Nutrition Interventions for Remission of Type 2 Diabetes: Potential Role for Diabetes-Specific Nutrition Formulas Free

The feasibility of diabetes remission pointed to a need to rethink the pathophysiological mechanisms of type 2 diabetes, which had been attributed to insulin resistance and loss of pancreatic islet β-cells (13). Early theories to account for remission after bariatric surgery invoked a role for incretin hormones, especially glucagon-like peptide 1 (GLP-1), but those theories were not supported by experimental evidence (14,15). Rather, it was shown that caloric restriction and accompanying declines in intrahepatic and pancreatic fat were causally related to restoration of β-cell function, leading to diabetes remission (15–17). This evidence led to the development of the twin-cycle hypothesis, which is supported by experimental and clinical evidence (10,16).

In the twin-cycle hypothesis (Figure 1) (10,16), chronic positive calorie balance increases de novo lipogenesis and fat accumulation in the liver, leading to a positive-feedback cycle of hepatic insulin resistance, elevated plasma glucose levels, and hepatic steatosis. Excess intrahepatic fat increases plasma levels of triglyceride-rich very-low-density lipoproteins (VLDLs), leading to fat accumulation in the pancreas and subsequent impairment of the β-cell response to glucose. These dual cycles ultimately lead to impaired β-cell function and clinical diabetes. Crucially, excess pancreatic fat causes β-cells to enter into a de-differentiated state, in which specialized functions such as insulin production are impaired; removal of the excess pancreatic fat allows β-cell function to recover, but only if fat removal occurs within the first few years of impaired β-cell function (18). Otherwise, impairments in β-cell function eventually become irreversible (18).

The twin-cycle hypothesis and its biological underpinnings lead to three clinically important inferences. First, it may be possible to temporarily reverse the progressive course of type 2 diabetes by negative energy balance sufficient to remove excess hepatic and pancreatic fat (19). Second, each person has a unique threshold beyond which hepatic and pancreatic fat accumulation occurs. This concept is known as the personal fat threshold. It helps to explain why individuals of normal weight may develop type 2 diabetes and why diabetes remission can be attained in some people after substantial weight loss even though they still have clinical obesity (19,20). Third, remission of diabetes may be possible only in people whose β-cell function can be recovered (i.e., those with a relatively short duration of diabetes). The clinical studies confirming these inferences have been reviewed previously (21).

As the scientific evidence of diabetes remission grew, it required a definition to be identified, accepted, and targeted as a treatment option. In 2021, the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD) issued a joint consensus report on the definition and interpretation of diabetes remission (Figure 2) (22,23); the consensus was based on the work of an expert panel, including representatives from the ADA, EASD, Diabetes UK, the Endocrine Society, and the Diabetes Surgery Summit. The expert panel defined diabetes remission as a return of A1C to <6.5% that persists for at least 3 months in the absence of glucose-lowering pharmacotherapy. Importantly, the panel noted that the glycemic levels used to define diabetes remission (A1C <6.5%) were an arbitrary point on a continuum and that the optimal cutoff value is not yet known. Indeed, Diabetes Canada recognizes two categories of diabetes remission, with the strictest category requiring normal glucose levels (<6.0%) (24). Additionally, many people can experience prolonged improvements in glycemic control and reduced cardiometabolic risk factors during adoption of a low-calorie diet and intensive lifestyle intervention even if they do not attain an A1C <6.5% (25–27). In this regard, diabetes remission could be a viable treatment target for many people, but it may not always be the outcome. The feasibility of maintaining diabetes remission for many years has not yet been established.

Total diet replacement involves replacing all meals during a specified period (e.g., 8–12 weeks) with formulas, often in the form of shakes or soups, supplying defined levels of macronutrients with the goal of low or very low energy intake (e.g., 800 kcal/day). Total diet replacement products also supply essential vitamins and micronutrients.

