Sleep Technology Intervention to Target Cardiometabolic Health (STITCH): a randomized controlled study of a behavioral sleep extension intervention compared to an education control to improve sleep duration, blood pressure, and cardiometabolic health among adults with elevated blood pressure/hypertension

According to data from the National Health Interview Survey,70.1 million US adults (29.2%) sleep < 6 h per 24-h period [1], below the consensus recommendation of “at least 7 h” needed for health and performance among adults [2]. Therefore, a large portion of US adults could potentially improve health and quality of life through extending sleep duration. Meta-analyses demonstrate compelling evidence that short sleep duration independently increases cardiometabolic diseases (CMD) including a 20% increase in hypertension, 40% increase in diabetes22, and 0.23% higher HbA1c in individuals who have type 2 diabetes [3], risk ratios comparable to other lifestyle risk factors such as overweight and physical inactivity. Experiment sleep restriction increases in BP among healthy and hypertensive participants [4, 5], as well as increasing inflammation [6], insulin resistance [7, 8], and caloric intake [9]. Taken together, the studies demonstrate biological and behavioral effects of short sleep duration on CMD development and management.

Despite the well-known negative effects of short sleep duration, surprisingly little attention has been paid to the development of sleep extension interventions and testing whether such treatments can improve CMD. Studies have demonstrated short-term sleep extension results in decreased insulin resistance [10, 11] decreased appetite [12] and lower intake of fat, carbohydrates, and sugars [13]. Together, this growing body of research demonstrates short-term sleep extension is feasible and may confer health benefits for CMD risk factors. However, these studies are limited by small sample size, high staff involvement in the intervention (e.g., intensive coaching and monitoring), and short-term follow-up. Therefore, our project moved this important research forward, taking it from proof-of-concept to efficacy studies to research conducted in more generalizable clinical populations [14].

The goal of this proposed study is to conduct a randomized controlled trial to test the efficacy of our behavioral sleep extension intervention on 24-h ambulatory BP (ABP) and other key biological and behavioral CMD risk factors among participants with short sleep duration and elevated BP/hypertension. The intervention will be completed over a 12-month period in 3 phases, intervention (weeks 0–8), maintenance (months 2–6), and follow-up (months 6–12), allowing us to test the short-term and extended effects of our intervention.

Compared to the education control group, we predict greater improvements in sleep duration in the intervention group in sleep duration (aim 1), 24-h ABP, and other measures of BP and CMD risk (aim 2) and health behaviors and patient-reported outcomes (aim 3). We predict improvements will persist at the 12-month follow-up, but the effects of the intervention will be attenuated over time.

This is a single-blind, parallel-group, randomized controlled trial with a 1:1 allocation ratio and superiority framework. The study will test a behavioral sleep extension intervention compared to an education control group. The primary endpoint is 8 weeks and the secondary endpoint is 12 months. A CONSORT-style flow chart for the trial is shown in Fig. 1.

This study will recruit participants with elevated BP/stage I hypertension and sleep duration < 7 h from community settings in Salt Lake City, Utah, and the surrounding area.

The schedule of time points, study assessments, and interventions is listed in Fig. 2.

After web or telephone screening, participants will complete a baseline/screening visit. Participants will undergo standardized BP screening to determine BP eligibility. The BP screening protocol involves 3 unattended BP recordings taken with an automated BP machine (HEM-907HL, OMRON Healthcare Inc.) after a 5-m quiet resting period and 3 m between measurements. If they meet the study criteria, they will review the consent form with the study staff and complete the informed consent prior to continuing the study visit. Participants will complete self-reported questionnaires and cognitive testing and then will wear the study devices at home over the next 10 days (Apnea link monitor for 1 night, ambulatory blood pressure monitor (OnTrak Ambulatory Blood Pressure Monitor 90,227, Spacelabs Inc.) for 24 h, and actigraphy and physical activity monitor for 10 nights). At the end of the baseline period, participants will return to the lab to download their devices and will have their blood drawn. If eligible for the study, they will be randomized to either the sleep extension intervention group or the education control group. They will receive a study letter and materials in the mail, as well as a call from the study coach informing them of their group assignment. The study interventions have 3 main periods. In the intervention period (study weeks 1–8), participants will receive 8-weekly intervention or control sessions and content. In the maintenance period (months 2–6), participants will receive monthly intervention or control sessions and materials. Finally, in months 6–12, participants will be observed, but there will be no interventions delivered.

