The primary aims of this pilot study were 1) to determine the proportion of approached, eligible patients who consented, 2) to determine the proportion of screened patients who consented, 3) to measure the number of days between surgery scheduling and admission date, 4) to determine the proportion of study participants retained, 5) to determine the proportion of study participants with adequate preoperative actigraphy data, 6) to determine the proportion of study participants who completed the sleep questionnaire battery, 7) to determine the proportion of study participants who completed the cognitive testing battery, and 8) to determine the proportion of postoperative days with daily delirium testing. The secondary aims of the study were 1) to investigate the prevalence of preoperative altered sleep patterns, 2) to determine the prevalence of the melatonin receptor SNP, and 3) to investigate the incidence of postoperative delirium and subsyndromal delirium.
Study design: We conducted a single-center, prospective, pilot, cohort study of older Hispanic/Latino patients undergoing cardiac surgery with and without preoperative sleep disturbances and postoperative delirium from February 2020 to December 2021.
Changes in methods after study commencement:
Our study was conducted during the height of the COVID19 pandemic in Miami, FL. During the various waves of the pandemic, when local incidence of COVID19 increased, recruitment would fall. To continue to assess the feasibility of the protocol, we amended our protocol to decrease the age for eligibility from 65 years of age to 50 years of age on June 17, 2021. In addition, we amended our protocol to allow a minimum of 5 days between recruitment and surgery instead of a 7-day requirement. Finally, after 6 participants the cognitive battery was reduced from 5 to 3 total tests due to multiple complaints of increased burden by the participants.
Eligibility requirements:
Sequential eligible people scheduled for cardiac surgery at the University of Miami, Miami Florida were recruited between February 2020 and November 2021. Inclusion Criteria were: 1), self-identify as Hispanic/Latino or preferred language as Spanish, 2) undergoing scheduled elective and urgent cardiothoracic surgery, and 3) 50 years of age and older. Exclusion Criteria were: 1) inability to consent, 2) preferred language other than English, Spanish or Portuguese, 3) emergency surgery, 4) anticipated discharge < 48 hours, 5) individuals < 50 years of age, 6) pregnant women, and 7) prisoners.
Setting: UHealth of the University of Miami is an academic tertiary center in Miami, FL with a large, 60%, Hispanic/Latino population.
Ethical approval, Registration and Funding:
The study was approved by the University of Miami Institutional Review Board (IRB) (eProst ID 20200942). The study was registered at ClinicalTrials.gov on March 8, 2021 (NCT04786899) This study received funding from the Miami Clinical and Translational Science Institute (Grant number: CTSI-Pilot-FY2021-05).
Recruitment and Informed Consent Process:
Each week, the study team reviewed a list of patients who were scheduled for cardiac surgery in the cardiothoracic clinic for age and estimated length of stay eligibility. The Internal Review Board (IRB) approved for the study team to search for each patient in the Electronic Health Record and review the patient demographics only. If a patient self-identified as Hispanic/Latino or their preferred language was Spanish, the study team then called the patient to assess for interest in participating in the study using IRB approved telephone scripts. Consent took place either virtually or in-person. The virtual consent process was approved by the IRB.
Study Procedures
Recruitment: We measured the number of eligible, approached, and consented patients in an on-going basis and at completion of the study. We gathered data on reasons for denial of participation, barriers to recruitment, and eligible: enrolled ratios to optimize recruitment in an ongoing basis. We measured the number of days between scheduling the surgery and the planned surgery date. The subjects were paid $50 after the preoperative study visit and an additional $50 at the completion of the study
Retention: We measured how many participants completed the study once consented. We gathered data on reasons for withdrawal. We gathered data from all participants data on burden level of each part of the study protocol.
