Cognitive and physical exercise to improve outcomes after surgery (COPE-iOS) study: protocol for a randomised, controlled trial in the USA examining the efficacy of a combined cognitive and physical exercise programme performed before and after major surgery in improving cognitive and functional outcomes for older adults

Abstract

Introduction

Surgery and its resulting hospitalisation are associated with subsequent cognitive and functional decline. Interventions to reduce this decline have exhibited limited success. Prehabilitation is the process of enhancing capacity and reserve before an acute stressor to improve tolerance of the acute physiologic insult. Older adults requiring major surgery are an ideal population for prehabilitation. Prehabilitation exercise studies have mostly focused on physical training to improve physical outcomes after specific surgery types, and data on cognitive outcomes and in broader surgical populations are needed. Computerised cognitive training (CCT) has been shown to enhance memory, processing speed, attention and multitasking. Combining CCT with a physical exercise may be most effective in reducing cognitive and functional decline in older patients undergoing major surgery, but has yet to be evaluated.

Methods and analysis

The COgnitive and Physical Exercise to improve Outcomes after Surgery (COPE-iOS) study is a randomised, controlled, participant and assessor blinded clinical trial testing the hypothesis that a pragmatic programme combining CCT and physical exercise throughout the perioperative (ie, preoperative and postoperative) period will improve long-term cognitive and disability outcomes in older surgical patients at high risk for decline. The trial aims to randomise 250 patients who undergo major surgery for a treatment period of approximately 1 month prior to surgery and 3 months after surgery, with a follow-up period of 12 months after surgery. The primary outcome is global cognition at 3 months after surgery. Key secondary outcomes include global cognition at 12 months after surgery and disability in activities of daily living and depression at 3 and 12 months after surgery.

Ethics and dissemination

Trial protocol has been approved by Vanderbilt Human Research Protections Programme (#202496) and an independent Data Safety Monitoring Board. Results will be presented at scientific conferences and submitted for publication.

Trial registration number

ClinicalTrials.gov Registry NCT04889417.

STRENGTHS AND LIMITATIONS OF THIS STUDY

• Randomised, controlled, patient and assessor blinded trial to evaluate the efficacy and safety of remote combined cognitive and physical exercise before and after major surgery in older adults.

• The inclusion of multiple surgery types, use of an active attention control and testing of a remote, scalable exercise intervention for both prehabilitation and rehabilitation.

• Robust assessments of cognition, disability and mental health at baseline and at 3 and 12 months after surgery.

• Surgery cancellations and inability to blind the research personnel performing the intervention and active control procedures.

Introduction

Cognitive decline impacts quality of life and contributes to functional disability and depression.^{1 2} Age is the strongest risk factor for cognitive decline, and maintaining cognitive health with ageing is vital to individuals and the public. Older adults often prioritise preservation of cognition and functional status over survival.³ Over 19 million major surgical cases are performed each year in the United States in patients >65 years of age,⁴ and this number is growing considerably with the ageing of the population. Surgery is performed to improve functional status, reduce pain or prolong life, yet surgery and its resulting hospitalisation are associated with significant brain dysfunction.⁵,¹¹ This long-term brain dysfunction has traditionally been labelled ‘postoperative cognitive dysfunction (POCD)’, though current recommendations suggest the term ‘persistent neurocognitive disorder’ as more appropriate and better aligned with Diagnostic and Statistical Manual of Mental Disorders (DSM criteria).^{12 13} For this research protocol and to be consistent with previous dementia and POCD research, we will refer to these changes as cognitive decline, defined as new impairments in cognitive function months to years after surgery. Overall, the rapidly growing number of surgical cases in older patients and the high likelihood of cognitive decline make it imperative to identify interventions to reduce this public health crisis.

Major risk factors for postoperative cognitive decline include age, education, pre-illness cognitive deficits and in-hospital delirium.⁷¹⁴,¹⁸ Previous studies have demonstrated that cognitive impairment after major surgery was not associated with the surgery and anaesthesia exposure but was predicted by in-hospital or post-operative delirium.⁵¹⁹,²¹ Improvement in preoperative cognitive status and reduction of postoperative delirium, therefore, are likely key targets to reduce cognitive decline after surgery. Furthermore, postoperative brain dysfunction (ie, delirium in the hospital and cognitive decline) is linked with functional disability and lack of independence,²²,²⁴ and postoperative cognitive decline has been associated with decreased employment and survival.910 14 22 25,²⁸ Thus, cognitive decline impacts quality of life, functional disability, depression, and caregiver and societal burden^{1 2} and impairs critical activities such as managing medication,^{29 30} handling finances^{31 32} and living independently.³³,³⁵

Surgery, in its own right, is also associated with restricted activity levels and functional decline, including in activities of basic and instrumental activities of daily living, that often persists for months after surgery.^{2224 36},³⁸ Depression, which is also common after surgery,³⁹ has been associated with poor emotional and physical recovery and impaired quality of life.⁴⁰ As such, functional disability and depression are major obstacles to recovery and survivorship from major surgery.

