Melatonin to improve sleep in patients undergoing major surgery: protocol for a double-blind, randomised placebo-controlled trial (MELROSE)

Abstract

Introduction

Sleep disturbance is common after major surgery and may impair recovery, immune function and postoperative outcomes. Environmental disruptions, pain and routine hospital care often contribute to poor sleep during hospitalisation. Sedative medications such as benzodiazepines are frequently used but may cause adverse effects. Melatonin, a hormone regulating circadian rhythm, has been shown to improve sleep disturbance with a favourable safety profile; however, evidence for its perioperative use remains limited.

Methods and analysis

The MELROSE trial is a prospective, double-blind, randomised placebo-controlled trial conducted at John Hunter Hospital. Adults undergoing elective major abdominal or orthopaedic surgery with an expected hospital stay of more than 48 hours are randomised 1:1 to receive melatonin (5 mg sublingual wafer) or placebo. Study medication is administered the night before surgery and nightly postoperatively at 21:00 hours for up to 7 days or until discharge. The primary outcome is the raw Patient-Reported Outcome Measurement Information System (PROMIS) Sleep Disturbance 8a score on postoperative day 3. Secondary outcomes include serum cortisol levels, postoperative complications (Clavien-Dindo classification), length of stay, delirium incidence, 30-day mortality and quality of life with the EuroQuol 5 Dimension 5 levels quality of life scale(EQ-5D-5L). The study aims to recruit 340 participants, with 170 participants per treatment group. The primary analysis will compare treatment groups using an Analysis of Covariance (ANCOVA) model adjusted for baseline PROMIS score and randomisation strata.

Ethics and dissemination

This study has been approved by the Hunter New England Human Research Ethics Committee (2019/ETH13791) protocol V11_ 22 January 2026. Study findings will be presented at local and national conferences and in scientific research journals.

Trial registration number

ACTRN12620000031965.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• Double-blind, randomised placebo-controlled design provides high-quality evidence on the effectiveness of melatonin for postoperative sleep.

• Sleep disturbance is assessed using the validated Patient-Reported Outcome Measurement Information System (PROMIS) Sleep Disturbance 8a instrument, ensuring standardised patient-reported outcomes.

• Evaluates a low-cost, widely available intervention with a favourable safety profile, supporting potential translation into routine perioperative care.

• Heterogeneous surgical population and exclusion of non-English-speaking patients limit generalisability.

• Formal chronotype, habitual sleep timing and diagnostic insomnia status are not assessed. Because melatonin’s chronobiological effect depends partly on timing relative to endogenous circadian phase, fixed administration at 21:00 may have variable effects between individuals.

Background and rationale

Quantity and quality of sleep is a fundamental biological requirement for human health. The duration of this quality of sleep in addition to quantity in a hospitalised patient after major abdominal or orthopaedic surgeries can be compromised at a time where healing is crucial in the recovery phase. Environmental factors such as noise, lighting, alarms and routine clinical observations frequently disrupt sleep in hospital settings, while pain and physiological stress related to surgery may further impair sleep disturbance.¹ These disturbances may impair immune responses, increase inflammatory activity and potentially worsen postoperative recovery.

Postoperative sleep disturbance is characterised by reduced total sleep time, sleep fragmentation, suppression of slow-wave and rapid-eye-movement sleep, and disruption of normal circadian organisation. Pain, surgical stress, anaesthetic and analgesic medications, environmental disturbance and overnight clinical care may all contribute.² Poor sleep during hospitalisation may have important physiological consequences. Experimental studies have demonstrated that sleep disruption can impair immune function, increase inflammatory responses and alter endocrine regulation.^{1 3} These effects are particularly relevant in the perioperative setting, where an optimal immune and inflammatory response is important for wound healing and recovery. Sleep disruption has also been associated with delirium, prolonged hospitalisation and poorer postoperative outcomes.⁴

What is sleep disturbance, quantity and circadian disruption

Distinguishing between sleep constructs:⁵

• Sleep quantity: total duration of sleep.

• Sleep disturbance or disturbance: perceived difficulty initiating or maintaining sleep, restlessness, non-restorative sleep and satisfaction with sleep.

• Sleep architecture: physiological distribution of sleep stages.

• Circadian disruption: misalignment or attenuation of endogenous biological rhythms, including melatonin and cortisol rhythms.