Preop cognitive testing: Preoperative cognitive impairment is a major risk factor for the development of delirium and a potential confounder. Baseline cognitive was to be tested in all subjects. Baseline cognitive status was evaluated by a cognitive battery. This battery was given in-person if feasible. If not feasible due to COVID-19 and social distancing requirements, we gave the battery over video-call platform if the patient had a capable device and internet connection. If they did not have video-call capabilities, we planned to administer a modified battery over the phone. The original cognitive battery included Montreal Cognitive Assessment (MOCA) (Spanish version and telehealth or blind version as necessary),24,25 Hopkins Verbal Learning Test Revised (HVLTR),26 copy figure from Repeatable Battery for the Assessment of Neuropsychological Status (RBANS),27 Geriatric Depression Scale (GDS) (available in multiple languages including English, Spanish and Portuguese),28 and the FRAIL scale.29 The modified cognitive battery only included the MOCA, GDS, and the FRAIL scale. MK, a neuropsychologist at the University of Miami, developed the protocol for virtual cognitive test visits according to her clinical practice. She trained all clinical research coordinators and EMG in the cognitive test administration over a series of 3 training sessions.
Preoperative Sleep testing
Actigraphy: We obtained 5 to 7 days of wrist actigraphy data using an Actiwatch Spectra Pro (Philips Respironics, Murrysville, PA), a research grade actigraph. Philips Actiware 6.1.1.3 (Respironics Inc, Murrysville, PA) software was used. The Spectra Pro is lightweight and waterproof. The participants were to press the event marker button on the device during wake-sleep transitions. Wake-sleep transitions were also determined from the actigraphy data if participants did not push the event marker button. Raw acceleration signals were converted into activity counts over 30 sec time epochs. Actiware automatically computed wake thresholds, the minimum number of activity counts required to score an epoch as awake. The Actiware software also has a sleep interval detection algorithm, and sleep duration was calculated once a rest interval was set. The wake-sleep transitions and sleep intervals were manually reviewed and validated by the research team. Sleep duration, wake after sleep onset (WASO), and sleep efficiency were our key variables. WASO is the total number of minutes scored as awake within sleep intervals during a 24-hour day. Sleep efficiency is the total sleep time divided by the time in bed and multiplied by 100. We defined normal sleep efficiency as >85%, slightly poor sleep efficiency as 75-84%, moderately poor sleep efficiency as 65-74%, and severely poor sleep efficiency as <65%.30 We derived 24-hour rest-activity rhythm metrics estimated from activity counts over multiple days to determine the timing and stability of such rhythms. Each participant was also asked to complete a sleep diary.
Sleep questionnaires:
The Pittsburgh sleep quality index (PSQI) was given to each patient to assess for the presence of poor sleep quality in the 4-week time interval prior to surgery. It is a validated and commonly used self-rated questionnaire. A cut-off score of 5 has a sensitivity of 89.6% and specificity of 86.5% in distinguishing good and poor sleepers.30 It consists of 19-item scale that provides seven component scores (ranges 0-3): subjective sleep quality, sleep latency, sleep duration, sleep efficiency, sleep disturbances, use of sleeping medications, and daytime dysfunction. The PSQI is available in multiple languages including English, Spanish, and Portuguese and includes a version specifically for Hispanic/Latino patients in the USA.31
The Insomnia Severity Index (ISI) was administered.32 The ISI, a 7-item self-report questionnaire, assesses the nature, severity and impact of insomnia. Severity of sleep onset, sleep maintenance, and early morning awakening problems, sleep dissatisfaction, interference of sleep difficulties with daytime functioning, noticeability of sleep problems by others, and distress caused by the sleep difficulties are evaluated. A 5-point Likert scale is used to rate each item (0 = no problem; 4 = very severe problem), for score of 0 to 28. We planned a cutoff of > 10 for insomnia (sensitivity of 86.1% and specificity of 87.7% in community samples).32 The ISI is available in multiple languages including English, Spanish and Portuguese.
The Epworth Sleepiness Scale (ESS) was administered. The ESS is a validated brief self-administered questionnaire which is used both clinically and research settings to measure the subject’s usual level of daytime sleepiness or average sleep propensity.33 The ESS asks the patient to rate the likelihood of falling asleep during 8 common situations like “watching television,” “sitting and reading,” and “sitting and talking to someone.” For each situation, the patient rates on a 4-point ordinal scale the chance of dozing from “No chance of dozing” to “high chance of dosing.” Scores of > 6 are consistent with excessive daytime sleepiness. The ESS is available in multiple languages including English, Spanish, and Portuguese.
