Things We Do for No Reason: Pharmacological Sleep Aids for Hospitalized Patients with Acute Insomnia

Background

Acute inpatient insomnia (AII) is common. Half of hospitalized patients report sleep disturbances, including more sleep interruptions and inferior sleep quality compared to home.¹ A 2017 study of 205 patients with no history of insomnia found that 36% developed new-onset insomnia while hospitalized.² Hospitalization affects sleep via multiple mechanisms including disruptive sleep environments, discomfort, anxiety, and medications that adversely affect sleep.^{1, 3} In turn, poor sleep precipitates worse health outcomes in hospitalized patients: impaired recovery, increased medical complications, functional impairment, delirium, and increased mortality.^1–3

The use of pharmacological sleep aids (PSAs) for the treatment of insomnia is common; however, the FDA has only approved PSAs for the treatment of chronic insomnia in the outpatient setting.^3,4 No medication has FDA approval for the treatment of AII.

PSAs are a heterogeneous group of medications. We define PSAs as benzodiazepines, benzodiazepine receptor agonists (e.g., zolpidem, eszopiclone, and zaleplon), barbiturates, tricyclic antidepressants, trazodone, centrally-acting antihistamines and second-generation antipsychotic medications (SGAs).^4–6 This definition excludes the melatonin receptor agonists (MRAs) melatonin and ramelteon.⁴

Why one might think PSAs are helpful

The high prevalence and potential adverse outcomes of AII may compel providers to provide a quick solution to hospitalized patients’ sleep concerns in the form of PSAs.^{1, 5} Indeed, providers commonly prescribe PSAs to hospitalized patients.^4–6 One retrospective study in a tertiary medical center found that 168 of 642 adult patients admitted to the general medical and surgical wards from January 2013-February 2013 received PSAs (26.2%), though the majority had no history of insomnia.⁷ A 2003 study of AII treatment found that providers prescribed benzodiazepines in 88% of cases of inpatient insomnia. Another 2017 study of elderly patients at a large urban medical center showed that insomnia complaints accounted for over 80% of newly prescribed benzodiazepines and sedative hypnotics.⁵ Providers prescribed over 60% of these newly prescribed PSAs during overnight hours,⁵ suggesting that they reacted to acute issues.

Recognizing that PSAs can be effective for the treatment of chronic insomnia in the outpatient setting, inpatient providers might assume that the same is true for AII based on knowledge of PSAs’ varied effects on sleep architecture. Benzodiazepines, the most extensively studied PSAs, reduce time to REM sleep, shorten sleep-onset latency, and decrease nocturnal awakenings.⁴ Benzodiazepine receptor agonists have similar effects on sleep but less frequent adverse events.⁴ Trazodone, antihistamines, and SGAs have varying effects on sleep including improved sleep quality, improved sleep maintenance and duration, and increased sedation, respectively.^4,8–10 A growing awareness of harms associated with benzodiazepines may lead providers to preferentially prescribe trazodone or other non-benzodiazepine PSAs perceived as safer alternatives.

Why inpatient sleep aids may not be helpful and are potentially harmful

Despite the ubiquity of PSAs, we lack evidence supporting their use in hospitalized patients, and we have evidence that they may cause harm.¹⁰ A 2016 systematic review summarizing the efficacy and safety of these agents found no consistent trends in their effects on improving sleep with respect to sleep latency, efficiency, quality, or interruptions.⁶ The review evaluated 11 randomized controlled trials and four observational studies comprising ward, ICU, and post-operative patients who received PSA medications including benzodiazepines, benzodiazepine receptor agonists, melatonin receptor agonists, and nonbenzodiazepine sedatives. The review concluded that “larger, better-designed trials in hospitalized adults are needed.”⁶ Similarly, other reviews emphasize wide heterogeneity in treatment of AII with little evidence supporting the efficacy of PSAs in hospitalized patients.^{9, 10} As such, societies have not provided evidence-based recommendations or guidelines for treatment of AII.

