Intermittent Versus Continuous Low-Energy Diet in Patients With Type 2 Diabetes: Protocol for a Pilot Randomized Controlled Trial

Background: Intensive face-to-face weight loss programs using continuous low-energy diets (CLEDs) providing approximately 800 kcal per day (3347 kJ per day) can produce significant weight loss and remission from type 2 diabetes (T2D). Intermittent low-energy diets (ILEDs) and remotely delivered programs could be viable alternatives that may support patient choice and adherence. Objective: This paper describes the protocol of a pilot randomized controlled trial to test the feasibility and potential efficacy of remotely supported isocaloric ILED and CLED programs among patients with overweight and obesity and T2D. Methods: A total of 79 participants were recruited from primary care, two National Health Service hospital trusts, and a voluntary T2D research register in the United Kingdom. The participants were randomized to a remotely delivered ILED (n=39) or CLED (n=40). The active weight loss phase of CLED involved 8 weeks of Optifast 820 kcal/3430 kJ formula diet, followed by 4 weeks of food reintroduction. The active weight


Introduction
Background An estimated 4.7 million people have type 2 diabetes (T2D) in the United Kingdom, with the number expected to rise to over 5.5 million by 2030 [1]. Diabetes related complications are common, and people with T2D die up to 10 years earlier than those without the disease [2]. Currently, 10% of the National Health Service (NHS) budget in the United Kingdom is spent on diabetes (approximately £10 billion (US $13.8 billion) per year) [1].
Approximately 80% to 90% of people with T2D have overweight or obesity [3][4][5][6]. Clinical guidelines for the management of T2D focus largely on multiple drug treatments to reduce blood glucose. They also recommend at least 5% to 10% weight loss [7,8] as this leads to improvements in glycemic control, insulin sensitivity, blood lipids, and blood pressure (BP) [9,10].

Continuous and Intermittent Low-Energy Diets
Intensive face-to-face weight loss programs using continuous low-energy diets (CLEDs) that provide approximately 800 kcal (3347 kJ) per day of formula-based total diet replacement for 8 to 20 weeks or longer are highly effective for large weight loss and remission from T2D [11][12][13][14]. The recently published Diabetes Remission Clinical Trial (DiRECT) tested an intensive CLED program in primary care and found that it was superior to standard best practice care (standard daily moderate energy restriction advice with minimal support). At 12 months, 45.6% (68/149) participants of the intervention group achieved remission with an average weight loss of 10 kg compared with 4.0% (6/149) and 1 kg loss in the control group (P<.001). Remission was highest (31/36, 86%) in those who achieved 15% weight loss [13] and is more likely in those diagnosed with T2D in more recent years [15].
The CLED approach using total diet replacement is thought to be effective because the initial rapid weight loss can be highly motivating [16], and formula diets remove decision making around food choices. In addition, subjective hunger may be reduced by the associated ketosis [17]. CLEDs have been shown to reduce excess fat in the liver and pancreas in patients with T2D, which is part of the proposed mechanism for T2D remission [11,18].
Possible drawbacks of the CLED approach are that it is not appealing to or achievable for everyone. Participants following CLED programs report shame and awkwardness in social situations centered around food [19]. Attrition in intensive CLED studies on people with overweight or obesity (+/−T2D) is approximately 25% [13,20], although higher rates have been reported in studies with less frequent health care professional (HCP) contact [14]. The prevention of weight gain following a CLED program remains a key challenge. In 2 years, participants in the CLED group in the DiRECT trial appeared to have regained approximately 40% of the weight they had lost after the initial total diet replacement phase despite regular face-to-face support and an intensive 2-year relapse program involving repeated spells of a CLED, partial meal replacements, and antiobesity medications [13,21].
An intermittent low-energy diet (ILED) is a potential alternative to CLED. This includes the same number of low-energy formula diet days as CLED, but these days are undertaken for 2 days per week over a longer period. ILED may provide an alternative approach for people who find CLEDs unappealing or difficult to maintain. Qualitative reports on people following CLEDs suggest that they would prefer an intermittent approach that may be easier to fit into life without the need for weeks away from normal food [22].
A recent randomized controlled trial (RCT) in people with T2D compared an ILED with regular, daily, modest energy restriction (1200-1500 kcal/5020-6276 kJ per day) for 12 months and showed similar improvements to glycemic control in both groups [23]. Early studies of CLEDs show that some of the improvements in insulin sensitivity and beta cell function are associated with acute energy restriction rather than weight loss [24]. As these benefits subsequently attenuate when subjects enter weight maintenance with euenergetic feeding [25], ongoing spells of intermittent energy restriction each week may be a way to maintain beneficial glycemic control. A recent RCT (n=46) of participants with prediabetes and overweight or obesity showed that an ILED every other day for 12 months produced greater reductions (P<.05) in fasting insulin −52% (SE 9%) and insulin resistance -53% (SE 9%) compared with isocaloric daily moderate calorie restriction (−14%, SE 9%; −17%, SE 11%) despite similar decreases in body weight [26].
An ILED with 2 low-energy days per week will lead to a slower initial weight loss compared with a CLED. An unanswered research question is whether an ILED may lead to improved weight loss maintenance compared with daily dietary approaches in patients with T2D.
The relative benefits of ILED versus isocaloric CLED regarding glycemic control, diabetes remission, and weight loss maintenance in people with T2D are unknown.

