Abstract
Background: New-onset insomnia and other sleep disturbances occur more frequently in the inpatient setting due to a variety of physical and psychological factors. Studies have found that non-pharmacologic interventions can be effective in treating insomnia in the inpatient setting, particularly in the ICU, to prevent adverse outcomes, but further research is needed to identify optimal pharmacologic interventions. Objective: To compare treatment outcomes of patients initiated on melatonin and trazodone to treat new-onset insomnia in non-ICU hospitalized patients based on the need for an additional sleep aid therapy during hospitalization and to compare the rate of adverse events of each agent. Methods: A retrospective chart review was conducted for adult patients admitted to a non-ICU general medicine or surgical floor at a community teaching hospital between July 1, 2020 and June 30, 2021. Patients were included if they were initiated on scheduled melatonin or trazodone for the treatment of new onset insomnia during their hospitalization. Patients were excluded if they had a previous diagnosis of insomnia, were prescribed 2 sleep aids simultaneously, or if their admission medication reconciliation included pharmacologic treatment for insomnia. Clinical data collected included non-pharmacologic interventions, sleep aid dose, number of doses of sleep aid administered, and total number of nights an additional sleep aid was needed. The primary outcome was the percentage of patients needing additional therapy defined as, administering an additional sleep aid between 2100-0600 or utilizing more than 1 sleep aid agent during hospitalization compared between melatonin and trazodone. Secondary outcomes of this study included the rate of adverse events such as difficulty awakening, daytime sleepiness, serotonin syndrome, falls, and development of in hospital delirium. Results: Of 158 included patients, 132 received melatonin and 26 received trazodone. Male sex (53.8% [melatonin] vs. 53.8% [trazodone]; P = 1), hospital length of stay (7.7 vs 7.7 days; P = .68), and administration of drugs that could cause insomnia (34.1% vs 23.1%vs; P = .27) were similar between sleep aids. Percentage of patients needing an additional sleep aid during hospitalization (19.7% vs 34.6%; P = .09), and patients prescribed a sleep aid at discharge (39.4% vs 46.2%; P = .52) were similar between sleep aids, respectively. Rates of adverse events were similar between sleep aids. Conclusions: There was no significant difference between the 2 agents in terms of the primary outcome, even though a higher rate of patients treated with trazodone for new-onset insomnia during hospitalization required an additional sleep aid compared to those treated with melatonin. No difference in adverse events was observed.
Keywords: trazodone, melatonin, insomnia, general medicine, delirium
Introduction
Lack of sleep during hospitalization is often a result of both environmental and nonenvironmental disturbances. 1 Factors such as patient care interventions, noise, and concomitant medications can all contribute to changes in sleep quality and quantity. 1 Consequences associated with poor sleep quality and loss of sleep during hospitalization have not been well studied. Sleep deprivation has been linked to delirium and declines in respiratory function within an intensive care unit setting, however, there is a lack of data generalizing these outcomes to patients outside of the ICU. 2 Current data in non-ICU patients demonstrates an association with worsening and shorter duration of sleep with higher risk of hyperglycemia. 3 Studies have found that various non-pharmacologic interventions can be effective in treating insomnia in the inpatient setting to prevent adverse outcomes, but further research is needed to identify optimal pharmacologic interventions. 4
It has been shown that melatonin levels in critically ill patients are significantly decreased, which led to the development of studies that found an improvement in sleep quality with the use of melatonin in ICU patients. 4 Additionally, it has been stated that benzodiazepines, such as lorazepam or temazepam, are potentially effective treatment options based on their pharmacokinetic profiles. When considering a benzodiazepine for insomnia treatment, it is important to note that all agents within the drug class have similar effects on sleep and it is their pharmacokinetics that set them apart and make some more desirable than others. 5 Zolpidem is often used as an alternative to benzodiazepines in both the inpatient and outpatient settings due its slightly different mechanism of activity that makes it more favorable when there is concern for potential abuse. 6 Trazodone has also been used as a potentially safer option with less concern for cognitive impairment and falls. 7 While each of these agents have been evaluated for either their efficacy or safety in individual studies, there is limited data that supports the use of one agent over another.