Table 1 is a summary of studies using very-low-calorie total diet replacement with the goal of inducing diabetes remission in people with recently diagnosed type 2 diabetes. In all studies except the ReTUNE (Remission of Type 2 Diabetes after Weight Loss in “Normal” Weight People) trial (28), participants had overweight or obesity (generally BMI ≥27 kg/m²). Interventions consisted of a period of total diet replacement as part of a comprehensive integrated program that also included guided food reintroduction and intensive education and clinical follow-up. Participants also discontinued glucose-lowering medications at the beginning of the interventions. The seminal study was DiRECT (Diabetes Remission Clinical Trial), which served as an intervention model for most of the subsequent studies (26,29).

The Counterpoint (Counteracting Pancreatic Inhibition of Insulin Secretion by Triglyceride) study (17) enrolled people (n = 11) with type 2 diabetes of <4 years’ duration, with an A1C of 6.5–9.0% and a stable BMI of 25–45 kg/m². The intervention was 8 weeks of total diet replacement plus nonstarchy vegetables, with a target energy intake of 600 kcal/day. At the end of the first week, mean fasting plasma glucose (FPG) declined from 166 ± 7 to 106 ± 7 mg/dL, and this change was sustained during the 8-week intervention. The intervention was also associated with improved hepatic insulin sensitivity, improved β-cell function, and reduced hepatic and pancreatic triacylglycerol content.

The Counterbalance (Counteracting Beta-Cell Failure by Long Term Action to Normalize Calorie Intake) study (30) was intended to test the durability of the diabetes remission attained in the Counterpoint study. The Counterbalance study enrolled people (n = 30) with type 2 diabetes, a BMI of 27–45 kg/m², and an A1C ≤9.5%. Unlike other studies, some participants had longstanding diabetes (up to 23 years). The 8-week intervention phase was similar to that of the Counterpoint study and was followed by a 6-month weight maintenance phase. At the end of the 8-week intervention, 40% of participants were in diabetes remission (according to an older definition of A1C <7.0%), and 43% were in remission at 6 months. The remission rate was substantially higher in people recently diagnosed with diabetes (<4 years) versus those who had a longer diabetes duration (60 and 21%, respectively).

DiRECT was an open-label, cluster-randomized, controlled trial conducted at 49 primary care practices in the United Kingdom (26,29). Individual practices were randomized to provide an integrated, structured weight management program (intervention) or guideline-directed care (control). Participants (n = 306) were adults (20–65 years of age) who were diagnosed with type 2 diabetes within the preceding 6 years and had overweight or obesity (BMI of 27–45 kg/m²). Exclusion criteria included insulin use or A1C ≥12%. The 12- to 20-week intervention involved total diet replacement supplying 825–853 kcal/day and required withdrawal from all glucose-lowering and antihypertensive medications. This was followed by stepped reintroduction of food over 2–8 weeks and structured support for long-term maintenance of weight loss.

At 12 months, diabetes remission (defined as A1C <6.5%) was reported in 46% of participants in the intervention group and 4% of those in the control group (odds ratio [OR] 19.7, 95% CI 7.8–49.8, P <0.0001) (29). At 2 years, 36% of participants in the intervention group and 5% of those in the control group maintained diabetes remission (adjusted OR 25.8, 95% CI 8.3–80.8, P <0.0001) (26). At both 1 year and 2 years, remission was more common among participants who lost more weight compared with those who lost less weight. For example, among those who experienced a weight reduction ≥15 kg at 1 year, 86% (31 of 36) attained diabetes remission, versus 34% (19 of 56) among those who attained a weight loss of 5 to <10 kg. Other factors associated with increased likelihood of diabetes remission were male sex, younger age, good mental health, use of fewer glucose-lowering medications before enrollment, shorter duration of type 2 diabetes, higher fasting insulin levels, and lower FPG levels (31). According to long-term follow-up data from patients in DiRECT, 10% (12 of 118) maintained diabetes remission at 5 years (32).