Sedative medications such as benzodiazepines are commonly used to manage sleep disturbance in hospitalised patients; however, these agents may cause adverse effects, including respiratory depression, cognitive impairment and delirium, particularly in older surgical populations.^{6 7} Safer pharmacological strategies to improve sleep during the perioperative period are therefore needed. Melatonin is a naturally occurring hormone that plays an important role in regulating the sleep–wake cycle.⁸ It may provide a safer alternative for managing sleep disturbance in hospitalised patients because of its favourable safety profile and low toxicity.

Previous trials of perioperative melatonin have generally been small and heterogeneous, and systematic reviews have produced inconsistent conclusions. This uncertainty supports the need for a sufficiently powered, placebo-controlled trial using a standardised patient-reported measure of postoperative sleep disturbance.⁹

Melatonin is an endogenous hormone that regulates circadian rhythm and the sleep–wake cycle (online supplemental file 1). In healthy individuals, melatonin secretion increases in the evening, peaks during the night and declines towards morning.³ Hospital environments may disrupt normal melatonin secretion due to exposure to artificial light at night and irregular sleep schedules. Exogenous melatonin has demonstrated benefits for sleep onset and sleep disturbance in several clinical populations and has a favourable safety profile compared with traditional sedative medications.^{10 11}

Previous studies investigating melatonin in hospitalised patients have primarily focused on intensive care populations or prevention of delirium.^12–14 Evidence for its use in the perioperative setting remains limited, with small studies and heterogeneous populations producing inconsistent results. Although melatonin has demonstrated potential benefits for sleep disturbance and anti-inflammatory effects, robust evidence in patients undergoing major surgery is lacking. Given the high prevalence of postoperative sleep disturbance and the potential consequences for recovery, interventions that improve sleep may enhance perioperative outcomes. Melatonin represents a low-cost and well-tolerated intervention that could be readily implemented in routine care if effective.

Knowledge gap

Despite the high prevalence of postoperative sleep disturbance, evidence supporting pharmacological strategies to improve sleep in surgical patients remains limited. To our knowledge, no large randomised controlled trial has evaluated perioperative melatonin for improving postoperative sleep disturbance in patients undergoing major abdominal or orthopaedic surgery.

Objective

The main objective is to determine whether perioperative administration of melatonin improves sleep disturbance in patients undergoing major abdominal or orthopaedic surgery.

Secondary aims are to assess whether melatonin reduces surgical stress response, postoperative complications, delirium, hospital length of stay and 30-day mortality and improves postoperative quality of life.

Methods

Trial design

This study is a two-arm, prospective, double-blind, randomised placebo-controlled trial comparing perioperative melatonin with placebo in patients undergoing major abdominal or orthopaedic surgery. Participants are randomised in a 1:1 ratio to melatonin or placebo. Patients are stratified by planned surgical approach, open versus laparoscopic and specialty gynaecology/gynae oncology, upper gastrointestinal (GI) or lower GI surgery and hip or knee orthopaedic surgery between August 2020 and December 2026.

Study setting

This is a single-centre study at a tertiary referral hospital within the Hunter New England Local Health District, Newcastle, New South Wales, Australia.

Inclusion criteria

• All English-speaking patients 18 years and over undergoing elective/semi-elective major abdominal or orthopaedic (including hips and knees) surgery with an anticipated length of stay greater than 48 hours.

Exclusion criteria

Patients from the above group will be excluded if they:

• Are unable to give written informed consent to take part in the study.

• Patients who require the use of a trained healthcare interpreter to be able to understand study information.

• Lack of decision-making capacity.

• Expected stay less than 48 hours.

• Concurrent use of benzodiazepine and non-benzodiazepine (Z-drugs).

• Alcohol consumption >13 units/week for women and >20 units/week for men.

• Currently taking melatonin.

• Drug interaction with melatonin.

• Allergy to melatonin or its excipients (may contain lactose).

• Decreased renal function as defined by an estimated glomerular filtration rate (eGFR) of <30 mL/min/1.73 m².

• Decreased liver function defined as an aspartate aminotransferase (AST) >500 U/L or alanine aminotransferaseALT >500 U/L.

• Pregnant or breastfeeding females.