At our institution, melatonin, ramelteon, trazodone, zolpidem, and benzodiazepines are the agents most frequently used for the treatment of new-onset insomnia in general medicine patients. However, to date, the institution has not adopted a standardized treatment protocol due to the lack of data supporting the use of certain agents over others in this patient population. This study aimed to provide guidance for a recommended pharmacotherapy agent for non-ICU patients suffering from new-onset insomnia requiring scheduled pharmacotherapy in the inpatient setting with a focus on trazodone or melatonin based on fewer adverse effects.
Methods
Project Design
This IRB approved, single-center, retrospective study was conducted for patients admitted to a community teaching hospital between July 1, 2020 and June 30, 2021. Patients were identified if they received melatonin or trazodone as a scheduled sleep aid during their hospitalization. The primary outcome was the percentage of patients needing additional therapy defined as, administering an additional sleep aid between 2100-0600 or utilizing more than 1 sleep aid agent during hospitalization. Secondary outcomes of this study included the rate of adverse events such as difficulty awakening, daytime sleepiness, serotonin syndrome, falls, and development of in hospital delirium.
Study Population
Adult patients admitted to a non-ICU general medicine or surgical floor and initiated on a scheduled sleep aid of melatonin or trazodone for the treatment of new onset insomnia during hospitalization, were included in the study. Patients needed to receive at least 2 doses of the sleep aid to be included. Exclusion criteria included if the patient had an ICU stay during admission, had a previous diagnosis of insomnia, if their medication reconciliation included pharmacologic treatment for insomnia prior to admission, or had 2 or more scheduled sleep aids initiated at the same time during admission.
Data Collection
Demographic data included age at admission, sex, race, primary reason for admission, length of hospitalization, administration of medications that may cause insomnia (corticosteroids, methylphenidate, nicotine, acetylcholinsterase inhibitors), diagnosis of other sleep disorders (sleep apnea, restless leg syndrome, narcolepsy), and significant past medical history. Clinical data collected included non-pharmacologic interventions, primary sleep aid dose, number of doses of sleep aid administered, total number of nights an additional sleep aid was needed, and adverse events (difficulty awakening, daytime sleepiness, serotonin syndrome, fall, new onset delirium), which was identified based on documentation in clinical progress notes. Additional sleep aids that were included were melatonin, trazodone, benzodiazepines, zolpidem, and ramelteon.
Statistical Analysis
Nominal data was analyzed using the Chi-square or Fisher’s exact test. Continuous data was analyzed using the t-test. The P-value was set a priori at .05 in order to determine statistical significance.
Results
Among 158 patients that met inclusion criteria, 132 (83.5%) were initiated on melatonin for the treatment of new onset insomnia and 26 (16.5%) were initiated on trazodone. The most common reason for exclusion was a medication reconciliation including a pharmacologic treatment for insomnia prior to admission. Patients included in the study were primarily white and resided in either progressive care or general medicine units. Male sex and median hospital length of stay were similar between the 2 groups, respectively. Patients prescribed melatonin tended to be older and less likely to be admitted to the hospital due to metabolic disturbances compared to the trazodone group (Table 1). Common metabolic disturbances on admission included electrolyte imbalances and hypoglycemia. Of patients in the melatonin group, the most common initial dose was 6 mg by mouth at bedtime (94/132, 71.2%) followed by 3 mg at bedtime (25/132, 18.9%). Of patients prescribed trazodone, 50 mg by mouth at bedtime was the most common dose (21/26, 80.8%) followed by 25 mg at bedtime (2/26, 7.7%) and 150 mg at bedtime (2/26, 7.7%).
Table 1.