Remote Follow-up
A potential strategy for increasing adherence efficacy and reach of low-energy diet (LED) programs may be to include high-frequency remote follow-up, which has been shown to be superior to low frequency face-to-face care in weight management interventions [27]. Remote care reduces participants' burden of attending face-to-face appointments and may be cost-effective compared with face-to-face care [28] while improving access to care. There is growing evidence to support the use of telehealth (including telephone and mobile phone-based apps) to monitor and provide feedback to patients with T2D and promote self-management of their condition [29][30][31].
The Manchester Intermittent versus Daily Diet Diabetes App Study (MIDDAS) incorporated high-frequency remote follow-up via the Oviva smartphone app and by telephone [32]. The app was used to facilitate self-monitoring of diet, weight, and blood glucose, and communication with HCPs, with the option of remote peer group support. Group participation in mobile apps has been shown to predict weight loss success [33]. Remote communities can be encouraging, motivating, and informative while remaining convenient and anonymous [34].

Goals of This Study
The primary aim of MIDDAS is to assess the feasibility and potential efficacy of remotely supported ILED and isocaloric CLED programs in patients with overweight or obesity and T2D. The feasibility of an RCT comparing the 2 approaches was also assessed. MIDDAS did not have a control group for comparison because CLED programs have already been shown to be superior to standard or best practice. [13] The initial estimates of acceptability (uptake and retention) and potential efficacy (change in weight and glycated hemoglobin [HbA 1c ]) of the programs will determine whether progression to a full RCT is indicated and will inform the feasibility of delivering ILED and CLED programs that incorporate remote follow-up.

Methods
The trial protocol (V5.0/08.04.19) was granted ethical approval by the North West Greater Manchester South Research Ethics Committee (ref:17/NW/0389). SPIRIT reporting guidelines were used [35].

Design and Setting
The study was a 12-month pilot 2-arm RCT, performed in patients with T2D and overweight or obesity, recruited from general practices, NHS hospital trusts and an NHS-supported voluntary research register in England. Participants attended trial assessments at Manchester University NHS Foundation Trust (MFT), Manchester, United Kingdom. Participants received dietary support remotely via the Oviva app and/or by telephone.