Physical exercise can improve cardiorespiratory fitness, cognition and neuropsychological performance and reduce cerebral atrophy and dementia risk.⁴¹,⁵⁰ It can also attenuate neurodegeneration, promote neuroplasticity and functional connectivity, improve endothelial function and protect against oxidative stress and disruption of the blood-brain barrier.^{4142 44},⁴⁹ Prehabilitation is the process of enhancing an individual’s capacity and reserve before an acute stressor. Prehabilitation exercise studies to date have primarily focused on physical training to improve physical outcomes in certain types of surgical cases and not on older patients specifically.⁵¹,⁵⁷ Personalised physical prehabilitation can reduce postoperative complications,⁵⁶ and data suggest that a progressive preoperative exercise programme combined with directed postoperative therapy may provide the best means for patients to return to normative function.⁵⁸ Additional studies have shown that preoperative cognitive training may lead to reductions in delirium and improved postoperative cognitive performance,⁵⁹,⁶³ and this is an area of ongoing research.^{64 65} Combining physical exercise with cognitive training may hold the most potential to prevent postoperative cognitive and functional decline. Similar multicomponent approaches have demonstrated improved cognitive capacity among older at-risk community persons, but have not been studied in surgical patients.⁶⁶,⁷⁴ Utilising a remote training intervention would increase participation and generalisability to enhance practicality, scalability and future implementation.

We hypothesise that participation in a remote combined physical and cognitive exercise programme prior to (prehabilitation) and after (rehabilitation) major surgery will improve long-term cognitive and disability outcomes after major surgery and will test this hypothesis in the COgnitive and Physical Exercise to improve Outcomes after Surgery (COPE-iOS) clinical trial.

Methods and analysis

Trial design, setting and registration

Trial design

The COPE-iOS study is a randomised, controlled, patient and assessor blinded trial to evaluate a remote combined cognitive and physical exercise programme before and after a major surgery in older adults compared with an active attention control. Reporting of the COPE-iOS protocol herein adheres to the Standard Protocol Items for Randomised Trials (SPIRIT)⁷⁵ statement as delineated in the online supplemental material SPIRIT Checklist.

Patient and public involvement

Patients or the public were not involved in trial design, conduct or dissemination plans of our research.

Setting

Quaternary academic hospital in the United States.

Registration

The trial is registered on ClinicalTrials.gov (NCT04889417).

Population, eligibility, screening and consent

Study population and eligibility

The target population includes 250 older adults (≥60 years) undergoing elective major non-cardiac surgery with expected hospitalisation for ≥3 days. This includes individuals planned for surgery with the following surgical services: general surgery, gynaecology, hepatobiliary, otolaryngology, surgical oncology, thoracic, urologic and vascular surgery. Eligible patients must be able to participate in the combined cognitive and physical exercise programme and assessments and have adequate lead time prior to surgery (ie, consented ≥2 weeks before scheduled surgery). Patients with mild cognitive impairment who can make their own healthcare decisions are eligible for study inclusion. Full inclusion and exclusion criteria are delineated in box 1.

Box 1. Inclusion and exclusion criteria.

Inclusion criteria

Adults ≥60 years of age

Undergoing elective major non-cardiac surgery with expected hospitalisation ≥3 days

Exclusion criteria

Blind, deaf or inability to understand English as these conditions would preclude the ability to perform the proposed programme and prevent assessment with the study instruments

Prisoners

Severe frailty or physical impairment that prohibits participation in the programme

Cognitively unable to consent for surgery (i.e., dementia or cognitive impairment of a severity that precludes ability to self-consent and, thus, also participation in study interventions)

Inability to obtain informed consent ≥2 weeks before scheduled surgery

Surgical team unwilling to allow physical activity or other components of the intervention

Inability or unwillingness to utilise a tablet device, laptop, or email

Co-enrolled in another interventional trial examining similar outcomes or current enrolment in a study that does not allow co-enrolment

Screening and consent

We recruit patients prior to surgery from the Vanderbilt University Medical Center High-Risk Surgical Encounter (Hi-RiSE) Preoperative Clinic, which is a collaboration of anaesthesiologists, surgeons and specialists designed to preoperatively evaluate and optimise elective surgical patients with high vulnerability for poor outcomes. Research staff screen the Hi-RiSE clinic appointments for eligible participants daily by identifying patients planned for surgeries that meet the minimal hospitalisation requirement. Research staff approach patients directly to obtain informed consent. We do not use surrogate consent given the eligibility requirements of this study. The consent includes active participation in the current study, including optional blood sample collection and MRI, as outlined in the protocol below, as well as inclusion of use of data and specimens for future ancillary studies. A complete copy of the consent is included in the online supplemental materials.

Timeline, sample size, and recruitment

Study timeline and flow

The study timeline and participant flow are summarised in figure 1. For all activities and measurements collected after surgery, the indicated time frame begins at the time of discharge from the hospital stay associated with the major surgery. To simplify language, we will refer to this time frame as ‘after surgery’ throughout the manuscript.