• Enrolment in other clinical trials involving an investigation agent.

• Taking quinolones, carbamazepine and rifampicin, fluvoxamine, 5-methoxypsoralen or 8-methoxypsoralen, cimetidine and oestrogens.

Recruitment and consent

Potential participants are identified by treating surgeons or research nurses through review of outpatient and operating theatre lists. Patients scheduled for elective major abdominal or orthopaedic surgery are screened for eligibility using predefined criteria and medical record review. Eligible patients are approached in the perioperative clinic or contacted by telephone prior to surgery and provided with a Human Research Ethics Committee-approved Participant Information Statement and Consent Form. Written informed consent is obtained before enrolment either in person or electronically via Research Electronic Data Capture (REDCap) e-consent.^{15 16}

Withdrawal protocol

Participants are withdrawn from the trial if:

• Adverse reaction to the medication melatonin.

• If the patient no longer wants to participate in the study.

• Clinical decision by treating surgeon.

Patient and public involvement statement

Patients and the public were not involved in the design, conduct, reporting or dissemination plans of this study. However, the study includes patient-reported outcome measures of sleep disturbance and quality of life to ensure that outcomes important to patients undergoing major surgery are evaluated. The findings of this study will be disseminated through peer-reviewed publications and presentations at scientific meetings.

Intervention and control groups

Preoperative protocol

All patients undergo a standardised preoperative preparation for surgery as per the usual protocol of the treating surgeons which will include routine ECG for all patients, chest X-ray and blood tests including full blood count, urea, creatinine and electrolytes and liver function tests. online supplemental file 2.

Participants are randomised and dispensed with a single dose of study drug (5 mg melatonin/placebo) in wafer form at the Perioperative Clinic, to take home and take sublingually the night prior to surgery or if in hospital at 21:00 hours as per instructions (online supplemental file 3).

For participants that have had surgery postponed after having taken the pre-op dose the night prior to planned surgery date, another single dose will be dispensed to the participant for ingesting the night prior to the new scheduled date of surgery. With a reported biological half-life of 35–50 min, it is felt safe to be able to do this.^{17 18} The additional drug will be the same as the randomised arm and participant will complete the Patient-Reported Outcome Measurement Information System (PROMIS) Sleep disturbance Questionnaire 8A again.¹⁹ All participants will receive a reminder SMS the day prior to operation to remind them to take study drug and complete PROMIS questionnaire either on paper or via emailed link.

Postoperative

Day 0—discharge from hospital

Participants are prescribed 5 mg melatonin/placebo in wafer form sublingually via Medchart, administered at 21:00 hours daily commencing the postoperative night for a maximum of 7 days while in hospital, or until discharge if <7 days. Compliance to study treatments is assessed by review of Medchart and medical record. Standard care follows Enhanced Recovery After Surgery pathways. online supplemental file 2.

Surgical follow-up as per treating team, at day 30 the patient will be contacted by phone by research nurse to assess complications and a reminder to complete the EuroQuol 5 Dimension 5 levels(EQ-5D-5L) quality of life questionnaire, via post or online.²⁰ Study-related questionnaires are to be attended in Perioperative Clinic and the night before surgery, day 1 to discharge (max to day 7) (online supplemental file 3).

Primary outcome measure

The primary efficacy outcome is a difference in PROMIS Sleep Disturbance 8a score at day 3.¹⁹

Secondary outcome measures

1. Mean difference between groups for serum cortisol levels at baseline, day of surgery and days 1, 2 and 3 postoperatively.

2. Incidence of postoperative surgical complications graded by the Clavien-Dindo Classification of Surgical Complications up to day 30 post-op as recorded in medical record.

3. Difference between groups for mean length of hospital stay (days) as recorded in medical record for index admission, up to the time of suitability for surgical discharge to home or rehabilitation.

4. All-cause mortality by day 30 as recorded in medical record or reported by treating surgeon.

5. Incidence of delirium as defined by a diagnosis of same by treating team assessed using daily the “Alertness, Abbreviated Mental test, Attention, and Acute Change” (4AT) assessment in ward patients or Confusion Assessment Method for ICU (CAM-ICU) assessment in the Intensive Care Unit(ICU) patients. Medical record documentation of delirium will be collected as an additional measure up to the time of hospital discharge or discharge from surgical service.

