This online, remote trial enrolls adults with an HbA1c of 7.0% or higher and assigns them to a 4-month intervention focused on teaching participants to adopt very low–carbohydrate breakfasts.
The intervention will be delivered remotely and is technology supported. Participants can live anywhere in the United States.
Key inclusion criteria include (1) a current HbA1c level of >7.0%, (2) the ability to read and speak English, (3) age 18-80 years, and (4) the willingness to regularly check blood glucose levels.
Exclusion criteria include (1) unable to provide informed consent; (2) pregnant or planning to become pregnant in the next 12 months, or breastfeeding or less than 6 months post partum; (3) vegan; (4) a C-peptide test result that suggests type 1 diabetes, a previous diagnosis of type 1 diabetes, or latent autoimmune diabetes of adults (if C-peptide is <0.75 ng/mL, participants are ineligible; if C-peptide is between 0.75 ng/mL and 1.5 ng/mL, these participants will need to get an additional test, Glutamic Acid Decarboxylase 65, to rule out latent autoimmune diabetes of adults; if the Glutamic Acid Decarboxylase 65 test result is ≥5 U/mL, participants will be ineligible); (5) cancer; (6) heart failure; (7) kidney failure; (8) liver failure; (9) an untreated mental health condition; (10) currently following a very low–carbohydrate diet or breakfasts; (11) abnormal thyroid levels, if baseline thyroid stimulating hormone is out of range, a free T3 or free T4 test may be ordered; (12) triglyceride >999 mg/dL; (13) a change in diabetes medication within the last 3 months; (14) alcohol use disorder; (15) previous bariatric surgery; (16) difficulty chewing or swallowing; (17) dependence on others for food preparation; and (18) currently enrolled in another investigative study that might conflict with this research.
Participants with a triglyceride level reported on the initial blood draw of 300-499 mg/dL will be enrolled with the requirement of a follow-up lipids test 2 weeks into the intervention. For triglyceride levels 500-999 mg/dL, we will have the participant first discuss the result with their primary care physician so that any medication or dietary changes can occur before starting the study. If any significant changes were made by the participant’s physician (eg, addition or removal of a medication), we would measure a new baseline at the appropriate time. If triglycerides are greater than 999 mg/dL, the participant will not be enrolled. Such tests may also be ordered after an initial meeting with the study physician, based on clinical judgment.
An authorized team member, who has undergone the necessary ethical training and received clearance from the institutional review board, will discuss the details of the study and consent forms, as well as respond to any questions. It will be emphasized that participation in the study is completely optional.
We propose to enroll 84 adults, after attrition, with type 2 diabetes and persistent hyperglycemia, in our 4-month online intervention. Assessments will take place at baseline and 4 months later, at the end of the intervention. Participants will complete an online course for a very low–carbohydrate breakfast diet program. This online program allows us to conduct the intervention remotely. Participants will receive weekly classes online, a food guide and recipe booklet through the mail, and a diet coach to support them. For details, see the SPIRIT (Standard Protocol Items: Recommendations for Interventional Trials) flowchart in
Timeline of planned contact for enrollment, study intervention, and assessments for pre-post program for adults with type 2 diabetes. CGM: continuous glucose monitor.
Study materials will encourage eating a very low–carbohydrate breakfast each day. The very low–carbohydrate diet intervention will focus on food choices, without calorie counting or calorie restriction. We will recommend that breakfast meals should have no more than 10 nonfiber grams of carbohydrates or no more than 20% of total calories from carbohydrates. To avoid confusion about how to interpret this, we will provide participants with copious food options that will vary in their complexity and cost. For examples, see
Classic eggs, bacon
Sausage breakfast casserole
Low-carbohydrate pancakes and waffles
Cottage cheese with fruit and nuts
Almond and coconut crunch cereal
Vanilla pecan bars
Cheesy breakfast pizza
Low-carbohydrate English muffin in a mug
In the event of any serious adverse events resulting from the intervention, participants will immediately stop the diet and will be included in the intention-to-treat analysis. If there are any changes to the participants’ medical treatment plan that conflict with the intervention, participants will also stop the intervention. Furthermore, participants will be notified that they have the right to retract their consent to participate at any time and may refuse to answer any questions from the study team.