Patient Demographics and Baseline Characteristics.
| All patients (n = 158) | Trazodone (n = 26) | Melatonin (n = 132) | P value | |
|---|---|---|---|---|
| Age at admission (years), mean (SD) | 68.9 (16) | 53.7 (18.6) | 71.9 (13.6) | <.00001 |
| Male sex, n (%) | 85 (53.8) | 14 (53.8) | 71 (53.8) | 1 |
| Race, n (%) | .99 | |||
| White | 126 (79.7) | 21 (80.8) | 105 (79.5) | |
| Black/African American | 26 (16.5) | 4 (15.4) | 22 (16.7) | |
| Other/Unknown | 6 (3.8) | 1 (3.8) | 5 (3.79) | |
| PMH, n (%) | ||||
| COPD | 26 (16.5) | 2 (7.69) | 24 (18.2) | .25 |
| OSA | 16 (10.1) | 2 (7.69) | 14 (10.6) | 1 |
| Obesity (BMI ≥ 30) | 21 (13.3) | 5 (19.2) | 16 (12.1) | .33 |
| Depression | 29 (18.4) | 6 (23.1) | 23 (17.4) | .5 |
| Anxiety | 27 (17.1) | 7 (26.9) | 20 (15.2) | .14 |
| Chronic pain | 13 (8.3) | 1 (3.85) | 12 (9.09) | .7 |
| Diagnosis of sleep disorder prior to admission, n (%) | ||||
| Sleep apnea | 1 (0.63) | 0 (0) | 1 (0.76) | 1 |
| Restless leg syndrome | 0 (0) | 0 (0) | 0 (0) | 1 |
| Narcolepsy | 0 (0) | 0 (0) | 0 (0) | 1 |
| Primary reason for admission | ||||
| Cardiovascular | 8 (5.06) | 1 (3.85) | 7 (5.3) | 1 |
| Respiratory | 12 (7.59) | 1 (3.85) | 11 (8.33) | .69 |
| Gastrointestinal | 11 (6.96) | 3 (11.5) | 8 (6.06) | .39 |
| Neurological | 47 (29.7) | 5 (19.2) | 42 (31.8) | .199 |
| Sepsis/Infection | 10 (6.33) | 2 (7.69) | 8 (6.06) | .67 |
| Trauma | 21 (13.3) | 5 (19.2) | 16 (12.1) | .329 |
| Metabolic | 5 (3.16) | 3 (11.5) | 2 (1.52) | .03 |
| Hematological | 9 (5.7) | 1 (3.85) | 8 (6.06) | 1 |
| Renal | 4 (4.43) | 0 (0) | 4 (3.03) | 1 |
| Pain | 6 (3.8) | 0 (0) | 6 (4.54) | .59 |
| Other | 25 (15.8) | 5 (19.2) | 20 (15.2) | .60 |
| Hospital length of stay (days), median (IQR) | 7.7 (7) | 7.7 (8.1) | 7.7 (6.6) | .68 |
| Duration of scheduled sleep aid (days), median (IQR) | 4 (4) | 4 (5) | 4 (3.5) | .569 |
| Administration of medications that may cause insomnia | ||||
| Total | 51 (32.3) | 6 (23.1) | 45 (34.1) | .27 |
| Corticosteroid, n (%) | 31 (19.6) | 3 (11.5) | 28 (21.2) | .42 |
| Methylphenidate, n (%) | 0 (0) | 0 (0) | 0 (0) | 1 |
| Nicotine, n (%) | 13 (8.2) | 3 (11.5) | 10 (7.6) | .45 |
| Acetylcholinesterase inhibitor, n (%) | 7 (4.4) | 0 (0) | 7 (5.3) | .60 |
| None of the above, n (%) | 110 (69.6) | 20 (76.9) | 90 (68.2) | .38 |
| Non-pharm interventions | ||||
| Encourage sleep/wake cycle, n (%) | 23 (14.6) | 1 (3.8) | 22 (16.7) | .13 |
| Promote sleep hygiene, n (%) | 11 (7.0) | 0 (0) | 11 (8.3) | .21 |
| Minimizing blood draws at night, n (%) | 17 (10.8) | 0 (0) | 17 (12.9) | .08 |
| Minimize vital signs at night, n (%) | 17 (10.8) | 0 (0) | 17 (12.9) | .08 |
| Use of eye masks or ear plugs, n (%) | 1 (0.6) | 0 (0) | 1 (0.8) | 1 |
| Minimizing noise disturbances, n (%) | 20 (12.7) | 0 (0) | 20 (15.2) | .046 |
| Encouraging daytime light exposure, n (%) | 35 (22.2) | 1 (3.8) | 34 (25.8) | .01 |
| None of the above, n (%) | 70 (44.3) | 25 (96.2) | 45 (34.1) | <.00001 |
Percentage of patients needing an additional sleep aid or change in sleep aid during hospitalization (19.7% [melatonin] vs 34.6% [trazodone]; P = .09) was similar between sleep aids (Table 2). Of the 25 patients in the melatonin group needing an additional sleep aid, the most common sleep aid prescribed was trazodone (9/25, 36%) followed by benzodiazepines (6/25, 24%) and ramelteon (6/25, 24%) In the trazodone group, the most common additional sleep aid administered was melatonin (5/9, 55.6%) followed by benzodiazepines (4/9, 44.4%). Patients prescribed a sleep aid or diagnosed with insomnia at discharge were similar between sleep aids. Patients prescribed trazodone were less likely to have documentation of a nonpharmacologic intervention compared to patients in the melatonin group. Rates of adverse events including difficulty awakening, daytime sleepiness, serotonin syndrome, falls, or development of in hospital delirium were similar between sleep aids (Table 2).
