Melatonin, a substance synthesized in the pineal gland, exhibits pleiotropic physiologic actions (1), and several recent studies have demonstrated a variety of effects of melatonin in vivo, including antioxidant and anti-inflammatory effects, as well as an effect on regulation of the circadian rhythm. Although the melatonin secretion rhythm has been shown to be severely depressed in patients admitted to the ICU because of various environmental factors as well as factors related to the patients’ life-threatening conditions themselves and the drugs used (2–4), few studies have been conducted yet to determine whether exogenous melatonin administration in the ICU can improve the outcomes in critically ill patients (56).
One of the most well-known functions of melatonin is to prevent delirium in critically ill patients (78). Delirium is an undesirable, but frequently encountered manifestation in ICU patients (9), and is known to be associated with worse outcomes (910), including longer ICU stay (11), higher mortality rate (12), longer time on the ventilator (13), and higher ICU and hospital costs (14). Considering that recent randomized controlled clinical trials have indicated that melatonin can prevent delirium (15–17), we considered that prophylactic administration of melatonin or melatonin agonists may shorten the duration of ICU stay and improve the outcomes in critically ill patients.
Therefore, we performed this single-center, triple-blinded, randomized placebo-controlled trial to examine whether prophylactic administration of ramelteon, a selective melatonin receptor agonist, is effective for shortening the duration of ICU stay in critically ill patients. As secondary endpoints, we measured the mortality rate and the occurrence rate and duration of delirium in ICU patients. To the best of our knowledge, this is the first randomized controlled clinical trial conducted exclusively in ICU patients to determine whether prophylactic administration of ramelteon can shorten the duration of ICU stay and decrease the occurrence rate of delirium in ICU patients.
MATERIALS AND METHODS
Setting and Participants
We conducted this single-center, triple-blinded (the testers, subjects, and statisticians were blinded), randomized placebo-controlled trial with the approval of the Research Ethics Board of Nagoya University Hospital (2015-0005). Nagoya University Hospital is an academic and educational hospital with 1,035 beds, including 10 emergency and medical ICU (EMICU) beds and 16 surgical ICU beds. This trial was performed in the EMICU, and the subjects were patients admitted via the emergency department (emergency ICU) or medical patients admitted via the general ward (medical ICU) who needed critical care treatment. The trial is registered in the University Hospital Medical Information Network Clinical Trials Registry as the Melatonin Evaluation of Lowered Inflammation in ICU Trial (UMIN000016541). We did not have any money from or speak on behalf of the manufacturer of the product used in our study.
The eligible subjects were adults (age ≥ 20 yr old) admitted to our EMICU between May 2015 and April 2017, who could receive their medications orally or through a nasogastric tube during the first 48 hours of admission to the ICU. Patients who were already receiving ramelteon or fluvoxamine maleate prior to their admission to the ICU were excluded from the study because of potential drug interactions. Patients with known allergy to ramelteon were also excluded, as also patients who refused to provide consent for participation in the study.
Randomization and Intervention
This study was performed as a triple-blinded, placebo-controlled randomized trial. Eligible participants were randomly assigned at a 1:1 ratio to the ramelteon group (ramelteon: 8 mg/d) or the placebo group (placebo; 1 g/d of lactose powder) at the end of the baseline assessment. The randomization was performed using stratified block randomization with a block size of four. The randomization list was made by an outsider before the start of enrollment, based on stratification according to the age (≥ 60/< 60 yr) and Acute Physiology and Chronic Health Evaluation (APACHE) II score (≥ 30/< 30 points) and according to whether the patients were intubated or not. Both ramelteon and placebo were pulverized in advance to make them indistinguishable from each other.
Depending on the group assignment, the patients received ramelteon or placebo in a blinded manner at 20:00 hours each day, until they were discharged from our ICU. All of the nurses in the ICU were also blinded to any information about the patients relevant to the study and provided equal care to all patients, including daily preventive care, avoidance of unnecessary immobilization, and regular verbal communication. Visits by the patients’ families were limited to twice a day: between 11 am and 12 noon and between 3 pm and 4 pm; after 10 pm, the lighting and noise in all the ICU rooms were reduced to the minimum. At the baseline, we recorded the demographic data, medical history, and medication history of the patients. Evaluation of the patients by the Richmond Agitation-Sedation Scale (RASS) was performed every 4 hours by trained ICU nurses (the assessment was skipped if the patients were sleeping).