To improve adherence to the intervention, there will be responsive coaching (via phone, video, and email) so that participants’ questions and comments are replied to in a timely and supportive manner. Reminders about targeted behaviors are tied to greater adherence, so participants will be sent text messages to provide feedback, as well as to motivate, educate, and remind them about targeted behaviors and skills about 5 times per week. Monthly group video sessions will also be offered as an optional way for participants to receive support.
Relevant concomitant care is permitted during the trial.
We will not be providing posttrial care. We do not anticipate harm due to trial participation.
At month 4, we will ask participants, “How would you rate your overall satisfaction with the program?” on a 0-6 Likert scale (0=very dissatisfied; 6=very satisfied).
Additionally, each week, participants will receive a check-in survey. The survey will inquire (1) how often they ate a very low–carbohydrate breakfast on a scale from 1 to 7 and how much they enjoyed eating a very low–carbohydrate breakfast on a scale from 1 to 7 (1=not at all to 7=very much so), (2) what their highest and lowest blood sugar readings were over the past week and what were the circumstances of the readings (only asked of those taking insulin, sulfonylureas, or meglitinides), and (3) any medication changes they may have made over the past 7 days.
The Diabetes Treatment Satisfaction Questionnaire, an 8-item scale, measures 2 factors: treatment satisfaction (along dimensions such as flexibility and convenience) and burden from hyperglycemia and hypoglycemia. It is a short, easy-to-answer scale with low participant burden. We will use the Diabetes Treatment Satisfaction Questionnaire at baseline and 4 months later [
Health-related quality of life is a central aspect of well-being. We will use the Patient-Reported Outcomes Measurement Information System 29 at baseline and 4 months later [
At baseline and 4 months later, we will also measure physical symptoms using a short, face-valid measure of physical symptoms [
In the 4-month survey, all participants will be asked open-ended questions about the aspects of the program participants enjoyed, which aspects could be improved, what made it easier to adhere to the program, and which factors made adherence more difficult.
To better inform the next phase of our study, we will conduct qualitative semistructured interviews at the end of the trial. This will allow us to explore the experience of the participants in greater depth, including their perceptions of the intervention, as well as barriers to and facilitators of their ability to make long-lasting dietary changes. We will model the interview guide and data analysis on prior research, using a semistructured interview guide that includes open-ended questions to encourage participants’ perspectives, thoughts, and beliefs about the intervention and its components. Participants will be interviewed by nonintervention staff from our research team. These interviews will last approximately 20-30 minutes. We will create a summary of the qualitative data, with evidence from participant quotes, to explore participant experiences.
As recommended by the CONSORT (Consolidated Standards of Reporting Trials) guidelines, we will report on the percentage enrolled from each recruitment method, the number excluded and why, the number who chose not to enter the trial and why, and the number and reasons for withdrawal [
We will assess dietary adherence with 1-2 unannounced 24-hour dietary recalls at baseline and at 4 months, conducted by a dietician. At the 4-month outcome mark, this will serve as a check on our intervention fidelity in terms of whether we have achieved our targeted macronutrient levels for breakfasts. We will explore what percentage of calories in participants’ breakfasts were derived from carbohydrates, as well as what percentage of participants ate breakfasts that had 20% or less of their calories from carbohydrates. We will also examine the macronutrient content and quality of other eating occasions throughout the day to assess whether breakfast quality impacted other eating patterns.
As appropriate, for the following measures, we will explore how each of these outcomes changed from the baseline to the end of the study. All blood tests will be done at Labcorp lab locations, with the rare exception that if participants are unable to go to a Labcorp location, we will mail them a DTI Laboratories, Inc home HbA1c kit for the 4-month outcome.
HbA1c is a standard measure of overall glycemic control in the clinical care of type 2 diabetes. A higher HbA1c is associated with increased risk of microvascular complications from diabetes [
The study team will ask participants about their current diabetes-related medication regimen using questions that have successfully been used in other studies [
We will explore metrics of diabetes control and report on the percentage of participants who (1) lower their HbA1c levels to below 7.0% without increasing their medication effect score, (2) lower their HbA1c levels to below 6.5% without increasing their medication effect score, and (3) lower their HbA1c levels to below 6.5%, without increasing their medication effect score, and take no antihyperglycemic medications other than metformin.