Figure 1. Figure 1 displays the timeline of the study. Patients who meet the outlined characteristics are approached, consented and enrolled at their preoperative clinic visit. Study personnel perform cognitive, physical function and mental health testing to assess baseline status. Baseline biomarkers are collected, and eligible participants undergo Anatomical and Functional Brain MRI. Participants then begin their cognitive and exercise programmes assigned at randomisation. Participation continues until the time of surgery and resumes when participants return home after surgery from the hospital. While in the hospital, study personnel collect information from the health record, perform daily delirium, pain and mobility assessments, and collect biomarker samples. Participants continue in the cognitive and exercise programmes until 3 months after surgery. Research personnel contact participants and collect information on recovery and complete cognitive, physical function and mental health testing at 3 and 12 months after surgery. At the 3-month assessment, participants who are able to complete a final blood draw for biomarkers and repeat anatomical and functional MRI. ADL, activities of daily living; CAM, confusion assessment method; CNS, central nervous system; CPOT, Critical Care Pain Observation Tool; COPE-iOs, Cognitive and to improve after surgery; IADL, instrumental activities of daily living; ICU, intensive care unit.

Sample size

The trial is powered to evaluate whether the COPE-iOS training programme, compared with active control, will result in less decline in global cognition scores 3 months after surgery, as measured by the Central Nervous System (CNS) vital signs neurocognitive battery.^{76 77} We plan to randomise 250 patients (125 in each group) prior to surgery that subsequently undergo surgery. Thus, enrolment and randomisation will overall be greater than 250 patients to account for surgery cancellation. After accounting for 5% mortality and 20% loss to follow-up, we expect to have at least 188 participants available for follow-up assessments. This will provide us with 90% power to detect a 7.1 point difference in the primary outcome of neurocognitive index (NCI) score (mean 100±15) using the SD of the adult population^{76 77} with a two-sided alpha of 0.05. This detectable difference has validity as a clinically meaningful difference in cognition.⁷⁸ Assuming a normal distribution for secondary outcomes of KATZ activities of daily living (ADL) (mean 1.1±0.2),⁷⁹ FAQ (mean 0.06±0.24),⁸⁰ and DASS Depression (mean 8.92±8.34),⁸¹ this sample size provides us at least 80% power to detect a change of 0.08 units on the Katz ADL, 0.09 units in the FAQ scale and 3.5 units in the DASS scale between the intervention and control groups at a 2-sided alpha of 0.05. We calculated sample size using a two-sample t-test, assuming normal distribution for the outcomes.

Recruitment

Recruitment started October 12, 2021 with completion of recruitment in October 2025. Queries to investigators, data cleaning and closure of the database will follow. Data analysis, manuscript preparation and submission for publication are anticipated to occur in 2026.

Assignment of interventions, blinding and masking

Randomisation

Consecutive participants are assigned randomly prior to surgery to one of the two groups: COPE-iOS programme or active control. Randomisation occurs using a computer-generated randomisation scheme (with permuted block sizes) in a 1:1 ratio. The randomisation schema was created by the study biostatistician and directly uploaded into the database randomisation module. The database is housed in Research Electronic Data Capture (REDCap), a secure web application for building and managing online databases.^{82 83} At the time of randomisation, research coordinators meet in person with participants to provide a tablet device, wearable fitness tracker and binder with detailed instructions regarding the assigned programme and educate participants on the features of the device, including how to connect to wireless internet, participate in video conferencing and use the intervention or control cognitive game applications. If participants are unable to learn how to use the device after randomisation, they will be withdrawn. Participants are provided with the wearable fitness tracker for heart rate monitoring.

Blinding and masking

Only research coordinators, therapists and research staff overseeing the administration of the intervention programme or active control know the treatment assignment. Participants are blinded to their group assignment, as are the clinicians, including the surgical team, caring for them. In addition, the neuropsychology professionals performing the assessments for the primary and secondary endpoints, research staff gathering in-hospital data, laboratory personnel and the neuroimaging team conducting measurements remain blinded to each patient’s treatment assignment throughout enrolment, follow-up and data analysis. The elements of the COPE-iOS programme potentially make blinding of participants difficult. We have, therefore, designed the control group to be an active attention control that includes a computerised game and supervised video exercise teleconferences. This will help ensure blinding of participants and that any differences observed between groups will be due to the specific computerised cognitive training (CCT) and progressive exercise components of the intervention.

Treatment approach, intervention and control

Treatment approach

This study will not affect the surgery or any of the medical care received before, during or after the surgery.

Intervention: COPE-iOS programme

The COPE-iOS programme is a training programme that combines tablet-based CCT and remotely delivered supervised progressive physical exercise. Participants are expected to perform three remote exercise sessions per week before surgery and on returning home after surgery, up until 3 months post-discharge.

