study-drug infusion. Two qualified investigators (X.-H.W. and F.C.) were responsible for attaching the electrodes to the patients. The polysomnogram included four-channel electroencephalogram (F3/A2, F4/A1, C3/A2, and C4/A1), two-channel electrooculogram (E1/A2 and E2/A1), and one-channel chin electromyogram (Chin1–Chin2). These data were processed automatically according to the American Academy of Sleep Medicine manual39 and stored in a computer disc. Electrocardiogram, invasive blood pressure, and pulse oxygen saturation were also continuously monitored.
Sleep architecture was scored manually (epoch by epoch) using standard criteria39 by a qualified sleep physician (C.Z.) who was blinded to the study protocol and did not participate in data collection and patient care. It was divided into wakefulness, NREM sleep, and REM sleep. NREM sleep was further divided into three stages, i.e., stage 1 (N1), stage 2 (N2), and stage 3 (N3). Total sleep time was defined as the sum of time spent in any sleep stage during the monitoring period. Sleep efficiency was calculated as the ratio between the total sleep time and the total recording time and expressed as percentage. The percentages of each sleep stage were calculated as the durations of each sleep stage divided by the total sleep time. Sleep fragmentation index was calculated as the average number of arousals and awakenings per hour of sleep.
Other Outcome Assessments
Sedation level was assessed using the RASS at 8:00 am before the end of study-drug infusion.40,41 Subjective pain scores at rest and with movement were assessed using the numeric rating scale (an 11-point scale, where 0 indicates no pain and 10 indicates the worst pain) at 3, 6, and 24 h after surgery. The subjective score of sleep quality was assessed at 8:00 am on the first, second, and third days after surgery using the numeric rating scale as well (an 11-point scale, where 0 indicates the best possible sleep and 10 indicates the worst possible sleep).42,43
Postoperative delirium was assessed using the confusion assessment method44,45 twice daily (at 8:00 am and 6:00 pm, respectively) during the first 7 days after surgery (appendix 1). Adverse events (including hypotension, bradycardia, hypertension, tachycardia, desaturation, and respiratory depression) were monitored until 24 h after surgery or until their disappearance (appendix 2). Intervention for hypotension included interruption of study-drug infusion, intravenous fluid bolus, and/or administration of medication. Intervention for bradycardia, hypertension, and tachycardia included interruption of study-drug infusion and/or administration of medication. Intervention for desaturation and respiratory depression included interruption of study-drug infusion, administration of oxygen, physical therapy, and/or noninvasive/invasive ventilation.
Postoperative complications were defined as any medical condition that required therapeutic intervention. The occurrence of postoperative complications was monitored twice daily during the first 7 days and then weekly until 30 days after surgery. Lengths of stay in the ICU and in hospital after surgery were documented. All-cause 30-day mortality was recorded.
Before the study, investigators (X.-H.W. and F.C.) who would perform patient recruitment and follow-up were trained to follow the study protocol according to the principles of good clinical practice. They were also trained to attach electrodes for polysomnographic monitoring by a sleep physician and to use the confusion assessment method by a psychiatrist.
Statistical Analysis
Sample Size Calculation.
Previous studies showed that in ICU patients requiring mechanical ventilation, sedation with benzodiazepines and/or opioids resulted in a percentage of stage N2 sleep of 40 ± 23% during the nighttime,11 whereas sedation with dexmedetomidine increased the percentage of stage N2 sleep to about 71%.38 We assumed that in postoperative ICU patients without mechanical ventilation, intravenous infusion of low-dose dexmedetomidine would increase the percentage of stage N2 sleep by 20% compared to placebo, with an SD of 23% in both groups. The calculated sample size that would provide 90% power to detect this difference based on a two-tailed significance level of 0.05 was 29 patients per group. Considering a dropout rate of about 25%, we intended to enroll 38 patients in each group.
Outcome Analyses.
Primary endpoint was the percentage of stage N2 sleep. Secondary endpoints included total sleep time, sleep efficiency, durations, and percentages of stage N1, N3, and REM sleep; sleep fragmentation index; and subjective sleep quality. Additional outcomes included the occurrence of postoperative complications, lengths of stay in ICU and hospital after surgery, and all-cause 30-day mortality.