Table 2.
Primary and Secondary Outcomes.
| All patients (n = 158) | Trazodone (n = 26) | Melatonin (n = 132) | P value | |
|---|---|---|---|---|
| Administration of additional sleep aid and/or change of sleep aid agent | 35 (22.2) | 9 (34.6) | 26 (19.7) | .09 |
| Change of sleep aid agent, n (%) | 1 (0.6) | 0 (0) | 1 (0.8) | 1 |
| Administration of additional sleep aid, n (%) | 34 (21.5) | 9 (34.6) | 25 (18.9) | .08 |
| New sleep aid at discharge, n (%) | 64 (40.5) | 12 (46.2) | 52 (39.4) | .52 |
| New diagnosis of insomnia at discharge, n (%) | 6 (3.8) | 2 (7.69) | 4 (3.03) | .26 |
| Adverse event | 14 | 1 (3.8) | 13 (9.8) | .47 |
| Difficulty awakening, n (%) | 4 (2.5) | 0 (0) | 4 (3.0) | 1 |
| Daytime sleepiness, n (%) | 5 (3.2) | 0 (0) | 5 (3.7) | .59 |
| Fall, n (%) | 1 (0.6) | 0 (0) | 1 (0.8) | 1 |
| New onset delirium, n (%) | 4 (2.5) | 1 (3.8) | 3 (2.3) | .52 |
| None of the above, n (%) | 145 (91.8) | 25 (96.2) | 120 (90.9) | .37 |
Discussion
This retrospective study provides insight on the management of new-onset insomnia during hospitalization, specifically within the non-ICU setting. While various non-pharmacologic and pharmacologic interventions for new-onset insomnia have been evaluated in the ICU setting, typically in relation to ICU delirium, there is limited guidance on pharmacologic interventions in the non-ICU population.
Medications that are often used to treat insomnia in the inpatient setting include: melatonin, trazodone, zolpidem, and hypnotics such as benzodiazepines or eszopiclone. 8 Previous studies have focused on the efficacy of these agents for the prevention of ICU delirium and the incidence rates of adverse events with their use, but there is limited data assessing the effectiveness of these agents for the treatment of new-onset insomnia in hospitalized non-ICU patients. 7 In general, melatonin and trazodone were found to be effective in our studied population based on a relatively low failure rate in terms of the primary outcome. These results align with those found in previous studies that evaluated these agents within an ICU population, however this study is one of the first to evaluate these agents within the adult non-ICU population. When evaluating our primary outcome, the percentage of patients requiring an additional sleep aid was not statistically significant, however it should be noted that the results had a numerical difference with 19.7% of patients treated with melatonin needing an additional sleep aid compared to 34.6% of patients treated with trazodone. It’s important to recognize that our study was limited by its sample size, and therefore did not meet power to detect the primary outcome based on a post hoc analysis that calculated power at 39.8% from our results and sample size. The difference in the size of the groups may be attributed to providers preferentially choosing melatonin for older patients compared to trazodone. This theory was supported by the statistically significant difference in average age of the 2 groups (71.9 vs 53.7 years, P < .00001). Therefore, larger studies are warranted to further explore the optimal treatment regimen for new-onset insomnia during hospitalization within this population.