6. Between-group difference in EQ-5D-5L index score and Euroqol (EQ) Visual Analogue Scale from baseline to day 30.

Sample size

Previous studies of the PROMIS sleep disturbance scale in hospitalised patients have shown an overall raw score mean of 9 points and SD of 9 points.¹⁹ Allowing for a 10% dropout rate, a 15% inflation in sample size to account for potentially skewed data and a correlation between baseline and follow-up of 0.6, a sample of 170 patients per arm at baseline will give the study 80% power to detect a clinically relevant decrease of 2.5 points in the intervention arm, with 80% power and 5% type 1 error rate.

Randomisation

After eligibility has been determined, patients will be randomised using REDCap,^{15 16} using a randomisation schedule generated by an independent statistician at the Hunter Medical Research Institute Clinical Research Design and Statistical Services. This provides a centralised, computer-based randomisation system. Treatments will be allocated equal ratios (1:1) within stratum using permuted blocks of size 4. Stratification will be for open versus laparoscopic surgery, and surgery type gynaecology, upper GI surgery or lower GI surgery or orthopaedic hip or knee surgery. Where a patient has a combined procedure, the primary intent of surgery will determine the surgery type.

Blinding

Participants, treating clinicians, research staff and outcome assessors will remain blinded to treatment allocation throughout the study. Melatonin and placebo wafers will be identical in appearance and packaging.

Enhanced Recovery After Surgery

Participants receive the specialty-specific perioperative pathway used by their treating service. These pathways are not identical across surgical specialties. Randomisation is stratified by surgical category and operative approach. Where routinely available, adherence to key components relevant to postoperative recovery and sleep—including mobilisation, oral intake, removal of invasive devices, opioid-sparing analgesia and avoidance of prolonged fasting—will be recorded and summarised by treatment group.

Data collection

Baseline demographic and clinical data are collected from the medical record and perioperative clinic assessment. Study outcomes are obtained through a combination of patient-reported questionnaires and medical record review.

Patient-Reported Outcome Measurement Information System Sleep Disturbance scoring and administration¹⁹

Sleep disturbance is assessed using the eight items from the PROMIS Sleep Disturbance 8a instrument. For the purposes of this study, the wording has been adapted so that participants answer each item with reference specifically to their sleep during the previous night. The postoperative day 3 assessment therefore refers to sleep during the night immediately preceding the assessment.

Questionnaires are administered once daily from postoperative day 1 to postoperative day 7 and completed before 12:00 midday. The primary outcome is the raw PROMIS Sleep Disturbance score at postoperative day 3. Item responses will be summed to produce a raw score ranging from 8 to 40, with higher scores indicating greater sleep disturbance.

A raw total score will be calculated only when all eight items have been completed. Where one or more items are missing, the raw total score will be recorded as missing and managed in accordance with the prespecified statistical analysis plan. PROMIS T-scores may be reported as a supplementary outcome using the applicable PROMIS scoring method, but the primary analysis and sample size calculation are based on raw scores.

The study is powered to detect an adjusted mean between-group difference of 2.5 raw PROMIS points. As a clinically meaningful difference has not been established for this modified previous-night version of the PROMIS Sleep Disturbance 8a instrument in an acute postoperative population, 2.5 points will be described as the prespecified target difference rather than a definitively established minimal clinically important difference.

Delirium

Incidence of delirium from Postoperative day (POD)1 to POD7 or hospital discharge, whichever occurs first, was identified using daily 4AT assessment in ward patients or CAM-ICU assessment in ICU patients. Medical record documentation of delirium will be collected as an additional measure.

Cortisol

Serum cortisol samples are collected between 07:00 and 09:00 at baseline, day of surgery and days 1, 2 and 3 postoperatively, with exact collection time recorded. Fasting status is documented but not mandated. Blood samples for cortisol measurement will be processed by the accredited hospital pathology laboratory using its validated routine clinical cortisol assay. The analytical platform, assay methodology, analytical range, precision and relevant quality control characteristics will be recorded and reported in the study findings. Administration of systemic corticosteroids, including dexamethasone used for postoperative nausea and vomiting prophylaxis, will be recorded with dose and timing.