Continuous variables were analyzed with independent samples Student’s t test or Mann–Whitney U test. Categorical variables were analyzed with chi-square analysis, continuity correction chi-square test, or Fisher exact test. Time-to-event outcomes were analyzed with Kaplan–Meier survival analyses, with differences between groups assessed by the log-rank test. The difference between two medians and 95% CI were estimated using the methodology of Hodges–Lehmann. In a post hoc adjustment, any baseline or perioperative factors that differed between the two groups (P < 0.10) together with the intervention factor (the administration of dexmedetomidine or placebo) were reanalyzed with the multivariate linear regression analysis against the primary outcome, in order to figure out the potential impact of confounding factors (if any) on the N2 sleep outcome. Statistical analyses were performed using SPSS 14.0 software (SPSS, USA) and SAS 9.2 software (SAS Institute). All tests were two sided, and a P value less than 0.05 was considered to be statistically significant.
Results
Patient Population
Between November 2012 and June 2013, 193 patients were screened for eligibility, 91 patients met the inclusion/exclusion criteria, and 76 patients were recruited into the study. During the period of study-drug infusion, polysomnographic monitoring failed in 10 patients because of electrode detachment or allergic reaction to the electrode paste (six in the placebo group and four in the dexmedetomidine group). Polysomnographic data were unanalyzable in five patients because of signal interference (two in the placebo group and three in the dexmedetomidine group). Data of these patients were excluded from the sleep architecture analyses (fig. 1). No blinding was unmasked during the study period.
The percentages of patients with preoperative hypertension among all enrolled participants as well as those who were included in the sleep architecture analyses were lower in the dexmedetomidine group than in the placebo group (P = 0.043 and 0.024, respectively). Other baseline and perioperative variables were comparable between the two groups (tables 1 and 2). The majority of patients (82.9%) underwent intraabdominal surgery with a mean (± SD) duration of 2.9 (± 1.5) h.
Sleep Architecture Analyses
Analyses of polysomnograms in the placebo group patients showed severe abnormal sleep architectures, i.e., shortened total sleep time, lowered sleep efficiency, increased percentage of stage N1 sleep, decreased percentages of stage N2 and N3 sleep, absent REM sleep, and high sleep fragmentation index (table 3 and fig. 2A).
Low-dose dexmedetomidine infusion increased the percentage of stage N2 sleep from median 15.8% (IQR, 1.3 to 62.8) in the placebo group to 43.5% (IQR, 16.6 to 80.2) in the dexmedetomidine group (median difference, 14.7%; 95% CI, 0.0 to 31.9; P = 0.048). When patients after intraabdominal surgery were analyzed separately, the percentage of stage N2 sleep was increased from 14.6% (1.4 to 60.5; n = 27) in the placebo group to 44.2% (17.2 to 78.4; n = 24) in the dexmedetomidine group (median difference, 16.2%; 95% CI, 3.9 to 38.1%; P = 0.019). Comparison between the two groups also showed that the total sleep time and the duration of stage N2 sleep were longer (P = 0.028 and 0.038, respectively) and the sleep efficiency was higher (P = 0.033) in the dexmedetomidine group than in the placebo group. Although the duration of stage N1 sleep was not significantly different between the two groups, the percentage of stage N1 sleep was lower in the dexmedetomidine group than in the placebo group (P = 0.038). Stage N3 sleep was present in 16.7% (5/30) of patients in the placebo group and in 25.8% (8/31) of patients in the dexmedetomidine group (P = 0.384). REM sleep was absent in the patients of both groups (table 3 and fig. 2B).
Since the stage N2 sleep percentage data were not normally distributed, a square root transformation was performed, and the transformed results fit normal distribution. After adjustment with the existence of preoperative hypertension and the duration of anesthesia using the multivariate linear regression analysis, low-dose dexmedetomidine infusion remained the factor that had a significant association with the increased percentage of stage N2 sleep (least-squares mean, 2.33; 95% CI, 0.62 to 4.05; P = 0.008).
Other Outcome Analyses
The subjective scores of sleep quality on the first postoperative morning of both all enrolled patients and those who were included in the sleep architecture analyses were lower in the dexmedetomidine group than in the placebo group (P = 0.004 and 0.005, respectively). But no significant differences were found between the two groups with regard to the subjective scores of sleep quality on the second and third postoperative mornings and the subjective pain scores both at rest and with movement at 3, 6, and 24 h after surgery (table 4).
The incidences of delirium and other complications after surgery were not statistically different between the two groups. The length of stay in the ICU after surgery of all enrolled patients was shorter in the dexmedetomidine group than in the placebo group (P = 0.038), whereas the lengths of stay in hospital after surgery of both all enrolled patients and those who were included in sleep architecture analyses were longer in the dexmedetomidine group than in the placebo group (P = 0.028 and 0.034, respectively; table 4).