During the first year of DiRECT, participants in the intervention arm had six noncardiovascular, three moderate or major cardiovascular, and 16 major diabetes-related serious adverse events (1). Corresponding numbers in the control arm were 3, 1, and 14, respectively (all differences were nonsignificant). In subsequent years, participants in the intervention arm had lower rates of noncardiovascular adverse events than those in the control arm, and this difference was statistically significant when all 5 years of follow-up were considered together (34 versus 69, estimated incidence rate ratio 0.5, 95% CI 0.3–0.8, P = 0.008). Five-year estimated incidence rates of cardiovascular or diabetes-related serious adverse events were not significantly different (1.1, 95% CI 0.6–2.1, and 1.0, 95% CI 0.7–1.3, respectively) (1). Because the participants in DiRECT were predominantly White Europeans, additional trials were needed to explore its applicability in other populations.

A feasibility study using an intervention similar to that of DiRECT enrolled people of Black or mixed racial backgrounds in Barbados (33). Participants (n = 25) were 26–68 years of age, had type 2 diabetes diagnosed within the preceding 6 years but were not using insulin, and had a BMI ≥27 kg/m². The intervention involved 4 weeks of total diet replacement specifically designed for use in diabetes; this was followed by a 4-week transition period to a solid diet with support and guidance from a local dietitian. An FPG <126 mg/dL was attained by 15 participants (60%), and there was a marked relationship with degree of weight loss (response observed in 9 of 11 [82%] of those who lost ≥10 kg and 6 of 14 [43%] of those who lost <10 kg).

The DIADEM-1 (Diabetes Intervention Accentuating Diet and Enhancing Metabolism) trial (34), which enrolled people from the Middle East and North Africa (n = 147), had many characteristics in common with DiRECT and was limited to younger participants (18–50 years of age) and those diagnosed with type 2 diabetes within the preceding 3 years. One difference from DiRECT was that participants in the DIADEM-1 trial were randomized on an individual basis rather than by cluster randomization. The intervention was similar to that of DiRECT, including total diet replacement supplying 800–820 kcal/day, but it also included physical activity support. At 1 year, 61% of participants in the intervention group attained diabetes remission compared with 12% in the control group (OR 12.03, 95% CI 5.17–28.03, P <0.0001). Five serious adverse events were reported in the control group versus none in the intervention group.

STANDby (South Asian Diabetes Remission Feasibility Trial) (35) followed the same intervention as DiRECT but in people of South Asian ethnicity (n = 25) with type 2 diabetes for ≤4 years, and it allowed a lower BMI (25–45 kg/m²). Body weight change was −7.7 kg (−7.2%) in the intensive intervention (total diet replacement) group and −1.2 kg (−1.4%) in the usual care control group (P = 0.005), with diabetes remission attained by 5 of 13 compared with 0 of 12, respectively (P = 0.039). No serious adverse events were reported.

DiRECT-Aus (Diabetes Remission Clinical Trial–Australia) (36) was an open-label, single-arm trial conducted across 25 primary care practices in New South Wales, Australia. Participants (n = 155) were 20–65 years of age with type 2 diabetes of ≤6 years’ duration and not treated with insulin, a BMI >27.0 kg/m², and a baseline A1C ≥6.5%. The intervention was similar to that in DiRECT. At 12 months, 56% of participants attained diabetes remission. Furthermore, rates of remission were correlated with the magnitude of weight loss; among those who lost >15% of body mass, 87% had diabetes remission at 12 months versus 38% in those who lost ≤5% of body mass. There were two serious adverse events of hypotension during the 12-month study period, both of which required hospitalization.