The cognitive portion of the intervention includes a CCT programme that is a tablet-based video game designed by Akili Interactive Labs called EndeavorRx, a Plasticity-Based Adaptive Cognitive Remediation (PACR) programme derived from the fields of brain plasticity and learning theory to strengthen functional connectivity and improve cognitive performance. EndeavorRx is designed to deliver sensory and motor stimuli that selectively activate specific cognitive neural systems in the brain to improve information processing, attention control, memory and executive functioning. EndeavorRx is the first FDA-approved treatment delivered through a video game experience. The programme provides an immersive action video game-like experience where the participant is encouraged to earn points and in-game rewards to advance, while simultaneously containing core therapeutic exercises. Adaptive closed-loop algorithms constantly adjust the difficulty and personalise the treatment experience for each individual, thus automatically adjusting the “dose” to each participant’s needs. The programme was specifically developed to target disorders with attention-related cognitive dysfunction which is affected by acute illness, ageing, dementia, depression and many other disorders. Participants are asked to perform CCT sessions daily.

The supervised physical exercise programme is led by licensed physical and occupational therapists and is conducted via video conferencing to allow real-time video and voice communication between the instructors and participants and permit qualitative and quantitative supervision in real time. After the initial in-person education session, participants complete an individual video conference session with one of the training therapists to acclimate them to the virtual experience. Participants then join the group physical exercises classes using video conferencing, where the therapists guide participants through exercise in real time. During the group exercise sessions, an additional monitor assists the instructor therapist in providing remote supervision through direct observation and feedback on the exercise (eg, technique, heart rate monitoring and exertion) to ensure proper performance and safety and to verify compliance (percent of sessions participated and percent of exercises performed per session). The instructor-therapist leads participants through a multimodal physical exercise programme designed to improve aerobic capacity, muscular strength and endurance. The programme includes seated aerobic, core strengthening and upper-body and lower-body strengthening exercises. Each session follows a consistent format outlined in table 1.

Table 1. Intervention and active attention control physical exercise.

Exercise session format
Intervention	Active Attention Control
Warm up and flexibility: 5 min	Warm up: 5 min
Upper body strength: 10 min	Flexibility and mobility: 10 min
Core strength: 10 min	Cool down: 5 min
Lower body strength and aerobic: 10 min
Combined upper and lower body strength and aerobic: 5 min
Cool-down and flexibility: 5 min

At the beginning, half-way point and at end of the sessions, the monitor asks participants to report their heart rate from the wearable fitness tracker and the patient-reported rating of perceived exertion (RPE), a 10-point scale that provides a subjective assessment of how hard the participant feels they are working during exercise. To achieve moderate-intensity exercise, we target a heart rate between 55% and 70% of the participant’s age-adjusted maximum heart rate and an RPE level of 5 to 6. The instructor aids participants in adjusting exertion during activities to achieve these targets.

Control: Active control

The active attention control includes a control computerised game and remotely delivered stretching exercises. Participants are expected to perform three exercise sessions per week before surgery and on returning home after surgery, up until 3 months post-surgery. Akili interactive selected the control computer game based on the following criteria: (1) to provide a face-valid approach to cognitive intervention, ensuring that participants remain blind to group affiliation; (2) to match the interventional treatment programme in overall programme use intensity, time-spent, reward delivery and engagement values; (3) will not adapt to participant performance by reducing/enhancing the speed and/or intensity of the game (as opposed to the adaptive CCT intervention). Participants are asked to complete control computerised game sessions daily. The supervised group exercise portion is conducted via video conferencing as in the intervention group. The exercises, however, include stretching and mobility with no emphasis on improving strength or cardiovascular endurance. The instructor therapist leads participants through a series of stretching exercises as outlined in table 1. During the group exercise sessions, an additional monitor assists the instructor therapist in providing remote supervision through direct observation and feedback on the exercise. At the beginning, half-way point and end of the sessions, the monitor asks participants to report their heart rate from the wearable fitness tracker and the patient-reported Rating of Perceived Exertion (RPE), a 10-point scale that provides a subjective assessment of how hard the participant feels they are working during exercise.

Primary, secondary and exploratory outcomes

Primary outcome

The trial’s primary outcome will be global cognition 3 months after surgery as determined by the neurocognition index (NCI) score from the CNS Vital Signs neurocognitive battery.^{76 77}

Secondary outcomes

Secondary outcomes include the following: (1) global cognition score (NCI score) measured by the CNS Vital Signs neurocognitive battery 12 months after surgery; (2) disability in basic ADLs measured by the Katz ADL Scale at 3 and 12 months after surgery; (3) disability in instrumental activities of daily living (IADLs) measured by the Pfeffer Functional Activities Questionnaire at 3 and 12 months after surgery; (4) depression measured by the Depression Anxiety and Stress Scale (DASS) at 3 and 12 months after surgery. Mechanistic secondary outcomes include the following: (1) plasma markers of endothelial and blood-brain barrier (BBB) injury; (2) MRI evaluation of tissue volume, white matter disease, functional connectivity, cerebral blood flow and BBB function.