Additional clinical outcomes, including postoperative complications, delirium, length of hospital stay and mortality, will be extracted from the electronic medical record. Quality of life will be measured using the EQ-5D-5L questionnaire.²⁰ All data will be entered into a secure REDCap electronic database with built-in validation and audit trails.

Data management

A REDCap^{15 16} database has been established through Hunter Medical Research Institute. REDCap is a secure, web-based software platform designed to support data capture for research studies. REDCap functions as a secure online tool, specifically designed for capturing data in research studies. It offers an easy-to-use interface for validated data entry, tracks data manipulation through audit trails, facilitates automated export processes for smooth data retrieval into common statistical packages and provides methods for seamlessly integrating and interacting with external data sources. Data will be analysed and reported in a deidentified and grouped format by the Clinical Research Design and Statistics team at Hunter Medical Research Institute (HMRI) using data sourced from the REDCap database.

Statistics analysis

The primary analysis will be in the intention-to-treat population. Participant characteristics will be summarised separately by treatment arm, reporting means and SDs (or medians and IQRs) for continuous variables, and counts and percentages for categorical variables. Differences between treatment arms in mean PROMIS scores will be assessed using an Analysis of Covariance (ANCOVA) linear regression model, adjusting for the baseline value of the outcome, stratification variables and the treatment variable (placebo as referent group). Least-square mean differences between treatment arms will be presented together with 95% CIs and p values. Multiple imputation will be used as the primary method of dealing with missing data.

Secondary outcomes will be compared between treatment arms using generalised linear models, with appropriate link and distribution families depending on the outcome type. Length of hospital stay is expected to have a right-skewed distribution and will therefore be analysed using a generalised linear model. A gamma distribution with a log-link function will be used, where supported by prespecified model diagnostics. Model fit will be assessed by examining the observed distribution, residuals and goodness-of-fit statistics. Where the gamma model provides an inadequate fit, an alternative prespecified distribution, such as a log-normal model, will be used and the rationale documented.

Rationale for Postoperative day 3

Postoperative day 3 was prospectively selected as the primary assessment time because it represents the early postoperative period after three postoperative study doses, while being less immediately affected than POD1 by residual anaesthesia, acute postoperative instability and incomplete recovery. Sleep disturbance is frequently most marked during the first several postoperative nights. PROMIS Sleep Disturbance scores collected on POD1–7 will additionally be examined in a prespecified longitudinal secondary analysis to characterise the trajectory of postoperative sleep.

Perioperative management will follow usual specialty-specific care and will not be standardised by the study protocol. To permit assessment of important co-interventions, the following variables will be collected where available: anaesthetic technique, use of neuraxial or regional analgesia, cumulative opioid administration, ketamine, dexmedetomidine, systemic corticosteroids, postoperative sedation or mechanical ventilation, and admission to intensive or high-dependency care. These variables will be summarised by randomised group. The primary analysis will remain adjusted for baseline PROMIS score and randomisation strata. A prespecified sensitivity analysis will additionally adjust for a limited set of clinically important perioperative variables selected before unblinding.

Subgroup analysis

To maintain allocation balance, randomisation is stratified by surgical approach and surgical category. The primary estimand is the adjusted mean between-group difference in the postoperative day 3 PROMIS Sleep Disturbance score among all randomised participants. Participants will be analysed according to their allocated treatment group, regardless of treatment adherence or protocol deviations, consistent with the intention-to-treat principle. The primary estimand therefore represents the overall average treatment effect of melatonin compared with placebo across the enrolled study population.

Prespecified subgroup analyses will assess whether the treatment effect differs according to:

1. Abdominal or gynaecological surgery versus orthopaedic surgery.

2. Open versus minimally invasive abdominal surgery.

3. Individual randomisation strata, where sample size permits.

4. Intensive care or high-dependency unit admission versus ward-only postoperative care.

5. Baseline sleep disturbance above versus below a prespecified clinically meaningful threshold, where available, or otherwise the cohort median.

6. Age group, where clinically justified and specified before treatment allocation is unblinded.

Subgroup analyses will be considered exploratory and will be evaluated using treatment-by-subgroup interaction terms rather than separate tests of treatment effect within each subgroup. Prespecified exploratory subgroup analyses will assess whether the treatment effect differs by surgical category and operative approach using interaction terms between treatment allocation and subgroup. Subgroup estimates will be reported with 95% CIs. The study is not powered to detect subgroup interactions, and these analyses will be interpreted cautiously.