Recently, the ReTUNE study (28) reported diabetes remission in people with nongenetic type 2 diabetes and a normal BMI (<27 kg/m²). Twenty participants (aged 20–79 years) with type 2 diabetes for <6 years, a mean BMI of 24.8 kg/m², and not using insulin were enrolled. The intervention involved two cycles of weight loss, with an optional third cycle and a goal of 5% weight loss per cycle. Each cycle consisted of 2–4 weeks of total diet replacement plus nonstarchy vegetables followed by 4–6 weeks of weight maintenance. All glucose-lowering medications were stopped before treatment initiation. Fourteen of the 20 participants (70%) attained diabetes remission, which was accompanied by declines in plasma insulin, liver fat, plasma VLDLs, and body weight and recovery of β-cell function. These results support the concept of the personal fat threshold, in which some people with normal body weight develop type 2 diabetes secondary to accumulation of ectopic hepatic and pancreatic fat (28).

There are three common observations across these trials. First, all of these studies used total diet replacement, were intensive, and had multiple visits with care management teams (with some exceptions for the proof-of-concept Counterpoint study). Second, the studies had similar outcomes despite differences in the racial background of participants. Finally, participants exhibited some weight regain after food reintroduction, but the intensive support provided served to limit weight regain.

In the past, a major barrier to achieving significant weight loss with a very-low-calorie diet has been the difficulty of sustained adherence to the diet while continually having to decide what to eat (37). In contrast, participants in DiRECT reported that adherence during the total diet replacement phase was easier than they had anticipated, and some participants even chose to continue this phase to attain greater weight loss (37). Other trials also reported high rates of study completion (34,36), supporting the feasibility of this treatment strategy in real-world settings. Despite the acceptability of total diet replacement, individualized behavioral support is an essential component of all treatment phases (37).

The ability to maintain a reduced body weight is crucial for maintenance of diabetes remission and for sustained improvements in other health outcomes associated with weight loss (1). Long-term maintenance of weight loss is challenging, however, and research on maintenance of reduced weight is less developed than research on weight loss (3–5). Nevertheless, after the initial weight loss phase, DiRECT used food reintroduction and long-term maintenance phases based on proof- of-concept studies (6,7) and evidence from randomized controlled trials (RCTs) and systematic reviews. Those prior studies had identified factors associated with successful maintenance of reduced weight (3,4,8–10,38).

Features of the food reintroduction and weight maintenance phases in DiRECT are shown in Table 2. These features are consistent with a recent systematic review (11), which identified several determinants with strong evidence for efficacy in weight loss maintenance, including high self-efficacy for weight management and exercise, reduced intake of unhealthy food, portion control, reduced energy intake, increased fruit and vegetable intake, and decreased consumption of sugar-sweetened beverages.

During DiRECT, guideline-directed protocols were followed for reintroduction of oral antidiabetic and antihypertensive medications (12). Additionally, participants had access to a “rescue plan,” in which they could return to a period of meal replacement if they regained >2 kg (2–4 weeks of partial diet replacement) or >4 kg (4 weeks of total diet replacement) (9). Half of the participants used one or more rescue periods during the first 2 years. Although orlistat was offered, no participants were using it at 1 year, and only three (2%) were using it at 2 years (8). At 5 years, no participants were using orlistat, and three (2%) were using a GLP-1 receptor agonist (1).

Continuous glucose monitoring (CGM) has been shown to improve glycemic control in both type 1 and type 2 diabetes (13), and it has been proposed as a potential tool for guiding and stimulating behavior change in people with diabetes, as well as in people with obesity (14). Research on this use of CGM to promote behavior change is still in its early stages, but early evidence suggests it could be valuable for this purpose, although its high cost may widen disparities (15). Ongoing programs (16) and studies (trial ID NCT05483140) are exploring the added benefit of CGM for maintaining weight loss and its associated improvements in glycemic control, including diabetes remission.

All the studies shown in Table 1 used total diet replacement during weight loss. The meal replacement products were formulated to provide essential macronutrients, micronutrients, and vitamins. However, except for the Barbados feasibility study, the meal replacement products were developed for use in the treatment of obesity and not specifically designed to manage dysglycemia. Although the studies showed average improvements in glycemic control during total diet replacement even after discontinuation of glucose-lowering medication, there is limited information available about the adequacy of glycemic control in participants who did not achieve diabetes remission. Indeed, there is evidence that practitioners have concerns about glycemic control during the total diet replacement phase (39).