Exploratory outcomes

We plan to assess and report several additional exploratory outcomes. These include delirium in the hospital measured using the Confusion Assessment Method for the intensive care unit (CAM-ICU),⁸⁴ mobility in the hospital measured using the Johns Hopkins Mobility Scale,⁸⁵ pain scores in the hospital measured using the Critical-Care Pain Observation Tool (CPOT)⁸⁶ and post discharge using the Behavioural Pain Index (BPI),^{87 88} length of hospital stay, discharge location (eg, home, rehabilitation centre and skilled nursing facility), healthcare utilisation after surgery and mortality at 12 months after surgery.

Observations and measures, data collection and data management

Observations and measures

At enrolment, we record patient demographics, home medications and pre-existing co-morbid conditions through patient interview and electronic medical record review. Trained neuropsychology professionals assess cognition, functional status, mental health and pain at enrolment and at 3 and 12 months after surgery. We collect blood samples to measure plasma biomarkers of endothelial and BBB injury at enrolment, the day of surgery, postoperative day 2, postoperative day 5 or discharge (whichever occurs first) and 3 months after surgery. Participants enrolled prior to 30 March 2025, were asked to complete an anatomic and functional research MRI at baseline and at 3 months after surgery. We collect data during the hospital admission from the electronic medical record, including surgery and anaesthetic details and medication administration, and in-person, including daily delirium and pain assessments with the CAM-ICU⁸⁴ and the CPOT⁸⁶ for up to 14 days after surgery. In addition, we track post-surgery hospital location (ie, ICU vs ward), need for mechanical ventilation, length of hospital stay and discharge location (eg, home, rehabilitation centre and skilled nursing facility). Further, we record standard of care mobility levels and participation in physical therapy in the hospital. A listing of key observations and measures and their associated time points during the trial is delineated in table 2.

Table 2. Summary of key observations and measures.

Variable	Enrolment	Intervention	In-hospital	Intervention	3 Months	12 Months
Demographics, comorbidities	X
Global cognition: CNS vital signs battery	X				X*	X
Disability, functional status: Katz ADL, Pfeffer FAQ and DASI	X				X	X
Depression, health status: FAQ, DASS and BPI	X				X	X
Blood draw: Plasma biomarkers of endothelial and blood-brain barrier injury	X		POD 0, 2, 5		X
Neuroimaging: Anatomical/functional MRI	X				X
Clinical assessment data:
Standard of care imaging, laboratory	X		Daily		X	X
Surgical course, hospital data, ICU requirement			Daily
Delirium (CAM-ICU) and pain scores (CPOT)			Daily
Discharge location			Daily
Readmission					X	X
Intervention performance and compliance:
CCT or control computer game		Daily		Daily
Supervised online physical training		Three times weekly		Three times weekly
Heart rate monitoring		Three times weekly		Three times weekly
Additional physical activity		Daily		Daily
Safety assessments		Three times weekly		Three times weekly

^*Global cognition measured at 3 months is the primary outcome. All other outcomes included in the table will be secondary or exploratory outcomes.

CAM-ICU, Cognitive Assessment Method for the Intensive Care Unit; CCT, Computerized Cognitive Training; CPOT, Critical Care Pain Observation Tool; DASI, Duke Activity Status Index; ICU, Intensive Care Unit; Katz ADL, Katz Activities of Daily Living; MRI, Magnetic Resonance Imaging; Pffeffer FAQ, Pfeffer Functional Activities Questionnaire.

Cognition

We assess global cognition with the CNS vital signs neurocognitive battery.^{76 77} CNS vital signs have been used to measure neurocognitive changes or deficits in most neuropsychiatric conditions and multiple other medical conditions, such as surgery and chemotherapy. It is sensitive to subtle deficits, as well as to progressive decline or improvement. The battery is reliable, well-validated, allows for remote assessment, is suitable for repeat testing, is translated into over 50 languages and has a large normative database. Its computer-based presentation, automatic scoring and norming and rating scales increase efficiency and accuracy of testing, and its standardisation provides comparisons across a wide range of cohorts.

The CNS vital signs neurocognitive assessment battery consists of seven subtests: verbal memory, visual memory, finger tapping, symbol digit coding, Stroop test, shifting attention test and continuous performance test. This battery objectively evaluates eleven basic brain functions: composite memory, verbal memory, visual memory, executive function, processing speed, psychomotor speed, reaction time, complex attention, cognitive flexibility, simple visual attention and motor speed. Standard scores are normalised from raw scores to provide an age-matched score relative to other people in a normative sample. CNS vital signs standardised scores have a mean of 100 and a SD of 15. The CNS vital signs neurocognition index (NCI) reflects the overall, or global, neurocognitive function of the patient. It is an average of the standard scores for five domains (composite memory, psychomotor speed, reaction time, complex attention and cognitive flexibility).