Data will be analyzed from a continuous glucose monitor at baseline and 4 months. The research team will download sensor data at the end of the measurement period. We will assess the glucose variability and the proportion of time spent in the euglycemic (3-7.8 mmol/L) and hyperglycemic (≥ 11.1 mmol/L) states, following previous standards for interstitial glucose concentrations [
We will give participants a body weight scale and ask them to weigh themselves weekly in order to help tailor coaching support. However, our outcome of interest will be weight change from baseline to 4 months. We will examine average weight change and the percentage of participants who lose 5% of their body weight, a clinically meaningful amount of weight loss.
At baseline and at 4 months, we will measure triglycerides and fractionated cholesterol using Labcorp’s Nuclear Magnetic Resonance LipoProfile [
At baseline and 4 months, fasting insulin and glucose will be used to estimate insulin resistance by calculating Homeostatic Model Assessment-Insulin Resistance. Homeostatic Model Assessment-Insulin Resistance uses a single fasting blood draw to estimate insulin resistance [
At baseline and 4 months, high-sensitivity C-reactive protein, which is an acute phase reactant, will be assessed via blood draw.
We will mail participants a blood pressure cuff with instructions from the American Heart Association to correctly measure blood pressure. To control for the white-coat effect, we will use a home-based measurement [
At baseline and 4 months later, we will measure this with the 8-item Weight Efficacy Lifestyle Questionnaire Short-Form, which assesses self-efficacy for dietary adherence when facing a variety of challenging situations, including poor mood and social situations [
At baseline and 4 months later, we will assess whether the intervention improves these 2 factors using two 4-item subscales of the Adult Eating Behavior Questionnaire, one for each factor [
Potential participants will make initial contact with the study by visiting the study website or reaching out to the study team through a dedicated study phone number or email address. Participants will be directed to a study website where they will take an initial screening survey. If participants appear initially eligible, the participant is invited to watch an online video describing the study and answer questions about their goals for participation and the pros and cons of participation (which previous research suggests can improve trial retention) [
If potential participants remain eligible after the blood draw, our study staff will consent them on the phone or via a video meeting and review the study schedules, procedures, and assessments. We will send a fax to the primary care provider explaining the study and that the participant is eligible unless the physician objects. If we are not contacted by the physician in 2 weeks, we assume the physician is supportive. Participants will then (1) complete a baseline survey online, (2) participate in 1-2 unannounced 24-hour dietary recalls over the phone, and (3) wear a continuous glucose monitor mailed to them, or attached in person if they live in the greater Ann Arbor area, and they will be taught how to place the continuous glucose monitor via a video call with a trained staff member. If the participant already has a continuous glucose monitor as part of their normal care, we will ask to be connected to their data. After wearing the continuous glucose monitor for 14 days, participants will mail the device back to the program staff, (4) report their blood pressure measured on the monitor provided by the study 8 times over 2 weeks, (5) measure their body weight on the scale provided twice within 5 minutes to provide an accurate average, and (6) meet online with a study physician to discuss medications and glucose testing, if they are using insulin or a secretagogue to manage their diabetes. Once all baseline steps have been completed, the program will begin asynchronously for each participant. For details, see the SPIRIT flowchart in
To estimate the number of subjects needed for the outcome of HbA1c, we note that the Diabetes Prevention Program showed a 0.1% decrease in HbA1c with an SD of the difference approximately equal to 0.1% [
We have denoted the progression criteria that would inform a decision to proceed to a full-scale trial. We would proceed if the 95% CIs for our measures of acceptability, feasibility, as well as for our health outcome of HbA1c, all meet the following benchmarks, respectively: a mean rating of at least 3.0 on the 0-6 satisfaction scale, at least 40% attendance of assigned sessions, and a decrease of at least 0.1% SD.