Recruitment
Potential eligible participants were recruited from 3 settings using the following methods: Those invited by letter had the opportunity to tell the trial team why they did not wish to take part via an anonymous reply slip. Interested patients were invited to an optional group session at MFT to receive more information on the diets and try the Optifast meal replacements before deciding to participate.
The inclusion and exclusion criteria for the trial is detailed in Textbox 1. • Severe anxiety or depression was assessed using the Generalized Anxiety Disorder Scale (GAD-7 [39], score ≥15) and Patient Health Questionnaire-9 (PHQ-9 [40], score ≥15). Hazardous or harmful drinking was indicated by the Alcohol Use Disorders Identification Test (AUDIT [41], score ≥16) • Active symptoms associated with emotionally unstable personality disorder, bipolar disorders, psychotic disorders, post-traumatic stress disorder, and current self-harm or suicidal behavior. Participants with these issues were potentially eligible, dependent on further information from their general practitioners and responses to the baseline study questionnaires • Current treatment with lithium, antipsychotics, or other psychotropic medications that may cause excessive weight gain • Chronic use of steroids • Medical conditions that in the opinion of the treating physician were at risk of deterioration (eg, severe systemic or organ disease, active cancer, liver, gall bladder disease, and pancreatitis) • Current participation in a diabetes drug trial

Participant Flow and Medication Management
The participant flow through the study is outlined in Figure 1. Informed consent was obtained by the trial research nurse at the baseline appointment. Eligible participants were randomized to ILED or CLED. All participants were invited to attend follow-up appointments at MFT for a repeat of clinical measurements at weeks 8, 12, 28, and 52. Changes to diabetes and antihypertensive medications specific to each treatment arm are detailed in the trial medication management plan in Table S2 of Multimedia Appendix 1. The CLED medication management protocol was devised by the research team and instructed participants to stop all diabetes medications, with the exception of metformin. Insulin was stopped or reduced, depending on the baseline HbA 1c level. Medication management of the ILED arm was adapted from a protocol tested in a recent ILED trial [42]. Medications in both arms were reintroduced or titrated during the trial according to clinical needs. General practitioners (GPs) were notified of the enrollment of patients and changes to their medications by letter.

Randomization and Blinding
Eligible participants were randomized 1:1 to ILED or CLED by a researcher independent of the intervention using a minimization program stratified by BMI ≥34 or <34 kg per m 2 (projected mean value from the counterpoint and counterbalance studies [11,12,15]), sex, duration of diabetes <4 years or ≥4 years, and whether participants were prescribed insulin. Due to the nature of the intervention, it was not possible to blind participants and clinicians to the treatment allocation. Clinical assessments were performed by an independent research assistant. Laboratory results were assessed by independent laboratory staff. Figure 1 shows a summary of the 2 dietary programs. Both included a combination of the Optifast LED and the Mediterranean diet over a period of 12 months. Both programs were designed to have exactly 56 Optifast LED days and an equivalent level of energy restriction during their active weight loss phase and weight maintenance or continued weight loss phase. The Mediterranean diet, as described previously by Harvie et al [43], was relatively high in protein (25% energy) with moderate carbohydrate (45% energy from low glycemic load carbohydrates), moderate fat (30% energy from fat: 15% monounsaturated, 8% polyunsaturated, and 7% saturated fatty acids), and limited alcohol to <10 units per week.

Weight Maintenance or Continued Weight Loss Phase (40 Weeks)
Participants who achieved a trial weight loss goal of 15%, and/or their target weight if greater, were advised to follow a euenergetic Mediterranean diet for weight maintenance. Those who had not achieved 15% weight loss or wished to lose more weight were asked to follow a 25% energy-restricted Mediterranean diet. Individuals' estimated energy requirements were calculated using the Mifflin equations [44] to estimate the basal metabolic rate×reported metabolic equivalent of the task. These calculations were based on the revised weight and activity levels after the completion of the active weight loss phase.

Active Weight Loss Phase (28 weeks)
The ILED group was asked to include 2 consecutive days per week of the Optifast LED plus 5 days of a Mediterranean diet for 28 weeks. An energy deficit was applied to the Mediterranean diet (up to a value of 265 kcal [1109 kJ] per day) to ensure that the ILED and CLED diets were isocaloric during the active weight loss phase.