Functional status

ADLs and IADLs will be assessed with the Katz ADL⁸⁹ and the Pfeffer FAQ,⁹⁰ respectively. A score >0 on the ADL and >1 on the FAQ indicates partial disability in at least one domain of daily living. Functional status will also be assessed by the Duke Activity Status Index (DASI), a self-reported questionnaire that measures ability to perform activities requiring varying levels of exercise capacity and provides an overall estimate of functional capacity.⁹¹

Mental health

Depression, anxiety and stress will be assessed with the DASS.⁹² The DASS was developed in response to concerns about the potential overlap between depression and anxiety and to independently measure a third factor, stress, which is common to both of those constructs. As such, it is a widely used screening tool to assess symptoms of depression, anxiety and stress. This instrument comprises three sub-scales: (1) the Depression sub-scale, which measures hopelessness, low self-esteem and low positive affect; (2) the Anxiety scale, which measures autonomic arousal, musculoskeletal symptoms, situational anxiety and subjective experience of anxious arousal; and (3) the Stress scale which measures tension, agitation and negative affect. Higher scores on each subscale indicate increasing severity of depression, anxiety or stress. Internal consistency for each of the subscales is typically high, and the scales are stable over time.^{93 94} It is valid in both community and clinical settings (eg, chronic pain, psychiatric patients and post myocardial infarction),⁹³⁹⁵,⁹⁹ including after major surgery and critical illness.¹⁰⁰,¹⁰⁴ It performs consistently for assessment of older patients, and data suggest that the DASS may be used with older adults in lieu of multiple scales designed to measure similar constructs.⁸¹

Delirium and pain in the hospital

Trained research personnel will assess patients in person with the CAM-ICU⁸⁴ for delirium and Critical Care CPOT⁸⁶ for pain twice daily if the patient is in the ICU and once daily on the floor for up to 14 days or until hospital discharge or death. Research personnel enter data directly into REDCap after each assessment.

Pain intensity and interference

Pain intensity levels and interference with daily living will be assessed using the BPI questionnaire.^{87 88} This questionnaire has been used to evaluate both pain from chronic diseases such as cancer, osteoarthritis and low back pain, and pain from acute conditions such as postoperative pain. To assess pain intensity, participants rate their pain on a scale of 0–10 at four occurrences: worst, least and average pain in the last 24 hours, and pain at the current moment. The overall impact of pain on daily living is assessed using the BPI interference score. This score assesses on a scale of 0–10 how much pain interferes with daily activities, including general activity, mood, normal work, relations with other people, sleep, enjoyment of life and walking ability. The overall BPI interference score is calculated as the mean of all answered questions among the seven questions on the interference portion of the assessment. Pain intensity and pain interference scores can be further categorised into mild (1-4), moderate (5-6) and severe (7-10).

Blood draw: Plasma biomarkers of endothelial and blood-brain barrier injury

We collected a blood sample of approximately 15 mL at baseline, postoperative days 0, 2 and 5 (or day of discharge, whichever comes first) and 3 months after surgery. The samples are centrifuged, processed and stored at −80°C for batched analyses. The biomarkers were chosen based on their importance to mechanisms of cognitive decline and their associations with delirium and cognitive impairment. We will measure E-selectin, which is specifically expressed by the endothelium and utilised as a marker for endothelial injury.³⁶ Plasma E-selectin has been associated with coronary artery disease, diabetic microangiopathy, cerebrovascular vasospasm, severity of sepsis, organ failure, death from cardiovascular events, delirium and cognitive impairment after critical illness.¹⁰⁵,¹¹⁰ S100B is expressed and secreted by astrocytes after CNS injury or ischaemia and cell death.⁴⁰ Levels of S100B have been validated as a measure of BBB injury against cerebrospinal fluid-serum albumin quotients and MRI after BBB injury.¹¹¹,¹¹³ Levels of S100B have also been associated with the development of delirium and cognitive changes after surgery, delirium during critical illness and cognitive impairment after critical illness.^{109110 114},¹¹⁸ We will additionally measure IL-1β, IL-6 and IL-10 to evaluate the pro-inflammatory and anti-inflammatory response to surgery and illness, as well as potential mediation of the associations with endothelial and BBB injury.

Magnetic resonance imaging

Research MRI scans are performed by certified MRI technologists, following standard safety screening procedures (eg, metal in or on the body, pacemaker and claustrophobia). MRI measurements incorporate standard anatomical imaging for tissue volume and white matter disease classification in the following protocol:¹¹⁹,¹²² 3D T₁-weighted MPRAGE, T₂-weighted turbo-spin echo (chronic white matter disease), T₂-weighted FLAIR (chronic white matter disease), and DWI (acute white matter disease). Additionally, we performed non-invasive magnetic resonance angiography of the major intracranial vessels for vasculopathy assessment. Haemodynamic scans including blood oxygenation level-dependent (BOLD) MRI, T2-relaxation under spin tagging and arterial spin labelling (ASL) are acquired to assess functional connectivity, oxygen extraction fraction and perfusion determination, respectively. We also acquired a pulsed arterial spin labelling scan with multiple delay times spanning the arterial and venous circulation times, which will allow for transit time and capillary permeability surface product to be estimated.¹²³,¹²⁵

Intervention performance, compliance and participant safety

We asked participants to perform CCT or control game play daily. Performance is measured by Akili Interactive and provided to our team in dashboards and reports that are uploaded into the REDCap database. We record both the number of days of CCT or control programme use, as well as the duration of play each day, and will determine overall compliance at the conclusion of the study.