The STOP-BANG score was administered to assess the risk of sleep disordered breathing. The STOP-BANG score is one of the most widely used scores to assess for the risk for obstructive sleep apnea (OSA).34 The questionnaire has 8 questions. The risk of OSA is scored as low, intermediate, or high depending on which questions are answered yes and the total answered yes. The Stop-Bang is available in multiple languages including English, Spanish and Portuguese.
Genetic testing for rs10830963 polymorphism: A blood sample was drawn during the patient’s hospital stay. The UHealth Center for Genome Technology provided DNA extractions, quantitation, and storage of the DNA. A taqman allelic discrimination assay for the variant was performed. All genotypes were done in duplicate with appropriate controls to ensure proper clustering and genotype calling. The prevalence of both the risk allele (G) and risk genotype (GG) were determined.
Postoperative Delirium Testing
The Confusion Assessment Method (CAM) was given twice per day along with chart review until discharge to identify the presence of delirium. All clinical research coordinators and research assistants were trained by MK, a neuropsychologist, and EMG, expert in delirium, according to the CAM training manual.35 The CAM-ICU was given to the patient while in the intensive care unit (ICU) and intermediate care unit.36 The CAM-ICU is the validated version to be used in this setting. The CAM-ICU is available in multiple languages including English, Spanish, and Portuguese. The Long CAM was planned to be given to the patient when not in an intensive care setting. Delirium was defined as the patient having evidence of both acute onset of change in cognition or fluctuating course (CAM Feature 1) and inattention (CAM Feature 2) and either altered consciousness (CAM Feature 3) or disorganized thinking (CAM Feature 4). Subsyndromal delirium was defined as the patient having any of the four features but not meeting criteria for delirium. The duration of delirium in days was recorded. Team meetings occurred once per month to review delirium testing and documented observations to adjudicate delirium results.
Data collection from both patient interview and medical record review included demographic variables, medical comorbidities, home medications, surgery type, anesthesia type, administered medications, perioperative complications, and pain scores.
Feasibility to proceed with future definitive trial
Feasibility of the study protocol was the objective of this pilot study. Investigators had established a priori threshold for specific feasibility criteria. These were the following: (a) the proportion of approached patients participating in the pilot study would be 30% or greater, (b) the proportion of enrolled subjects who were retained until the end of the study would be 70% or greater, (c) the proportion of 5 days or more of actigraphy data would be 80% or greater, (d) the proportion of complete sleep questionnaire batter would be 80% or greater, (e) the proportion of complete cognitive batter would be 80% or greater, and (f) the proportion of subjects with daily delirium testing would be 80% or greater.
Sample size: Since this was a pilot study, a sample size calculation was not performed. The researchers aimed for 15 to 30 participants because it was felt this would be a large enough sample to inform them about the practicalities of recruitment, retention, study procedures in a perioperative study on sleep and delirium in older Hispanic/Latino patients undergoing cardiac surgery. While a sample size of a future definitive study at this time cannot be calculated, the goal would have a pilot sample of about 10% the size of the larger scale study.
Blinding: Because the primary goal of the pilot study was to assess feasibility of the protocol and because we had limited research staff, blinding of the results of the sleep questionnaire and cognitive batteries during delirium testing was not always possible. However, the clinical research staff were blinded to the results of the actigraphy and melatonin receptor SNP data during delirium testing.
Analytical Methods:
For the primary feasibility aims of the study descriptive statistics were performed for eligible patients who consented, approached patients who consented, number of days between scheduling of the surgery and surgery date, the retention of study participants, the adequacy preoperative actigraphy data, the completion of the sleep questionnaire battery, the completion of the cognitive battery, and the completion postoperative delirium testing. For the secondary patient specific aims, descriptive statistics were performed for sleep duration, wake after sleep onset, sleep efficiency, poor sleep quality, insomnia, excessive daytime sleepiness, melatonin receptor SNP risk allele and genotype, and postoperative delirium. We also included age, education level, surgery type, and length of postoperative stay. Descriptive statistics were conducted for all these study variables.