Treatment Safety
The RASS score before the end of study-drug infusion did not differ between the two groups. The incidence of hypotension of all enrolled patients was higher in the dexmedetomidine group than in the placebo group (P = 0.009), whereas the incidences of hypertension of both all enrolled patients and those who were included in sleep architecture analyses were lower in the dexmedetomidine group than in the placebo group (P = 0.002 and 0.026, respectively). There were no significant differences between the two groups with regard to the incidences of other adverse events (bradycardia, tachycardia, and desaturation), the percentages of patients requiring intervention for any adverse events, and the rate of study-drug interruption because of adverse events. No respiratory depression occurred in either group during the first 24 h after surgery. No patient died within 30 days after surgery (table 5).
Discussion
Our results showed that, in elderly patients who were admitted to the ICU after noncardiac surgery and did not require mechanical ventilation, prophylactic low-dose dexmedetomidine infusion increased the percentage of stage N2 sleep. It also prolonged the duration of total sleep time, increased the sleep efficiency, decreased the percentage of stage N1 sleep, and ultimately improved patients’ subjective sleep quality. However, the incidence of hypotension was increased by the treatment.
In the current study, our data further demonstrated that severe sleep disturbances developed in this group of patients during the first night after surgery, which were manifested as shortened total sleep time, increased percentage of stage N1 sleep, decreased percentage of stage N2 sleep, decreased or absent stage N3 and REM sleep, and severe sleep fragmentation (for structure of normal sleep, see appendix 3).46 These results were consistent with the polysomnographic data reported in the ICU medical11–13,47 and surgical patients.9,10,48,49 However, it seems that the degree of sleep disturbances was more severe in our patients than in those previously reported who were not admitted to the ICU after surgery.50–53 This was manifested by the facts that in our patients, total sleep time was shortened more significantly, stage N3 sleep was absent in the majority of patients (78.7%), and REM sleep was absent in all patients. This could be explained as follows. First, the severity of sleep disturbances is closely related to the surgery type (minor or major). In fact, previous studies found that, during the first postoperative night, stage N3 sleep and REM sleep were severely or completely suppressed in patients after major intraabdominal surgery,9,54 whereas stage N3 sleep was decreased and REM sleep was not changed in patients after laparoscopic cholecystectomy.55 The majority of our patients underwent major intraabdominal surgery and thus developed more severe sleep disturbance. Second, age and comorbidity likely contribute to the development of postoperative sleep disturbances. Ageing is associated with sleep pattern changes, and indeed, it is more difficult for the elderly to adjust their sleep to environmental changes.48,56 In addition, comorbid illness and medications are more common in the elderly, which may also contribute to the degree of sleep disruption.56–58 Our patients were of advanced age (mean age, 75.0 ± 5.5 yr), and most of them had preoperative comorbidity. Thus, they were more prone to develop sleep disturbance. Third, the medical environment, medications (especially sedatives and analgesics), and nursing activities in the ICU were also contributors to the sleep disturbance development in our patients.15–17,59 Last, previous studies performed polysomnogram monitoring during nighttime,50–53 whereas our sleep monitoring began from daytime in the afternoon (5:00 pm). Considering the low sleep quality during daytime, prolonged monitoring time (used as a denominator in calculating sleep efficiency) and possibly the resulting prolonged total sleep time (used as a denominator in calculating percentages of each sleep stage) might have resulted in worse sleep parameters.