Weight Maintenance or Continued Weight Loss Phase (24 weeks)
Participants were asked to follow a food-based ILED. Participants who had achieved 15% weight loss, and/or their target weight if greater, were asked to follow the food-based LED described above for one day per week and a euenergetic

Relapse
Trial participants who regained 2 kg or more during their weight maintenance or continued weight loss phase were offered one opportunity to resume their initial active weight loss diet to help reverse the weight gain. The CLED group was offered the 820 kcal (3430 kJ) formula Optifast LED for one week, and the ILED group was offered the Optifast LED 2 days per week for 3 to 4 weeks. The relapse program included increased support and monitoring by the trial dietitian and support from the trial psychologist to explore reasons for weight regain and to prevent further relapse.

Dietitian and Nurse
The diabetes specialist dietitian provided education on the diets and how to use the Oviva app, either by telephone or face-to-face. All participants were provided with written support materials, including recipes and meal plans. The diabetes specialist nurse provided support for patients who were on diabetes medications other than metformin, on antihypertensives, or who were hypertensive at baseline. All changes to diabetes and antihypertensive medications were agreed upon by the trial doctor and communicated to participants and their GP by a diabetes specialist nurse.

Physical Activity
Participants were encouraged to undertake physical activity (PA) throughout the trial to limit the loss of fat-free mass and promote weight loss maintenance. PA was supported by an exercise specialist, and suitability for PA was assessed at baseline using the Physical Activity Readiness Questionnaire (PAR-Q) [45] and signed off by the trial consultant endocrinologist. Participants were encouraged to aim for 5×30 minutes of moderate-intensity cardiovascular PA per week and resistance exercises for the legs, arms, and trunk 3 times per week. They were signposted to local PA services as appropriate and were educated on minimizing the risk of hypoglycemia. Diabetes medications were managed alongside exercise in response to the reported blood glucose readings.

Psychological Support
Enhanced psychological support from the trial psychologist was available for participants whose baseline scores indicated moderate scores for binge eating (score 18-26 on BES [37]), self-efficacy (score 36-45 on the Weight Efficacy Lifestyle Questionnaire Short Form (WEL-SF) [38]), anxiety (score 10-14 on the generalized anxiety disorder scale (GAD-7, [39]), depression (score 10-14 on the Patient Health Questionnaire-9 [PHQ-9]; [40]), or risk of alcohol-related problems (score 8-15 on the Alcohol Use Disorders Identification Test [AUDIT]; [41]). Psychological support was also available to participants who relapsed or had been identified by the team as experiencing difficulties impairing their ability to adhere to the programs. Psychological intervention was informed by motivational interviews, cognitive behavioral therapy, behavioral activation, mindfulness, emotional regulation, and distress tolerance skills.

Remote Behavioral Support
All participants received regular remote support from a multidisciplinary team including a diabetes specialist dietitian, nurse, and exercise specialist. The frequency and mode of behavioral support are detailed in Tables S1 and S2 in Multimedia Appendix 2. For example, the dietitian contacted the participants weekly via Oviva app messaging in weeks 1 to 12, fortnightly in weeks 13 to 28, and monthly in weeks 29 to 52. Follow-up telephone calls with the dietitian were performed at weeks 8, 12, 28, and 52.
Participants were invited to communicate with the multidisciplinary team via the Oviva app functional on iOS and Android smartphones and tablets. The app facilitates written messages and self-monitoring of diet, weight, blood glucose, weight, and activity levels. Participants were also invited to take part in group messaging on the app with other participants from their allocated diet group. Use of the app was optional, so if participants chose not to use it, then their scheduled contacts were done by telephone.
Participants signed a treatment contract to monitor their blood glucose and BP according to the protocol (Table S1 in Multimedia Appendix 1) and reported these values by telephone or via the Oviva app to the multidisciplinary team. The number of participants in both arms who requested face-to-face dietary consultations during the trial was recorded.
The multidisciplinary team was trained in motivational interview techniques (a well-established model of supporting behavioral change with proven efficacy in facilitating weight loss) to support dietary behavioral changes during both LED programs and in the longer term. Both programs used behavior change techniques such as problem solving and feedback on behavior, identified in a recent systematic review as being effective in reducing HbA 1c [46]. The programs also used established behavior change techniques, such as goal setting and self-monitoring [47].