The following are the inclusion criteria:

The following are the exclusion criteria:

Participants will attend a screening visit, and prior to consent, the research staff will take their resting blood pressure using a standardized protocol. Participants who are eligible based on their blood pressure readings will complete informed consent. Consent will be completed by the trained study staff (the research coordinator or research assistant) during an in-person visit to the University of Utah. The informed consent document will be reviewed with participants after a standardized blood pressure reading to determine preliminary eligibility.

This study includes blood draws for cardiometabolic risk markers (e.g., lipids, glycated hemoglobin (HbA1c)) as well as Alzheimer’s-related biomarkers (targeted metabolomics, lipidomics) and the genetic risk gene apolipoprotein E4 testing (APOE4). In the consent process, participants will have the opportunity to opt out of blood draws, opt out of genetic testing, and the option to opt out of receiving results of genetic testing.

Study outcomes and assessment time points are listed in Fig. 2.

Primary outcome: The primary outcome is a change in sleep duration from baseline to 8 weeks. Sleep duration at baseline and 8 weeks will each be calculated as averages over 10-day periods. Sleep duration will be measured by 8–10 days of actigraphy at baseline/screening, week 8, and 6 and 12 months with the Actiwatch Spectrum device (Philips Respironics Inc.). Devices will be configured using default settings and scored using the Acitware Software (Philips Respironics, Inc.). Actigraphy will be manually scored using a standardized procedure [16].

Main secondary outcome: The main secondary outcome will be change in 24-h SBP from baseline to 8 weeks. This measure will be calculated as an average for the 24-h period at baseline and 8 weeks. BP will be assessed by 24-h ABPM at home using the Spacelabs ambulatory BP monitor (Spacelabs Healthcare, Hertford, UK). Participants will wear the BP monitor at baseline/screening, week 8, and 6 and 12 months. Monitors will be programmed to take readings in random 20-min intervals during the day and 30-min intervals at night.

Other secondary outcomes: Changes from baseline to 6 months and from baseline to 12 months will be measured to evaluate the maintenance of sleep duration changes. In addition, changes in 24-h SBP from baseline to 6 and 12 months will be measured to evaluate the maintenance. We will also measure the change in 24-h diastolic blood pressure (DBP) from baseline to 8 weeks and 6 and 12 months. We will measure diet quality by administering the Automated Self-Administered 24-h Dietary Assessment Tool and calculating the Healthy Eating Index (HEI-2018) [17], a composite measure derived from the 24-h recall measurement that uses food groups and nutrients collected in the diet recall to compute the US dietary guidelines into a single measure. Changes in the HEI reflect the overall change in the healthfulness of diet. We will also compute macronutrients and key food group/diet behaviors (e.g., fast food, fruit/vegetable intake) from the ASA-24 data for use in exploratory analysis.

Objective 24-h physical activity data including sedentary behavior will be collected using the ActivPAL4 activity monitor [18, 19]. The main measure of interest will be minutes of moderate and vigorous physical activity per day (MVPA).

Other measures of cardiometabolic risk will be measured at screening, week 8, and months 6 and 12 including BMI and body fat estimation using a bioimpedance scale (Tanita of America), waist and hip circumference, lipids, and HbA1c. We will also evaluate the effects on nocturnal and diurnal BP and on non-dipping BP status defined by ABPM levels as mean nighttime to daytime SBP/DBP ratio of > 0.90.

Participants will complete measures of patient-reported outcomes at session baseline, week 8, and months 3, 6, 9, and 12, including the PROMIS sleep disturbance, sleep-related impairment and sleepiness [20], and mood [21].

We will assess the intervention acceptability and user engagement to inform our findings and plan for future studies (Fitbit usage, coaching session attendance and duration, and email lesson engagement and use surveys and open-ended questions at the end of the intervention and follow-up periods).

A description of the participant timeline is listed in Fig. 4.