Dexmedetomidine has been demonstrated to improve sleep quality in mechanically ventilated ICU patients.37,38 However, the administration of sedative dose of dexmedetomidine is associated with the occurrence of side effects (e.g., hypotension and/or bradycardia),34 and the effect of dexmedetomidine on human hemodynamic changes is dose-dependent.60 Keeping this drawback in mind and to eliminate the potential effects of sedation on the sleep architecture, a “nonsedative dose” of dexmedetomidine (0.1 μg kg−1 h−1) was, therefore, administered to our patients who did not require mechanical ventilation. Our results demonstrated that this low-dose dexmedetomidine infusion prolonged the total sleep time and ameliorated the sleep architecture by decreasing the percentage of stage N1 sleep and increasing the percentage of stage N2 sleep. This treatment did not restore normal sleep architecture because stage N3 sleep was decreased and REM sleep was absent in both groups of our patients. Our current results were in line with those in mechanically ventilated patients undergoing dexmedetomidine sedation.37,38 These effects are likely associated with its pharmacologic activation of the endogenous sleep-promoting pathway to produce a state resembling physiologic stage N2 sleep.35,36,61 It is worth noting that the treatment effect in the current study (median, 14.7% increase of stage N2 sleep) was less than we expected in calculating the sample size (20% increase in stage N2 sleep), which made our study less powerful to detect the difference between groups, and in contrast to a previous study,37 our low-dose dexmedetomidine infusion did not decrease sleep fragmentation index. One possible reason is that the dose of dexmedetomidine was too low to produce sedation (median RASS score, 0) in our study, whereas light sedation (RASS score, −1 to −2) was achieved with dexmedetomidine in the previous study.37 We also note that although the median values of many sleep parameters are statistically different between the two groups, the IQRs are very large. This is perhaps because we used a fixed dose of dexmedetomidine and did not consider the individual variability. An individualized dose–response trial is needed to explore the optimal dosing regimen in similar patient populations. On the other hand, considering the importance and significance of stage N3 sleep (restorative sleep) and REM sleep,62,63 the lack of change in these two stages of sleep implies that the beneficial effects of dexmedetomidine infusion may be minimal.
In addition to polysomnography, a subjective evaluation of the sleep quality42,43 was used to monitor sleep quality during the previous night. In accordance with the polysomnographic data, our study also found that the subjective sleep quality during the first postoperative night was improved by the low-dose dexmedetomidine infusion. However, no differences were found between the two groups with regard to the subjective pain scores, whereas other studies reported that the use of dexmedetomidine improved the effect of analgesia and decreased the requirement of opioids.64,65 This discrepancy is very likely due to the dose of dexmedetomidine (low vs. relatively high) and the use of patient-controlled analgesia (yes vs. no).
Our data showed that low-dose dexmedetomidine infusion did not produce significant sedation. It increased the occurrence of hypotension and, although not statistically significant, slightly increased the occurrence of bradycardia. However, considering the facts that the occurrence of these adverse events was transient and the interventions required for these adverse events were low and similar between groups, the use of low-dose dexmedetomidine in clinical practice may be acceptable, but further study is required to clarify the benefit–risk ratio.
In a recent study of healthy volunteers, it was found that sedative dose of dexmedetomidine (a bolus of 0.59 ± 0.25 μg/kg for 10 min followed by an infusion of 0.45 ± 0.06 μg kg−1 h−1) depressed the hypoxic and the hypercarbic ventilatory responses.66 Our results showed that, although desaturation developed in 9.2% (7/76) of all enrolled patients, there were no significant differences between the two groups and none of them required therapeutic intervention. None of our patients developed respiratory depression (hypercapnia or bradypnea). Therefore, the dosing regimen used in this pilot study was safe for respiration. However, this warrants further study since hypoxic or hypercarbic ventilatory responses were not assessed and the sample size was relatively small in our study.
Despite ameliorated sleep quality, clinical outcome, e.g., the incidence of postoperative delirium, was not improved by the treatment in our patients. This is likely because the sample size of the current study was small and hence not large enough to detect the difference between the two groups. This indeed is the case; our recent large sample randomized trial found that low-dose dexmedetomidine did reduce delirium incidence in elderly ICU patients after noncardiac surgery.67
There were several limitations to our study. First, all our participants were screened and enrolled after ICU admission; no baseline sleep study was performed. Therefore, we cannot preclude the potential bias produced by the imbalance of baseline sleep status; strict randomization may have helped to balance this factor between groups. Second, the sample size was relatively small, and 19.7% (15/76) of our patients were excluded from sleep architecture analyses because of failed polysomnographic monitoring or unanalyzable data due to artificial interference. This may result in data bias. Third, the polysomnographic monitoring was only performed on the first night after surgery. Therefore, we could not preclude the residual effects of anesthetics and the effects of postoperative analgesics/sedatives on sleep architecture. For example, benzodiazepines and opiates all decrease the percentages of stage N3 and REM sleep.32,33 However, this should not represent a bias since patients of both groups were randomly recruited and hence they had equal opportunity to receive such interventions. In addition, we could not provide objective evidence for comparison between the groups on the night after the infusion, although we collected subjective reports on sleep quality for subsequent days. Fourth, we did not include a group receiving a commonly used sedative, for example, a γ-aminobutyric acid mimetic. It could be possible that what our study found was due to an “extra” sedative, which provided better sleep per se. However, previous studies reported that γ-aminobutyric acid–mediated sedatives deteriorate, rather than ameliorate, the architecture of sleep.32,33 γ-aminobutyric acid–mediated sedatives, especially benzodiazepines, have not been suggested for use in ICU patients.