To reach our target sample size, we will advertise on social media to reach participants across the United States. We will hang fliers in community centers in various locations in the United States. We will also search the medical records of Michigan Medicine for potentially eligible participants. Each year of the study, we will initially contact the pool of potentially eligible participants by sending them a letter and flyer using a Health Insurance Portability and Accountability Act (HIPAA)–compliant company describing the study, which will refer them to a study webpage and an initial online screening questionnaire. All recruitment materials will be written at no higher than a fifth-grade reading level. We may follow up with a phone call.
Assessment data will be reviewed for accuracy and completion by study staff. Loss to follow-up and incomplete data will be recorded.
Participants will be paid US $100 for completing the 4-month assessments and will be able to keep any of the intervention-related items we send them, including a body weight scale, a blood pressure cuff, and a glucometer. We also allow them to choose a few kitchen items, such as a skillet, a baking sheet, or ingredients, to send to them at the beginning of the study.
Trial data will be collected through online questionnaires via Qualtrics. Labcorp will send laboratory data via electronic files to be added to the study database. There will be a regular review of cases during the enrollment process to ensure proper eligibility. All study data will be stored in REDCap (Research Electronic Data Capture; Vanderbilt University), a cloud-based HIPAA-compliant database. REDCap allows for specified ranges and automatic calculations to reduce errors. Data will be cleaned by the research team upon completion of data collection.
Participants will have blood samples drawn at baseline and 4 months later. All blood samples will be drawn, analyzed, and then destroyed by Labcorp, excluding the rare exception where participants will be mailed a DTI Laboratories kit for their 4-month HbA1c measurement.
A biostatistician will perform all outcome statistical analyses. An intention-to-treat analysis will be performed on all participants. A per-protocol analysis will also be used on participants who had high compliance with the very low–carbohydrate breakfast intervention. We defined high compliance as breakfasts that have 20% or less of their calories from net carbohydrates.
Descriptive statistics will be computed for all outcome measures. Continuous variables will be reported using means, medians, IQRs, and SDs based on the distribution. Categorical variables will be presented using frequencies and percentages.
For this aim, we will explore how measures of acceptability and feasibility change from baseline to the end of the study using a 2-tailed paired
For the primary outcome measure of acceptability, we have denoted a mean rating of 4.0 on the 0-6 satisfaction scale as our
Thus, we have at least 85% power for both primary hypotheses of Aim 1.
For continuous health and psychological measures, we will explore the change from baseline to the end of the study for each outcome, using a 2-tailed paired
There are no planned interim analyses or stopping rules for this trial. All intervention-related serious adverse events will be reviewed by the Data and Safety Monitoring Board to determine if the study should be stopped.
We will analyze whether baseline characteristics modify the benefits of the low-carbohydrate breakfast diet on HbA1c. We will explore whether there seem to be differences in the magnitude of HbA1c changes from baseline to 4 months later across 3 predefined subgroups: levels of obesity, levels of insulin resistance, and women versus men. Linear mixed models similar to those described above will be used, with the addition of candidate moderators (each in separate models), and interactions between the moderator, intervention arm, and time. These models will be used to estimate the change in the outcome within subgroups and differences in change between subgroups. The focus of these exploratory analyses will be the magnitude and direction of change within and between subgroups. We will also report the statistical test on the interaction term, which represents an overall test of the moderation effect.
To assess the impact of missing data on our outcomes, baseline characteristics of participants will be compared with and without missing data. Primary outcome analyses will use random-intercept-random-slope mixed effects models to approximate changes from the start to the end of the study, 4 months later. Mixed effects models using maximum likelihood estimation allow an appropriate assessment of repeated measures despite the effects of missing data.
After the trial’s prespecified outcomes have been published, a deidentified dataset will be available to other investigators by request. For access to the dataset, the request must align with institutional review board protocols, and the trial steering committee must agree to the request.
The primary decision-making body of this study is the investigative team comprising the principal investigator (PI; LRS) and the coinvestigators. The PI is accountable for the overall management of the study. She directs the operations of the study, reviews questions or problems that arise from the study between team meetings, and presents issues to the research team for decision. The PI acts as the intermediary with the funding organization, handling tasks such as submitting annual reports, overseeing fiscal and administrative operations, and managing the coordination and implementation of the study.