Our power calculations are based on the assumption that the standard deviation (SD) for change in sleep duration from baseline to 8 weeks will not exceed 37.4 min, which is 10% higher than the SD for change in sleep duration of 34 min reported after 40 days in our previous study [22]. We also assume the Pearson correlation between baseline and follow-up sleep duration will not exceed 0.75, and that the attrition rate will not exceed 20%. Under these assumptions, 120 randomized participants (60 in each of the treatment and control groups) will provide at least 80% power with 2-sided α = 0.05 to detect a treatment effect of 20.2 min.

If we also assume that (1) the SD for baseline SBP will be \(\le\) 11.8 mmHg, which is 10% higher than the average baseline SD reported in two previous studies [refs]; (2) the Pearson correlation between the baseline and 8-week follow-up SBP values will be \(\ge\) 0.70; and (3) the attrition will be \(\le\) 20%, and the sample size of 60 patients per treatment group will additionally ensure at least 80% power for the main secondary analysis to detect a treatment effect on the change in SBP of 4.86 mmHg.

The study will use both active and passive recruitment methods. For active recruitment, patients from the University of Utah Health (UU Health) primary care practices who meet the initial criteria (age, BP) will be sent letters explaining the study and providing information for online screening. The study staff will follow-up letters with a phone call and email to potential participants. We will work with local community leaders and community health workers to enhance the recruitment of diverse study participants, such as providing flyers and information about the study to share with their communities. We will attend local health fairs and conduct BP screening and share study information for interested participants. For passive methods, we will post flyers on the University of Utah campus and nearby areas (e.g., grocery stores and coffee houses) and will post ads on social media (e.g., Facebook). If recruitment slowdowns happen, the team could send more letters, expand to a larger number of clinics, post more frequently on social media, consider in-person recruitment in primary care visits, and attend more health fair events.

The primary outcome of the proposed study is the change in the mean sleep duration from baseline to 8 weeks. The study sample size has been determined based on the primary analysis of covariance (ANCOVA) that will compare mean sleep duration at 8 weeks, with statistical adjustment for the baseline sleep duration and the sex and age randomization strata. The proposed ANCOVA approach is recommended for the evaluation of treatment effects on continuous outcomes in randomized trials because it optimally accounts for random imbalances in the outcome at baseline and regression to the mean, thus providing superior statistical power compared to the analyses of change scores or follow-up scores without adjustment for baseline levels [36].

The main secondary outcome for this study is the change in SBP from baseline to 8 weeks. Evaluation of the 24-h SBP outcome at 8 weeks will assure that few patients will have modified their antihypertensive medication regimen over the 8-week interval prior to the primary outcome assessment.

Evaluation of the 24-h SBP outcome at 8 weeks will assure that few patients will have modified their antihypertensive medication regimen over the 8-week interval prior to the primary outcome assessment. Similar ANCOVA models as used with sleep duration will also be used to evaluate the acute effects on this and other numeric secondary and tertiary outcomes.

We will use linear mixed models [37, 38] to analyze changes in sleep duration, BP measures, sleep, and other numeric tertiary outcomes over all follow-up visits, including the 8-week and 24-week assessments. These analyses will evaluate the persistence of treatment effects over the full 24-week follow-up period and will allow us to utilize all data collected (i.e., baseline, 8 weeks, and 24 weeks) while allowing for within-subject variation.

There are no interim analyses planned for this study.

In addition to our primary analyses, we plan to analyze the results of participants who completed the intervention per protocol, in that they attended all the coaching or health education sessions. We will also conduct exploratory analyses based on clinically significant improvements in sleep and analyze the results among participants in the intervention group that extended their sleep duration by at least 30 min and for those who met criteria at 8 weeks of having sleep duration over 7 h. Finally, we plan to test the impact of obstructive sleep apnea as a moderator of sleep and cardiometabolic outcomes. We will conduct moderator analyses of participants to test the results of the intervention among participants with mild obstructive sleep apnea (5–15 apneas per hour at screening) compared to those with < 5 events per hour.

We plan to apply multiple imputation to impute missing outcome data, using comprehensive imputation models that include both non-missing measurements from the variables included in the data analysis as well as additional auxiliary variables that are considered to be predictive either of the risk of missingness or the values of the variables whose missing values are being imputed. The multiple imputation approach has been established as superior to alternatives, including complete case analysis or simplified ad hoc single imputation approaches [39, 40].

De-identified data and a codebook will be provided upon written request to the contact PI (Dr. Baron).