Primary
1. Uptake: To achieve an uptake of at least 10% from a primary care mail out. 2. The proportion of subjects in the ILED and CLED groups who successfully lost and maintained >15% weight loss at 12 months, as determined by intention-to-treat (ITT) analysis. 3. The proportion of subjects in both groups who achieved HbA 1c <48 mmol per mol (6.5%) at 12 months using ITT analysis. 4. Retention: Aiming for a retention of 60% (48/79) completion as measured by attendance at the 12-month appointment. This is the acceptable completion rate in NICE guidance for commissioning weight management services in England [48].

Secondary
Process Measures 1. Participant adherence to the protocol including self-reported adherence to LED days, preference for face-to-face contact with the dietitian, preference for food-based LED days over Optifast, and attendance at follow-up appointments. 2. Download and usage of the Oviva app for self-monitoring.

Exploratory Measures
Change in the following measures across the 12-month study period: 1. Body fat and fat-free mass (bioelectrical impedance) 2. Waist and hip circumference 3. BP, lipid profile, and fasting blood glucose levels 4. Number and dosage of diabetes and BP medications 5. Self-efficacy for eating, anxiety, depression, and quality of life (as measured by WEL-SF [38], GAD-7 [39], PHQ-9 [40] and Audit of Diabetes-Dependent Quality of Life (ADDQoL) [49] questionnaires, respectively) 6. Quality of diet on non-LED days (as measured by the Mediterranean diet score questionnaire [50]) 7. PA (as measured by the Scottish Physical Activity Questionnaire [51]) 8. Self-reported satisfaction with weight loss was measured using a 7-point Likert scale. This is highly relevant for the comparison of ILED and CLED, where the CLED will experience faster initial weight loss and achievement of their weight loss goal than the ILED arm [52] 9. Serious adverse events reported up to the end of the 12-month trial Table 1 provides a summary of the measurements collected at baseline and weeks 8, 12, 28, and 52 by a research nurse.

Participant Characteristics
At baseline, we collected information on participants' age, sex, marital status, number of dependents living at home, ethnicity, education history and employment status, IMD score based on their postal code, relevant medical history, and current medications.

Physical Measurements
Weight and body composition were measured using Tanita BC-300MA calibrated scales to the nearest 0.1 kg. Height was measured using a portable stadiometer. Waist circumference was measured halfway between the point of the lowest rib and the iliac crest, and hip circumference was measured at the maximum circumference of the buttocks [53]. All measurements were taken to the nearest 1 mm. BP was measured with patients seated at rest for at least 10 minutes. All assessors were trained in accordance with the departmental protocols.

Fasting Blood Sample
A fasting venous blood sample was collected for HbA 1c , plasma glucose, lipid profile, serum urea and electrolytes, eGFR, creatinine, and liver function tests. A urine pregnancy test was performed in women of childbearing age to exclude pregnancy. Pregnant women were excluded from the trial. Laboratory results were assessed by independent laboratory staff.

Retention and Withdrawal
Participants had the right to withdraw from the trial at any time. Participants were considered as withdrawn from the trial if they withdrew from the study intervention voluntarily or if they failed to return for follow-up assessments. Participants could also be removed by the principal investigator if this was considered necessary for medical reasons or due to ineligibility arising during the study (eg, pregnancy). Reasons for withdrawal were recorded, and their GP was notified with recommendations for follow-up care where appropriate. Participants who withdrew continued to have weight and HbA 1c collected from their routine diabetes clinic or GP visits for the duration of the 12-month trial unless they did not consent to this at baseline. No incentives were provided to the participants to promote retention and follow-up.

Adverse Events
All adverse events in the 12-month study period were recorded by following the Good Clinical Practice and Health Research Authority processes. Nonserious adverse events such as constipation, fatigue, or hair loss were recorded when participants informed the trial team.