In conclusion, our pilot study showed that in nonmechanically ventilated elderly patients who were admitted to the ICU after noncardiac surgery, prophylactic low-dose dexmedetomidine infusion modestly ameliorated the sleep architecture and improved the subjective sleep quality. It increased the occurrence of hypotension, but not the requirement of intervention. Future studies with large sampling size are required to verify the benefit–risk ratio and long-term outcomes of low-dose dexmedetomidine infusion.
Acknowledgments
The authors gratefully acknowledge Xin-Yu Sun, M.D., Department of Psychiatrics, Peking University Sixth Hospital, Beijing, China, for her help in psychiatric consultation and Chun-Mei Deng, M.D., Cong Fu, M.D., and Hao Kong, M.D. (Department of Anesthesiology and Critical Care Medicine, Peking University First Hospital, Beijing, China) for their help with data collection.
Research Support
Supported by Wu Jieping Medical Foundation (Beijing, China; Yan Hua Court 2-601). Study drugs were manufactured and supplied by Jiangsu Hengrui Medicine Co, Ltd, Jiangsu, China. The sponsors have no role in the study design and conduct; the collection, management, analysis, and interpretation of the data; or the preparation and approval of the manuscript.
Competing Interests
Dr. Wang reports that he has received lecture fees and travel expenses for lectures given at domestic academic meetings from Jiangsu Hengrui Medicine Co, Ltd, Jiangsu, China. The other authors declare no competing interests.
Reproducible Science
Full protocol available from Dr. Wang: wangdongxin@hotmail.com. Raw data available from Dr. Wang: wangdongxin@hotmail.com.
Appendix 1: Diagnostic Criteria of Adverse Events
Hypotension was defined as systolic blood pressure less than 90 mmHg or a decrease of more than 20% from baseline (if the baseline value [before study-drug infusion] is lower than 113 mmHg) or diastolic blood pressure less than 50 mmHg.
Bradycardia was defined as heart rate less than 50 beats/min or a decrease of more than 20% from baseline (if the baseline value is lower than 63 beats/min).
Hypertension was defined as systolic blood pressure more than 160 mmHg or an increase of more than 20% from baseline (if the baseline value is higher than 133 mmHg) or diastolic blood pressure more than 100 mmHg.
Tachycardia was defined as heart rate more than 100 beats/min or an increase of more than 20% from baseline (if the baseline value is higher than 83 beats/min).
Desaturation was defined as pulse oxygen saturation less than 90% or a decrease of more than 5% (absolute value) from baseline (if the baseline value is lower than 95%).
Respiratory depression was defined as arterial blood carbon dioxide partial pressure more than 50 mmHg or respiratory rate less than 10 breaths/min.
Appendix 2: The Confusion Assessment Method
The confusion assessment method is a delirium assessment tool published in 1990. It detects four features of delirium including (1) acute onset of mental status changes or a fluctuating course, (2) inattention, (3) disorganized thinking, and (4) altered level of consciousness. To diagnose delirium, a patient must display the first two aforementioned features plus either the third or fourth aforementioned feature.44,45
Appendix 3: Structure of Normal Sleep
Normal sleep has a significant circadian rhythm with a certain period of duration and cycles of the non–rapid eye movement sleep and rapid eye movement (REM) sleep (normally in the order of N1 → N2 → N3 → N2 → REM). Stage N1 sleep is usually called light sleep and makes up 5 to 10% of total sleep in adults. Stage N2 sleep occupies 45 to 55% of total sleep in adults. Once sleep becomes deeper, stage N2 will progress to stage N3. Stage N3 sleep accounts for 15 to 25% of total sleep in adults and is also called slow-wave sleep or deep sleep. REM sleep usually occupies 20 to 25% of total sleep in adults. During REM sleep, most muscles are paralyzed, and heart rate, breathing, and body temperature become unregulated, and the sleeper may experience vivid dreams. Appropriate sleep structure and time are important to keep health status. For healthy older adults of 75 yr, normal sleep usually requires a total sleep time of about 6 h; of which, 11%, 57%, 12%, and 20% are the stage N1, N2, N3, and REM sleep, respectively.46,68,69