Beyond the lack of proven inpatient efficacy, multiple studies cite the potential harm of PSAs.^{4, 5, 9, 11} Studies associate benzodiazepines, benzodiazepine receptor agonists, barbiturates, SGAs, and antihistamines with delirium, cognitive impairment, fall-related injuries, and death, particularly in the elderly. Each of these medication classes appear in the American Geriatrics Society’s Beers Criteria for Potentially Inappropriate Medication Use in Older Adults.^{4, 5, 9} For instance, one observational study found an association between inpatient use of zolpidem and falls, a major driver of patient morbidity and hospital costs, with a hazard ratio of 4.37.¹² This finding is particularly concerning given that providers prescribe PSAs four times more commonly for patients aged 80 and older as compared to patients aged 20–39 years.¹³

Moreover, patients initiated on a PSA in the hospital often receive a discharge prescription for this medication. In one retrospective study, 40 of 168 (34%) hospitalized patients started on a PSA for the first time left the hospital with a new PSA prescription, despite a lack of evidence that AII persists after discharge.^{2, 7} This practice increases patients’ risk of the long-term use of these agents following discharge, predisposing them to further complications including those related to polypharmacy and dependence.¹² Inappropriate PSA prescriptions cause harm to patients and the healthcare system at large. One study assessed the national financial burden from associated mortality and adverse drug effects of all PSAs to exceed 10 billion dollars per year.¹⁴

What we should do instead

Providers should begin investigations of AII with a focused patient interview. Questions should elicit historical characteristics of a patient’s insomnia, current and prior medications, and relevant history including substance use and medical comorbidities (e.g., undiagnosed obstructive sleep apnea).^{3, 9, 12} Providers should then identify factors interfering with sleep, including the most common sleep-disturbing factors in the hospital—noise and interruptions by clinical staff as well as inadequately controlled pain.^{1, 12, 15}

Rather than prescribing PSAs, hospitalists should address causes of hospital-acquired insomnia by managing patients’ acute illness and pain, treating withdrawal, and pursuing environmental modifications and bundling of care (Table 1). A 2018 prospective study of two general medicine units at a single institution found that a multi-disciplinary intervention of staff education and electronic medical record prompts aimed at improving patient sleep significantly reduced nightly interruptions by 44% and improved patient-reported outcomes of sleep preservation and pain control.¹⁵ Another 2019 study found that holistic, noise-reduction interventions reduce inpatient PSA prescriptions by 18–21%.¹² Importantly, evidence suggests that patients welcome these efforts. A 2016 study of 100 cases of AII found that 67% of patients were open to trying NPAs, though only 11% of patients were offered information on them.⁸

Table 1:

Summary of evidence for non-PSA interventions for sleep promotion.

Evidence for Non-PSA Interventions to Promote Sleep
Intervention	Study Description	Outcome
Eye masking and earplugs in ICU	Cochrane Systematic Review of 30 trials, low quality evidence.12	Lower incidence of delirium (HR 0.55, p = 0.002); increased sleep time (2.19 hours, p = 0.02)
Sleep hygiene protocol among adult inpatients	Systematic Review of 13 intervention studies, including 4 RCTs, low quality evidence12	PRN sedative use decreased by 49% (p = 0.004).
Sleep hygiene protocol among ICU patients	Pre and post study of a bundled intervention targeting fewer sleep interruptions and less noise.12	Decreased noise ratings (mean ± SD: 65.9 ± 26.6 vs. 60.5 ± 26.3, p = 0.001); decreased delirium/coma (OR 0.46; 95% CI, 0.23–0.89; p = 0.02); delirium/coma-free days (OR: 1.64; 95% CI 1.04 – 2.58; p = 0.03)
Bundled intervention of staff education and electronic medical record prompts to reduce nighttime awakening	Prospective, 2-center cohort study of 1083 patients15	44% less nighttime room entry (p < 0.001); fewer disruptions for nighttime vital signs (70% vs 41%, p = 0.05) or medications (84% vs 57%, p = 0.031); survey scores increased by 7% for “Quiet at night” and 9% for “Pain well controlled” categories
“Quiet Hours” Protocol introduced, email prompts sent to physicians and staff.	prospective 8-month pre-post intervention. N = 161 pre-intervention, N = 106 post-intervention13	Decrease in PRN sedative medications from 37.1% to 16% (95% CI, 0.056 – 0.26) among all patients.
Nonpharmacologic sleep aid promotion to both staff and patients with video clips and written material.	prospective 2-month pre-post intervention. N = 118475 pre-intervention, N = 120,973 post-intervention.13	Decrease in proportion of patients previously prescribed PSAs who discharged with a prescription for PSAs from 57.0% to 46.8% (RR 0.82; 95% CI, 0.79 – 0.86).
Nonpharmacologic sleep protocol including warm drinks and massages for patients with AII.	Prospective 6-month pre-post study. N = 94 preintervention, N = 111 post-intervention.13	Decrease in use of sedative-hypnotics 54% to 31% (p < 0.002).
Nursing staff educated on nonpharmacologic sleep aids and bundled care.	Non-blinded randomized trial, N = 59.13	No change in sleep hours. Average number of PSAs prescribed in intervention arm was 1.5 vs 2.2 in control arm (p < 0.05).