Qualitative Evaluation
In-depth semistructured interviews were conducted with a subset of 10 ILED and 10 CLED participants at the end of the active weight loss phase (ILED week 28, CLED week 12). HCPs (n=6) delivering the programs were also interviewed near study completion. All interviews were conducted by an independent research assistant trained in qualitative interview techniques. Interviews explored the participant and HCP experiences of the ILED and CLED programs and the use of the Oviva app.
Purposive sampling was used to select trial participants with a range of success in terms of actual weight loss and HbA 1c reduction, as well as participants' perceived success. Participants with ethnicities other than White-British were included where possible, with a fairly even split between men and women. All participants gave written informed consent to be interviewed and were assured that their data would be anonymized. The interviews were audiorecorded and transcribed verbatim for thematic analysis.

Sample Size and Statistical Analysis
The total number of participants recruited for the study was 79. The sample size was selected to allow an estimate of uptake within ±6.6% of the target uptake of 10% [55] with 95% confidence while allowing the research team to obtain sufficient data on the feasibility and potential efficacy of the ILED and CLED.
This study will not undertake significance tests of changes to the primary outcome measures. Descriptive, graphical (summary), and basic inferential statistics of outcomes will be presented as appropriate, for example, frequencies and percentages, mean and SD, or median and quartiles. Confidence intervals (95%) will be calculated to show the change from baseline in the outcomes for each group.
Questionnaires used as outcome measures are quantitative and will be analyzed using appropriate descriptive statistics as per standard.
Changes to diabetes medication will be presented using the medication effect score (MES). The MES for a participant is the sum of the MES for each of their individual medications, where MES=actual drug dose/maximum drug dose×drug mean adjustment factor. A decrease in MES corresponds to a decrease in the use of diabetes medications [56]. Changes to BP medications will be presented using a Treatment Intensity Score (TIS) defined as the actual drug dose/maximum drug dose [57]. The TIS for a participant is the sum of the TIS for each of their medications, and a decrease in TIS indicates a decrease in BP medications.
An ITT analysis using multiple imputations will conducted for percentage weight loss, HbA 1c , and MES. All other outcomes will be presented only for those who completed the trial.

Data Management
Data were recorded on hard copy case report forms and subsequently transferred to a database with ranges and programmed validation checks to aid reliable data entry. Data are held on secure servers at MFT.

Trial Steering Committee and Trial Management Group
The trial steering committee provided oversight for participant safety and included 2 co-principal investigators (BI and MH), an external endocrinologist and independent external advisor with experience of LEDs in the management of T2D. The committee met every 3 months to review and ensure the safety aspects of the trial. The trial management group, which comprised the chief investigators, diabetes specialist dietitians, a diabetes specialist nurse, and research nurses, evaluated all adverse events. The trial could have been stopped by the sponsor, chief investigators, or the trial management group or trial steering committee on the basis of new safety information or for other reasons given by the research ethics committee, but this was not required. The trial was subject to inspection and audit by MFT as the trial sponsor.

Results
The project was funded in May 2017, ethical approval was obtained in August 2017, and enrollment began in February 2018. In total, 79 participants were recruited and randomized to the ILED (n=39) and CLED (n=40) arms of the trial. Data collection was completed in February 2020. Data analysis began in June 2020, and the first results are expected to be submitted for publication in 2021.

Discussion
This is the first study to compare an ILED with an isocaloric CLED to achieve and maintain weight loss and normoglycemia among patients with T2D and overweight and obesity. The study will inform the acceptability and potential efficacy of high-frequency remote follow-up in patients with T2D and overweight and obesity undertaking low energy diets. It will also contribute to the limited data on the safety and efficacy of patients with T2D on insulin undertaking a ILED or CLED.
The study did not have a standard or best practice control group for comparison and was not powered to show statistical differences between the groups. However, the planned quantitative and qualitative analyses will assess the feasibility of the programs and inform the case for a future definitive trial.