No multiplicity adjustment will be applied to secondary outcomes because these analyses are exploratory. Emphasis will be placed on effect estimates and CIs rather than isolated p values. Findings from secondary and subgroup analyses will be interpreted as hypothesis-generating.

Trial monitoring

Monitoring will include centralised review of electronic case report forms (eCRFs) and other study documents for protocol compliance, data accuracy and completeness. Central data monitoring will occur every 6 months, with review of recruitment, consent, eligibility, treatment adherence, missing outcome data, adverse events and data completeness. The monitoring staff will be given direct access to source documents, eCRFs and other study-related documents. By signing the informed consent form, the participant gives consent direct access to their medical records and the study data by study monitors and auditors.

Data monitoring committee

A formal data monitoring committee has not been established. The intervention under investigation (melatonin) has a well-established safety profile and the trial involves minimal risk to participants. All SAEs will be reviewed by the principal investigator and reported to the approving Human Research Ethics Committee in accordance with institutional requirements.

Safety reporting

All potential SAEs will be reported immediately to the principal investigator who will report to the approving Human Research Ethics Committee within 72 hours. Adverse events will also be reported annually to the Human Research Ethics Committee (HREC) and be reported and discussed in any publication that may result from this research.

Serious adverse event definition

Any untoward and unexpected medical occurrence or effect that²¹:

• Results in death.

• Is life threatening.

• Requires hospitalisation, or prolongation of existing inpatient’s hospitalisation.

• Results in persistent or significant disability or incapacity.

Deviations

Protocol deviations will be prospectively classified before unblinding as major or minor. Major deviations may include:

• Ineligibility discovered after randomisation.

• Receipt of the incorrect treatment.

• Failure to receive any study medication.

• Unblinding before outcome ascertainment.

• Primary-outcome assessment outside the prespecified window.

• Prohibited concomitant medication.

The primary analysis will remain intention-to-treat. A prespecified per-protocol sensitivity analysis may exclude participants with major deviations affecting exposure to the intervention or ascertainment of the primary endpoint. The number and type of deviations will be reported by treatment group.

Data safety and monitoring

A general physician with clinical pharmacology training will be appointed to review all SAEs and make a determination as to the relationship of the SAE and the study medication. This will be reported to the approving HREC. There will be no interim analysis of this project due to small accrual target.

Data sharing and access

Following publication of the primary study findings, the investigators will consider requests for access to deidentified individual participant data that underlie the reported results. Supporting documentation, including the study protocol, data dictionary and statistical analysis plan, may also be made available where appropriate.

Data will not be made publicly available in an open repository. Researchers seeking access must submit a written request to the principal investigator outlining the proposed research question, objectives, analysis plan, qualifications of the research team and intended use of the data. Requests will be assessed by the principal investigator and study team on the basis of scientific merit, feasibility, participant confidentiality, ethical requirements and consistency with the original participant consent.

Data will only be shared following receipt of any required ethics and institutional approvals and completion of a formal data sharing agreement. The agreement will specify the approved use of the data, data security requirements, publication arrangements and requirements relating to the destruction or return of the data following completion of the approved analyses.

Only deidentified data necessary to address the approved research question will be provided. Direct participant identifiers will not be shared. Data sharing will remain subject to applicable privacy legislation, institutional data governance requirements and any restrictions imposed by the approving Human Research Ethics Committee.

Ethics and dissemination

This study will be conducted according to the Note for Guidance on Good Clinical Practice (CPMP/ICH/135/95) in compliance with applicable laws and regulations. The study will be performed in accordance with the National Health and Medical Research Council (NHMRC) Statement on Ethical Conduct in Research Involving Humans (Commonwealth of Australia 2007) and the principles laid down by the World Medical Association in 2008. The investigators shall comply with the protocol, except when a protocol deviation is required to eliminate an immediate hazard to a participant. This study has been approved by the Hunter New England Human Research Ethics Committee (2019/ETH13791). The study’s findings will be communicated locally to hospital teams and shared on a national scale through conferences and presentations. Additionally, results will be disseminated via publication in scientific journals. Deidentified and summarised outcomes will be provided to external parties upon specific requests made to the chief investigator.