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 [3-6]. 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].

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 [11-14]. 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.

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 [29-31].

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].

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.

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.

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.

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.

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 m² (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.

Table 1 provides a summary of the measurements collected at baseline and weeks 8, 12, 28, and 52 by a research nurse.

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.

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.

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.

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 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.

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.