Each video exercise session (intervention or control) is administered by a therapy team member and supervised by a monitor. To determine exercise participation and compliance, monitors record participant attendance and completion of each session in REDCap. In addition, if participants choose to complete additional exercise activities, they are recorded in a diary that is collected at study completion and documented in REDCap. To monitor compliance and safety, participants use wearable activity monitors to actively monitor heart rate during exercise session participation, both for targeting exertion levels during intervention and for verifying that the maximum heart rate is not exceeded. The therapist and session monitor observe participants throughout the exercise sessions for adverse events including, but not limited to, emotional distress, physical injury (eg, discomfort, falls, myocardial ischaemia, syncope and dyspnoea) and fatigue.

Data collection and management

During all study phases, data are entered into electronic case report forms (eCRFs) in a password-protected REDCap database. Some portions of the neuropsychological battery are collected on paper CRFs as they require copyrighted forms. The data obtained from paper CRFs is subsequently entered into eCRFs for storage in the database.

Education and training, study withdrawal, adherence and monitoring

Education and training

Our lead physical therapist trains other therapists and monitors in the supervised exercise programme and control exercise programme. Monitors complete a checklist of the components delivered during each session and make note of any protocol deviations. The research team engages in weekly conference calls to reinforce knowledge and skills, discuss logistical, clinical and scientific issues that arise, and provide support and accountability for the research staff administering the programmes.

Withdrawal

Study participants may withdraw from the study at any time, for any reason, or for no stated reason. Participants will be withdrawn from the study if their scheduled major surgery is cancelled. We anticipate that a potential reason for study withdrawal will be the time required to use the programme. For participants seeking to withdraw for that reason, we will offer the alternative of discontinuing use of the programme and scheduling and completing the assessments only. Given this high-risk surgical patient population, some patients have their surgeries cancelled after enrolment and randomisation due to disease progression, change in treatment course or decision to no longer pursue surgical intervention. As such, those who do not complete the requisite surgery are not eligible for postoperative cognitive decline, disability or depression assessments and will not be included in the analysis cohort. These patients are withdrawn once the decision to no longer pursue surgery is made. Participants may also be withdrawn from study participation if the attending physician requests that the subject be withdrawn. In rare cases, the primary investigator may decide to prematurely discontinue a participant’s involvement in the study for safety, a change in circumstance that would prevent completion of protocol-required assessments, participant relocation, if the participant displays inappropriate behaviour toward study staff members, loss of funding, or clinical determination that continuation in the study is not in the participant’s best interests.

Adherence and fidelity

To maintain the integrity of the exercise sessions, our protocol includes the following: (1) staff training and assessment of competence in delivering the interventions and in the importance of fidelity; (2) use of detailed treatment manuals and (3) ongoing supervision to ensure accurate and consistent treatment delivery (provided via clinical team meetings). All video conference exercise sessions are recorded and a random selection of 10% of the recordings was selected for review during the initial months of the study. The recordings are compared with the programme checklist by an unblinded investigator who does not lead or monitor exercise sessions. After checklists reached 90% accuracy, a random selection of 5% of the subsequent recordings is reviewed (evenly balanced across sessions) and will continue until the intervention phase ends. If deviation from the programme checklist is detected, the study therapist(s) and/or monitor(s) will be re-trained and 100% of recordings will be reviewed until the deviation is addressed and fidelity returns to>90%.

Monitoring

A data safety monitoring board (DSMB) of four independent members who are not study investigators and have no financial, scientific, or other conflict of interest with the trial meets annually to monitor overall study activity. The DSMB members are experts/representatives from anaesthesiology, surgery, exercise science, clinical trial methodology and biostatistics. Prior to the initiation of the trial, the DSMB met and reviewed the entire IRB-approved study protocol, case report forms and manual of standard operating procedures to evaluate subject safety, recruitment, randomisation, intervention, data management, quality control and analysis. Also, they reviewed the informed consent document to determine applicability and readability. The DSMB is responsible for identifying problems related to safety, requesting additional data relevant to safety, proposing analyses of safety endpoints as needed and considering the rationale for continuation of the study in light of safety data, progress of enrolment, retention, protocol adherence and data management. Masked (ie, treatment group assignment will not be revealed unless asked) reports are provided to the DSMB annually. Interim analyses for efficacy or futility will not be performed given the early-stage nature of the trial, long-term outcome primary and secondary aims, mechanistic aims and exploratory outcomes of interest. The trial intervention being evaluated is minimal risk; thus, interim statistical analyses for safety will only occur if deemed necessary by the DSMB.

The study’s principal investigator (PI) oversees all trial operations. The Vanderbilt Coordinating Centre (VCC) within the Critical Illness, Brain Dysfunction and Survivorship (CIBS) Center works closely with PI to manage all study personnel and direct all study operations, including assurance of timely and appropriate enrolment and quality control, protocol compliance and optimal daily study conduct. A Steering Committee assists the PI in the design, conduct, analysis and interpretation of this trial. The Steering Committee includes experts in clinical trials, cognitive training and neuropsychiatric follow-up, physical and psychological training, neuroimaging, acute surgical illness and biostatistics and data management.