Abbreviations: ICU, Intensive Care Unit; PSA, Pharmacologic Sleep Aid; RCT, Randomized Controlled Trial; OR, Odds Ratio; 95% CI, 95% Confidence Interval; RR, Risk Ratio; HR, Hazard Ratio.

If nonpharmacological interventions fail, MRAs may present a safer alternative to PSAs. Limited data suggest that melatonin improves sleep quality, decreases sleep latency, improves sleep efficiency, and increases sleep duration in medical inpatients.¹⁰ According to a review of sleep medication safety among elderly outpatients, ramelteon may decrease sleep latency by 13–29 minutes without causing general central nervous system depression.⁴ MRAs have similar side effects to placebo in randomized controlled trials.⁴ The American Academy of Sleep Medicine has not issued guidance for MRA use but continues to study their efficacy and safety. Last, if insomnia persists despite nonpharmacological interventions and a trial of MRAs, providers should engage in a risk-benefit discussion with the patient before prescribing a PSA for AII.

Hospitalists should advocate for systemic change to optimize patient sleep. Recent evidence-based guidelines from the Society of Hospital Medicine and JAMA Internal Medicine outline such interventions, including care protocols that encourage daytime activity, limit daytime sleep, maintain normal light-dark cycles, modify environmental factors including noise, and minimize nighttime and early-morning patient care activities.^{9, 11, 12} Other recommended interventions include pharmacist-assisted medication review, nurse and physician “champions” of nighttime sleep, and care-team education programs.¹² These systemic nonpharmacological interventions likely decrease the use of PSAs in noncritically-ill hospitalized patients.¹²

Recommendations

• Consider the differential diagnosis for AII including disruptive sleep environments, underlying illness, inadequate pain management, anxiety, medication side effects, and withdrawal.

• Avoid initiating PSAs in hospitalized patients with AII.

• If AII persists despite nonpharmacological interventions, consider prescribing MRAs as a safer alternative to PSAs.

• From a systems standpoint, prioritize creating “sleep-friendly” environments (e.g., light, temperature, and noise) and minimize nocturnal disruptions by bundling patient care and delaying early morning patient care activities.

Conclusion

To prevent harms associated with PSAs, physicians should take a holistic, patient-specific, multi-disciplinary approach to inpatient insomnia, prioritizing non-PSA interventions. In the clinical scenario, the hospitalist responds to the nurse’s page by coming to the bedside to conduct an interview. The hospitalist discovers that the patient has no history of insomnia, but does have inadequately controlled postoperative pain and nocturia. The patient also reports frequent interruptions from vital sign checks and medication administration. The hospitalist optimizes the pain management regimen, stops IV fluids, bundles overnight vital sign checks and medications, and offers eye covers and ear plugs. The hospitalist adds that if these interventions fail to improve her symptoms, they can discuss prescribing an MRA.

What do you do? Do you think this is a low-value practice? Is this truly a “Thing We Do for No Reason”? Let us know what you do in your practice and propose ideas for other “Things We Do for No Reason” topics. Please join in the conversation online at Twitter (#TWDFNR)/Facebook and don’t forget to “Like It” on Facebook or retweet it on Twitter.

Clinical Scenario.

The overnight hospitalist receives a page requesting sleep medication for a 78-year-old patient recovering from a surgical repair of a hip fracture. The patient’s medications include as-needed acetaminophen and maintenance intravenous fluids.

The “Things We Do for No Reason” series reviews practices which have become common parts of hospital care but which may provide little value to our patients. Practices reviewed in the TWDFNR series do not represent “black and white” conclusions or clinical practice standards, but are meant as a starting place for research and active discussions among hospitalists and patients. We invite you to be part of that discussion.