Outcomes and Data collection
The primary endpoint of the study was the duration of ICU stay. The indications for admission and suitability for discharge of the patients were determined at a multidisciplinary conference of ICU doctors and the patients’ attending physicians, held twice every day. In both patient groups, the suitability for discharge from the ICU was determined based on the following: no or minimal need for sedative drugs, the oxygen saturation as measured by pulse oxymetry greater than 90% at an Fio2 of less than 0.5 in the absence of mechanical ventilatory support, no or minimal need for inotropes or vasopressor agents, and urine output more than 0.5 mL/kg/hr. Patients who were intubated and not yet tracheotomized and patients who needed continuous hemodiafiltration were not discharged. Patients were considered suitable for discharge when they could be disconnected from all the ICU monitors.
The secondary endpoints of the study were the occurrence rate and duration of delirium in the patients during their ICU stay and the clinical status at discharge. Delirium was assessed every 4 hours using the Confusion Assessment Method for the ICU (18), by trained ICU nurses (the assessment was skipped if the patients were sleeping) and by the ICU doctors as needed.
In regard to analysis of the sleep variables, the number of awakenings per night, the proportion of nights without awakenings relative to the total number of nights in the ICU, and the mean hours of sleep of the nonintubated patients were determined by retrospectively reviewing the electronic charts in both groups. We reviewed the nursing observations and records and the rater observations to collate the data.
Treatment for Delirium
Delirium occurring in the ICU was primarily managed by nonpharmacologic approaches, including early mobilization and creation of an environment conducive to comfortable sleep, together with the removal of any potential cause of delirium, such as pain, discomfort, dyspnea, and anxiety. In cases judged as requiring pharmacologic treatment for the control of delirium, risperidone (1 mg, oral) or continuous infusion of dexmedetomidine at progressively increasing doses as necessary (maximum dose: 0.7 µg/kg/hr) was used initially, according to the clinical preference of the attending physician. In the event of emergency, in patients who were judged by the treating physician as needing more immediate control of delirium than that afforded by oral treatment, we used haloperidol (5 mg, IV). Additional administration of risperidone, haloperidol, or other medications was considered depending on the severity of the delirium.
For the primary efficacy measure (duration of ICU stay), we assumed a mean of the log-transformed value of the duration of ICU stay of 4.17 (65.28 hr), with a sd of 1.09 in the placebo group, based on the past sample data of patients admitted to our ICU (December 1–31, 2014). It was estimated that a sample size of 91 per group would be needed for detecting a reduction of the duration of ICU stay by 1 day (24 hr) in the ramelteon group, which we assumed as a clinically significant effect size, with at least 80% statistical power at a two-sided significance level of 5% by Student two-sample t test. During the sampling period (December 1–31, 2014), 50 patients were admitted to our ICU, and the number of eligible patients was 26. We assumed that about 1 year would be needed for collecting 182 patients and set the study period at May 2015 to May 2016. Because much fewer patients than expected actually met the eligible criteria during the set study period, we did not achieve the target sample size by May 2016 and extended the registration period for the study to April 2017. The target sample size was, however, not met even by the end of the additional registration period.
The statistical analysis was performed on an intention-to-treat basis, and the data of subjects who received ramelteon or placebo at least once during their ICU stay were included. We performed Student t test to evaluate the significance of differences in the log-transformed values of the duration of ICU stay. To incorporate possible imbalances in the patients’ baseline characteristics between the ramelteon and placebo groups, a multivariate linear regression analysis was conducted to identify factors that were independently associated with the log-transformed value of the duration of ICU stay. The baseline characteristics that were adjusted for in the analysis were those that were considered as potentially influencing the duration of ICU stay: age, APACHE II score, intubation status (intubated/not intubated), presence/absence of dementia before admission, and the mean RASS score during the ICU stay. For analysis of the secondary endpoints, Fisher exact test was used for the categorical variables (occurrence of delirium and mortality at discharge) and Student t test for the continuous variables (duration of delirium and sleep variables). All reported p values were two sided, and p value of less than 0.05 was regarded as denoting a statistically significant difference. All analyses were conducted using the SAS software, Version 9.4 (SAS Institute, Cary, NC).