It has been shown that sleep disturbance in the inpatient setting is caused by a variety of physical and psychological factors and leads to a diagnosis of new-onset insomnia in approximately 36% of hospitalized patients. 9 Of note, a study performed by Ho et al found that insomnia diagnosed during a hospital admission often resolves within 2 weeks of discharge. 9 Both pharmacologic and non-pharmacologic interventions are used in attempt to treat this diagnosis. A study performed by Frighetto et al found that hypnotic prescriptions are initiated on approximately 30% of hospitalized patients that have no history of hypnotic use prior to admission. 10 When considering treatment for new onset insomnia, it is important to consider both the short-term and long-term implications of pharmacologic therapy and to demonstrate responsible prescribing. A retrospective review identified that about 35% of patients that are newly diagnosed with insomnia during their hospitalization are discharged home on a sleep aid. 11 Results of our study aligned those in previous literature with a total of 64 (40.5%) patients being prescribed a new sleep aid at discharge, and was similar between treatment groups (39.4% vs 46.2%; P = .52). Although a significant number of patients were prescribed a new sleep aid at discharge, only 6 (3.8%) received a new diagnosis of insomnia at discharge. These findings further support that judicious prescribing of sleep aids at discharge and appropriate follow-up should be a focus in clinical practice in order to limit unnecessary long-term impact from pharmacologic agents, specifically in relation to new-onset insomnia.
In studies and clinical practice, the importance of incorporating non-pharmacologic interventions in the management of insomnia in the hospital setting has been recognized. 4 Interestingly, our study found that patients who were treated with trazodone were less likely to have documentation of non-pharmacologic interventions compared to those treated with melatonin. At our institution there is no standard non-pharmacologic intervention and is up to the discretion of the provider. This impacted the use and documentation of nonpharmacologic options in our study. In addition to nonpharmacologic interventions, studies evaluating insomnia and delirium in the hospital setting have evaluated the contribution and adverse effects that benzodiazepines display in those conditions. 7 When evaluating the administration of an additional sleep aid, it was found that 5 (55.6%) patients in the trazodone group received a benzodiazepine as their additional sleep aid compared to 6 patients (22.6%) in the melatonin group (P = .11). While the difference is not statistically different, like the primary outcome, it should be noted that the results trended toward a numerical difference and further studies may be warranted to find a true difference.
This study was strengthened by its consideration of potential confounding variables such as administration of medications that may cause insomnia, hospital unit admitted to, and nonpharmacologic interventions, none of which were statistically significant between the 2 treatment groups. This study also evaluated adverse effects and potentially unfavorable outcomes in addition to the primary outcome that allows prescribers to assess both efficacy and safety of these agents. Rates of adverse effects and the prescribing of a new sleep aid at discharge were similar between the 2 groups, however evaluation of these outcomes may also be limited by the sample size and the potential lack of documentation of adverse drug reactions based on the retrospective nature of this study, therefore a true difference between them may not have been appropriately identified.
This study was limited by its retrospective nature which provides opportunity for data collection error and missing information due to lack of documentation. One area that this may have impacted the inclusion or exclusion of a patient from the study was the accuracy of the admission medication reconciliation. As previously mentioned, our outcomes should be considered exploratory as the results of the outcomes were limited by the sample size and may not have met power to detect. A total of 132 patients were included in the melatonin group and 26 patients in the trazodone group, however other sleep aid agents such as ramelteon and zolpidem were not able to be assessed within this study due to a limited number of patients meeting study eligibility for either of those agents. The majority of the patients included in this study were white with an average age of 68.9 years, and therefore may limit the external validity of our findings. When evaluating the administration of sleep aids, as-needed doses that were administered prior to the initiation of a scheduled sleep aid were not included in this study and therefore may have led to an inadequate evaluation of the total number of sleep aid doses.