Throughout the study, we monitor for emotional distress, physical injury (eg, discomfort, falls, myocardial ischaemia, syncope and respiratory insufficiency), and fatigue. For emotional discomfort related to learning about the existence or severity of cognitive deficit, we will anticipate and respond to this possible scenario in two ways. First (in accordance with American Psychological Association Ethical Guidelines), we provide a feedback session to all study participants who request this, during which time we both provide participants with information about their test results and discuss with them any issues or concerns raised by the testing. Second, we have a list of referral sources available for individuals who request or are in need of further counselling or support (including those who were not negatively affected by neuropsychological assessment per se but who report ongoing emotional difficulties such as post-traumatic stress disorder symptoms related to the events of their intensive care hospitalisation).

Protocol amendments

The protocol details described herein are based on COPE-iOS study protocol V. 1.7, dated 30 March 2025. The timing of this protocol publication was informed by prior protocol amendments and a desire for the published protocol to most closely reflect and align with the eventual trial results published.

COPE-iOS study protocol V. 1.0 (dated 7 December 2020) was the first approved by the IRB. COPE-iOS study protocol V. 1.1 (dated 23 March 2021) clarified expected hospitalisation, timing of the follow-up assessments, changed to DASS for mental health assessments, timing of study training sessions, and added fidelity monitoring expectations. COPE-iOS study protocol V. 1.2 (dated 10 September 2021) removed contrast for the MRI. This is the version under which patient enrolment began.

COPE-iOS study protocol V. 1.3 (dated 13 September 2022) added remote consent and clarified randomised versus enrolled for statistical analyses. COPE-iOS study protocol V. 1.4 (dated 12 July 2023) clarified timing of programme around surgery and heart rate monitoring on the activity monitor. COPE-iOS study protocol V. 1.5 (dated 9 April 2024) changed enrolment target to 250 randomised who have surgery (approximately 275 overall), clarified future DNA work, and added one blood draw on the day of surgery. COPE-iOS study protocol V. 1.6 (dated 15 January 2025) changed enrolment target to approximately 315 overall to account for surgery cancellations. COPE-iOS study protocol V. 1.7 (dated 30 March 2025) added that patients enrolled after 30 March 2025 will not participate in neuroimaging. This change is due to funding and no cost extension status.

Data analysis

Analyses

To determine the effect of the COPE-iOS training programme compared with active control on cognition 3 and 12 months after surgery (Aim 1), disabilities and depression 3 and 12 months after surgery (Aim 2), endothelial and BBB injury 3 months after surgery (Aim 3), and cerebral atrophy and white matter connectivity changes 3 month after surgery (Aim 3), we will use the Student’s t-test or Mann-Whitney U test depending on the distribution of the outcome. Although randomisation should, in theory, ensure balance between the treatment and the control group, we will conduct analyses using multivariable regression to adjust for imbalances in baseline factors and strong prognostic factors to verify that any differences seen between the two groups are truly as a result of the treatment. Since adjusting for strong risk factors for outcomes reduces measurement error, this also leads to an increase in statistical power.¹²⁶ As such, the adjusted analysis will be the primary analysis. Baseline measures of the follow-up outcomes will be included in the corresponding outcome models when available. Including the baseline counterparts of the outcome in the model as a covariate (eg, baseline CNS Vital Signs scores when evaluating 3 month cognition as the outcome) accounts for variation in post-discharge outcomes due to baseline status.¹²⁷ Potential covariates to include in the model will be a priori specified in the statistical analysis plan and adjusted for in the models. The distribution of the outcomes will be evaluated before choosing the appropriate model. For normally distributed outcomes, we will fit multivariable linear regression models and evaluate model diagnostics. For outcomes that are skewed, we will fit a proportional odds model and use graphical techniques to evaluate the assumptions for a proportional odds model.¹²⁸ We will not adjust for multiple comparisons when examining secondary, a priori defined outcomes, in keeping with authoritative recommendations.¹²⁹,¹³² Restricted cubic splines with three knots will be incorporated into the multivariable regression models (when appropriate) to model non-linear relationships between the outcome and the continuous covariates. Mortality analysis will be conducted using actual time to death, with censoring at date of last follow-up. Kaplan-Meier curve with log-rank tests will be generated and multivariable analysis will only be conducted if there are sufficient events to adjust for covariates. To evaluate the robustness of our results, we will conduct sensitivity analyses, including complete case analysis. Descriptive and exploratory outcomes will be analysed using complete case analysis.

Ethics and dissemination

This trial protocol has been approved by Vanderbilt Human Research Protections Program (#202496) and an independent Data Safety Monitoring Board. The complete trial protocol and statistical analysis plan can be accessed via email to the investigators. We will submit trial results for publication to peer-reviewed journals and present results at one or more scientific conferences, with an expected timeframe for publication of 2027. A de-identified dataset will be shared for potential collaboration pending project review. We will consistently update ClinicalTrials.gov with any amendments to the protocol as per SPIRIT Guidelines.