A total of 98 subjects were screened to determine if they met the eligibility criteria for this study; six of these patients who refused to provide consent for participation in this study were excluded. There were no patients who were already receiving ramelteon or fluvoxamine maleate before their admission to the ICU or were known to have allergy to ramelteon. The 92 eligible patients were then randomized to the ramelteon or the placebo group. Of these, four patients (two from the ramelteon and two from the placebo group) were discharged from the ICU even before the start of the intervention (three patients [two from the ramelteon group and one from the placebo group] were transferred to the general ward because of stabilization of the general condition, and one patient [placebo group] died). The remaining 88 patients were included in the intent-to-treat analysis (Fig. 1).
The baseline characteristics of the enrolled patients in the ramelteon group and placebo group are summarized in Table 1. The distribution of the patient characteristics was generally similar between the two groups, especially in respect of the three allocation factors for the randomization (age, APACHE II score, and intubation status), indicating that the randomization was carried out appropriately. However, we observed slightly greater proportions of subjects with the following characteristics in the ramelteon group: male subjects and subjects with dementia, sepsis, history of habitual heavy use of alcohol, and history of medication.
In the primary endpoint analysis, the median (range) duration of ICU stay was 4.56 days (2.10–7.07 d) days in the ramelteon group and 5.86 days (2.97–14.16 d) in the placebo group, with no statistically significant difference between the two groups, as assessed by the Student t test comparing the log-transformed values of the duration of ICU stay (p = 0.082) (Fig. 2). In the multivariate analysis conducted with adjustments for the five prespecified baseline characteristics, prophylactic administration of ramelteon was identified as being significantly and independently associated with a decrease in the duration of ICU stay (p = 0.028) (Table 2). In addition, we also performed sensitivity analyses using all the variables shown in Table 1 for adjustment, which also revealed a statistically significant association between prophylactic ramelteon administration and decrease in the duration of ICU stay (Supplemental Table 1, Supplemental Digital Content 1, http://links.lww.com/CCM/D470).
The results for the secondary endpoints are shown in Table 3. The occurrence rate of delirium was 24.4% (11/45) in the ramelteon group and 46.5% (20/43) in the placebo group, the difference between the two groups being statistically significant (p = 0.044; odds ratio, 2.69 [1.09–6.65]). Furthermore, the duration of delirium was also statistically significantly shorter in the ramelteon group as compared to that in the placebo group (0.78 vs 1.40 d; p = 0.048). The mortality rate at discharge from the ICU was 6.7% (3/45) in the ramelteon group and 7.5% (3/43) in the placebo group, with no significant difference between the two groups (p > 0.999).
We also compared the following sleep variables in the nonintubated patients between the ramelteon and placebo groups: the number of awakenings per night, the proportion of nights without awakenings relative to the total number of nights in the ICU, and the mean hours of sleep. In the univariate analyses, the ramelteon group showed statistically significantly fewer awakenings per night (0.80 times per night vs 1.31 times per night; p = 0.045) and a statistically significantly higher proportion of nights without awakenings (51% vs 30%; p = 0.048) as compared to the placebo group, while there was no significant difference in the mean hours of sleep between the two groups (7.29 vs 6.78 hr; p = 0.252). Even after adjustments for the baseline characteristics in the multivariate analysis, similar results were obtained (p = 0.036, p = 0.019, and p = 0.123, respectively) (Supplemental Table 2, Supplemental Digital Content 2, http://links.lww.com/CCM/D471).
In this study, prophylactic administration of ramelteon, a selective melatonin receptor agonist, was associated with both a reduced occurrence rate and shortened duration of delirium in the ICU. Although no statistically significant difference in the duration of ICU stay was observed between the two groups, multivariate analysis performed with adjustments for the baseline patient characteristics revealed that prophylactic administration of ramelteon was significantly and independently associated with a shortened duration of ICU stay. To the best of our knowledge, this is the first report to demonstrate a beneficial effect of ramelteon on the duration of ICU stay in critically ill patients.