Conclusion
A variety of factors can contribute to the optimal management of new-onset insomnia in hospitalized adult patients. While there was no statistically significant difference between the 2 treatment groups in terms of the primary outcome, this study suggests that melatonin may have a more favorable treatment outcome than trazodone based on the need for an additional sleep aid. Further research is needed to assess the true efficacy of melatonin and trazodone as well as comparing the 2 agents to other sleep aid agents available. In regards to other aspects of treatment optimization, there was no difference found in the rates of adverse events, however patients treated with trazodone were less likely to have documentation of nonpharmacologic treatment interventions. Similar to efficacy outcomes, a larger study is warranted to determine the true difference in safety between the 2 agents.
Footnotes
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding: The author(s) received no financial support for the research, authorship, and/or publication of this article.
ORCID iDs: Emily Letsinger 
https://orcid.org/0000-0002-8646-6148
Lindsay Saum
https://orcid.org/0000-0001-8595-2059
References
- 1. Young JS, Bourgeois JA, Hilty DM, Hardin KA. Sleep in hospitalized medical patients, part 1: factors affecting sleep. J Hosp Med. 2008;3(6):473-482. doi: 10.1002/jhm.372 [DOI] [PubMed] [Google Scholar]
- 2. Thille AW, Reynaud F, Marie D, et al. Impact of sleep alterations on weaning duration in mechanically ventilated patients: a prospective study. Eur Respir J. 2018;51(4):1702465. doi: 10.1183/13993003.02465-2017 [DOI] [PubMed] [Google Scholar]
- 3. DePietro RH, Knutson KL, Spampinato L, et al. Association between inpatient sleep loss and hyperglycemia of hospitalization. Diabetes Care. 2017;40(2):188-193. doi: 10.2337/dc16-1683 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Huang HW, Zheng BL, Jiang L, et al. Effect of oral melatonin and wearing earplugs and eye masks on nocturnal sleep in healthy subjects in a simulated intensive care unit environment: which might be a more promising strategy for ICU sleep deprivation? Crit Care. 2015;19(1):124-211. doi: 10.1186/s13054-015-0842-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Lenhart SE, Buysse DJ. Treatment of insomnia in hospitalized patients. Ann Pharmacother. 2001;35:1449-1457. doi: 10.1345/aph.1a040 [DOI] [PubMed] [Google Scholar]
- 6. Kelly J. Insomnia treatment for the medically ill hospitalized patient. Ment Health Clin. 2014;4(2):82-90. doi: 10.9740/mhc.n190102 [DOI] [Google Scholar]
- 7. Joseph SG. Melatonin supplementation for the prevention of hospital-associated delirium. Ment Health Clin. 2017;7(4):143-146. doi: 10.9740/mhc.2017.07.143 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Morse AM, Bender E. Sleep in hospitalized patients. Clocks Sleep. 2019;1:151-165. doi: 10.3390/clockssleep1010014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Ho A, Raja B, Waldhorn R, Baez V, Mohammed I. New onset of insomnia in hospitalized patients in general medical wards: incidence, causes, and resolution rate. J Community Hosp Intern Med Perspect. 2017;7(5):309-313. doi: 10.1080/20009666.2017.1374108 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Frighetto L, Marra C, Bandali S, Wilbur K, Naumann T, Jewesson P. An assessment of quality of sleep and the use of drugs with sedating properties in hospitalized adult patients. Health Qual Life Outcomes. 2004;2(1):17-10. doi: 10.1186/1477-7525-2-17 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Gillis CM, Poyant JO, Degrado JR, Ye L, Anger KE, Owens RL. Inpatient pharmacological sleep aid utilization is common at a tertiary medical center. J Hosp Med. 2014;9(10):652-657. doi: 10.1002/jhm.2246 [DOI] [PubMed] [Google Scholar]