The trial steering committee will consist of the PI (LRS), trial coordinator (KR), and study physician (SK). Weekly meetings with the study team will be conducted to review study implementation and discuss adverse events. Additionally, monthly meetings will be held with other study investigators and staff to address any overarching study issues and assess trial progress. Ongoing communication through group email will facilitate discussions related to enrollment inquiries and address issues like recommending medication changes for participants.
The day-to-day operations of the study, encompassing recruitment, data collection, and intervention processes, are overseen by project coordinators and research assistants. Their responsibilities extend to coordinating institutional review board revisions, managing data, monitoring reports, and documenting the completion of necessary trainings. Staff members are tasked with participant recruitment and screening, securing informed consent, and scheduling and conducting follow-up assessments. The lead project manager supervises the development of the study’s data tracking system and surveys.
A data safety monitoring board, composed of 3 individuals, has been established. Biannual board meetings will be conducted through video conference to assess recruitment, retention, and safety reports, as well as preliminary results.
In the event of a serious adverse event with a likely connection to study participation, a special meeting with the data and safety monitoring board will be convened to assess the necessity for any required modifications or potential early termination of the trial. An adverse event is broadly defined as any occurrence that causes or elevates the risk of harm to the participant or others. Serious adverse events encompass instances leading to death, inpatient hospitalization, or extension of existing hospital stay, persistent or significant disability or incapacity, or the emergence of a congenital anomaly or birth defect. The study team meticulously evaluates all potential adverse events reported by participants, considering their relation to the study intervention, expectedness, and severity.
Following a reduced-carbohydrate diet may result in reduced glucose levels. For individuals taking medications like insulin or a secretagogue, this could elevate the risk of hypoglycemia if appropriate medication adjustments are not implemented. Medical adjustments, particularly reductions (or discontinuations at low doses), will be supervised by study physicians. The order of prioritizing reductions or discontinuations typically follows the sequence outlined: (1) insulin, (2) secretagogues, (3) meglitinides, (4) sodium-glucose transport-2 inhibitors, (5) glucagon-like peptide -1 agonists, glucagon-like peptide 1/glucose-dependent insulinotropic polypeptide dual agonists, and dipeptidyl peptidase-4 inhibitors, (5) alpha glucosidase inhibitors, and (6) thiazolidinediones.
Upon enrollment in the study, we will provide participants’ primary care physicians with study information.
The study team will consult the study endocrinologist for potential medication changes if the participants report a low blood glucose level below 90 mg/dL on their weekly survey or if they report their blood glucose as dropping below 110 mg/dL on 2 separate occasions. If the endocrinologist advises medication changes, this will be communicated to the participant. We will fax any changes to the primary care physician.
If participants decide not to accept the medication changes, the study team will also fax the primary care physician our recommendation and the participants’ rejection of the suggestion.
The trial will be closely monitored by study investigators and the data and safety monitoring board, convening twice annually. The study team will submit annual progress reports to both the institutional review board and the National Institute of Diabetes and Digestive Kidney Diseases, the study sponsor.
The trial results will be disseminated through conference presentations, uploaded onto ClinicalTrials.gov (NCT05986097, initially registered on August 2, 2023), and published in peer-reviewed journals. All final peer-reviewed manuscripts will be submitted to the digital archive PubMed Central. Additionally, relevant data will be deposited in suitable public repositories where applicable.
This trial uses a single institutional review board, with the primary application approved by the University of Michigan Institutional Review Board (HUM00225646; Office for Human Research Protections Institutional Review Board registration number IRB00000244). When communicating important protocol modifications, the study team will request approval from the University of Michigan Institutional Review Board via eResearch and communicate approved changes to the relevant parties via email, phone, or online meetings. We plan to confirm that all groups understand and agree with the proposed changes. All trial participants provide informed, written consent using forms approved by the institutional review board. We will deidentify and code all surveys and forms with a unique participant number. To encourage participant retention, participants will be paid US $100 for completing the 4-month assessments.