Ramelteon was demonstrated in an in vitro study as having six-fold and three-fold higher affinities for melatonin receptor 1 and melatonin receptor 2, respectively, compared with melatonin (19). Hatta et al (16) reported a statistically significant decrease in the occurrence rate of delirium associated with the use of ramelteon in elderly patients. However, there are no reported studies on the effect of ramelteon conducted exclusively in ICU patients, although two large randomized clinical trials are ongoing (20, 21). Considering that the occurrence rate of delirium in the ICU is much higher than that in general wards (22), it may not be unreasonable to carefully extrapolate the results of the two aforementioned previous studies to ICU patients. The current study is the first to suggest that prophylactic administration of ramelteon also prevents the occurrence of delirium in ICU patients, as in general ward patients.
Interestingly, our results showed that the ramelteon group had statistically significantly fewer awakenings per night and a higher proportion of nights without awakenings as compared to the placebo group, although unexpectedly, there was no statistically significant difference in the mean hours of sleep between the two groups. This result may provide a clue to explain the biological mechanism underlying the prevention of delirium by ramelteon observed in this study. That is, ramelteon probably reduces delirium by increasing the duration of restful sleep, although further study will be needed.
This study was a small single-center study, which may limit the generalizability of the results. We calculated the power a priori with the goal of obtaining preliminary data for a larger randomized controlled study. We think that a large multicenter randomized clinical trial is needed in the future. Although we had estimated in advance that a total of 200 subjects would be required, only 88 subjects could be enrolled within the set registration period because fewer patients than we expected actually met the eligibility criteria during the study period. Therefore, the study may have been underpowered. During the study period, visits by the patients’ families were limited to twice a day, even though an open-door visiting policy is promoted in ICUs nowadays for reducing the occurrence of delirium. In this study, we did not assess the type of delirium, hyperactive, hypoactive, or mixed type delirium. Finally, this dose of haloperidol (5 mg) may be much higher than the usual dose used around the world as rescue treatment for delirium patients.
It is interesting that the estimated effect size of ramelteon was 1.30 days in the unadjusted comparison, which was greater than the effect size of 1.00 day that we had calculated in advance, during the sample size calculation, as a clinically significant effect size. Although there was no statistical significance in the unadjusted comparison analysis, this large estimate of the effect size could be one of the main reasons for obtaining p value equals to 0.082, which is close to the value of α equals to 0.05, and also a statistically significant result (p = 0.028) in the adjusted analysis, in spite of the smaller than expected sample size. It is noteworthy that in our study also, we observed no adverse events associated with ramelteon administration, which is consistent with the results of previous studies (2324). Prophylactic administration of ramelteon may have a potential advantage in terms of the safety and efficacy among ICU patients, and it would be great interest to investigate the cost-effectiveness of prophylactic administration of ramelteon in the ICU.
Ramelteon administration was associated with a tendency toward a decreased duration of ICU stay, as well as significant decreases in the occurrence rate and duration of delirium in patients admitted to the ICU.
We thank all the patients who participated in the trial. We also thank Mr. Yusuke Sakai for his assistance in implementing the randomization and Drs. Takanori Yamamoto and Hideki Tokuyama for monitoring this study to ensure that it was going according to plan.
1. Brzezinski AMelatonin in humans. N Engl J Med 1997; 336:186195
2. Olofsson K, Alling C, Lundberg D, et al.Abolished circadian rhythm of melatonin secretion in sedated and artificially ventilated intensive care patients. Acta Anaesthesiol Scand 2004; 48:679684
3. Miyazaki T, Kuwano H, Kato H, et al.Correlation between serum melatonin circadian rhythm and intensive care unit psychosis after thoracic esophagectomy. Surgery 2003; 133:662668
4. Shilo L, Dagan Y, Smorjik Y, et al.Patients in the intensive care unit suffer from severe lack of sleep associated with loss of normal melatonin secretion pattern. Am J Med Sci 1999; 317:278281
5. Bourne RS, Mills GH, Minelli CMelatonin therapy to improve nocturnal sleep in critically ill patients: Encouraging results from a small randomised controlled trial. Crit Care 2008; 12:R52
6. Mistraletti G, Umbrello M, Sabbatini G, et al.Melatonin reduces the need for sedation in ICU patients: A randomized controlled trial
. Minerva Anestesiol 2015; 81:12981310
7. Walker CK, Gales MAMelatonin receptor agonists for delirium
prevention. Ann Pharmacother 2017; 51:7278
8. Tsuda A, Nishimura K, Naganawa E, et al.Ramelteon
for the treatment of delirium
in elderly patients: A consecutive case series study. Int J Psychiatry Med 2014; 47:97104
9. Inouye SKDelirium in older persons. N Engl J Med 2006; 354:11571165
10. Ely EW, Gautam S, Margolin R, et al.The impact of delirium
in the intensive care unit on hospital length of stay. Intensive Care Med 2001; 27:18921900
11. Barr J, Fraser GL, Puntillo K, et al.American College of Critical Care
Medicine: Clinical practice guidelines for the management of pain, agitation, and delirium
in adult patients in the intensive care unit. Crit Care Med 2013; 41:263306
12. Shehabi Y, Riker RR, Bokesch PM, et al.SEDCOM (Safety and Efficacy of Dexmedetomidine Compared With Midazolam) Study Group: Delirium
duration and mortality in lightly sedated, mechanically ventilated intensive care patients. Crit Care Med 2010; 38:23112318
13. Pisani MA, Kong SY, Kasl SV, et al.Days of delirium
are associated with 1-year mortality in an older intensive care unit population. Am J Respir Crit Care Med 2009; 180:10921097
14. Milbrandt EB, Deppen S, Harrison PL, et al.Costs associated with delirium
in mechanically ventilated patients. Crit Care Med 2004; 32:955962
15. Sultan SSAssessment of role of perioperative melatonin in prevention and treatment of postoperative delirium
after hip arthroplasty under spinal anesthesia in the elderly. Saudi J Anaesth 2010; 4:169173
16. Hatta K, Kishi Y, Wada K, et al.DELIRIA-J Group: Preventive effects of ramelteon
: A randomized placebo-controlled trial. JAMA Psychiatry 2014; 71:397403
17. Al-Aama T, Brymer C, Gutmanis I, et al.Melatonin decreases delirium
in elderly patients: A randomized, placebo-controlled trial. Int J Geriatr Psychiatry 2011; 26:687694
18. Ely EW, Inouye SK, Bernard GR, et al.Delirium
in mechanically ventilated patients: Validity and reliability of the confusion assessment method for the intensive care unit (CAM-ICU). JAMA 2001; 286:27032710
19. Miyamoto MPharmacology of ramelteon
, a selective MT1/MT2 receptor agonist: A novel therapeutic drug for sleep disorders. CNS Neurosci Ther 2009; 15:3251
20. Burry L, Scales D, Williamson D, et al.Feasibility of melatonin for prevention of delirium
in critically ill patients: A protocol for a multicentre, randomised, placebo-controlled study. BMJ Open 2017; 7:e015420
21. Martinez FE, Anstey M, Ford A, et al.Prophylactic Melatonin for Delirium
in Intensive Care (Pro-MEDIC): Study protocol for a randomised controlled trial. Trials 2017; 18:4
22. McNicoll L, Pisani MA, Zhang Y, et al.Delirium
in the intensive care unit: Occurrence and clinical course in older patients. J Am Geriatr Soc 2003; 51:591598
23. Uchimura N, Ogawa A, Hamamura M, et al.Efficacy and safety of ramelteon
in Japanese adults with chronic insomnia: A randomized, double-blind, placebo-controlled study. Expert Rev Neurother 2011; 11:215224
24. Erman M, Seiden D, Zammit G, et al.An efficacy, safety, and dose-response study of Ramelteon
in patients with chronic primary insomnia. Sleep Med 2006; 7:1724