Diabetes-specific nutrition formulas (DSNFs) are specialized medical nutrition therapies designed for use in the management of dysglycemia, including use as meal replacements (40,41). DSNFs may contain complex carbohydrates with low glycemic indices, fiber, high levels of monounsaturated fatty acids and/or polyunsaturated fatty acids, proteins, amino acids and their metabolites such as β-hydroxy-β-methylbutyrate, as well as vitamins and minerals (42). These nutrients are essential for optimizing metabolic function and minimizing cardiometabolic risk and for controlling caloric intake (41–43).

Extensive research on DSNFs has established their beneficial effects versus standard nutrition formulas on glycemic control and numerous cardiometabolic conditions and risk factors (41–44). Across a wide range of studies in various medical conditions, DSNFs have lowered health care utilization and costs (40–43).

Currently, the Barbados feasibility study is the only trial of diabetes remission using a DSNF for total diet replacement. Although the results were comparable to the other studies in Table 1, more studies are needed to determine the relative advantages or disadvantages of using a DSNF in place of other total diet replacement products in this setting. The ongoing RESET (Digitally-Enabled Weight Management Program on T2DM) study (trial ID NCT05483140) in the United Kingdom is assessing the feasibility and real-world impact of a digitally enabled lifestyle program with a DSNF as part of a low-calorie diet. It may provide additional insights on this issue, as may the DROP (Diabetes Remission Outcome Protocol) program described in more detail below.

Meal replacement products other than those listed in Table 1 are untested in the setting of diabetes remission. Furthermore, their ability to manage dysglycemia or provide essential micronutrients varies widely (40,41,45).

Ongoing studies of diabetes remission involving a period of total diet replacement include DIGEST (Digital Diabetes Remission Trial; trial ID NCT05647226) and CARBCOUNT (Dietary Strategies for Remission of Type 2 Diabetes; trial ID NCT04943926).

Recently, the Joslin Diabetes Center developed the DROP program for people with type 2 diabetes who meet three criteria: diagnosed within 5 years, treated with no more than two glucose-lowering medications and not on insulin, and having an A1C <8% (46). DROP consists of two phases: a 12-week intervention phase followed by a 9-month maintenance phase.

During the intervention phase, participants follow a combination of two nutrition-based methods previously shown to induce weight loss and improve glycemic control: 1) time-restricted eating, in which participants fast for 16 hours each day, and 2) a very-low-calorie diet (800–1,000 kcal/day) using a DSNF and whole food. Participants also receive cognitive behavioral therapy and an individualized exercise plan aimed at maintaining muscle mass. Participants use CGM as a tool for monitoring the effects of lifestyle interventions on glycemic levels. Participants are advised to stop glucose-lowering medications at the beginning of the intervention phase, but they may reintroduce them during the first 2 weeks based on prespecified CGM criteria.

During the maintenance phase, participants are instructed to resume a diet of 1,500–1,800 kcal/day with the aim of maintaining weight loss and diabetes remission. They may also repeat 1–2 weeks of the intervention protocol if their glycemic time in range (70–180 mg/dL) drops below 85%.

In a small feasibility study (n = 5), participants lost an average 9 kg (8.1%) of body weight by 12 weeks, and their mean A1C decreased to 6.14 ± 0.48% from a baseline of 6.44 ± 0.3% (46). Participants also reported significant improvements in quality of life and diabetes distress scores (P <0.05) and a significant mean increase in the ratio of lean muscle mass to total body weight (0.03 ± 0.02, P = 0.03). Only one patient required a low dosage of metformin (1,000 mg/day).

An ongoing RCT is evaluating the DROP protocol in a larger cohort. For several reasons, DROP is a promising approach for inducing diabetes remission and improving quality of life in selected people with type 2 diabetes. First, the diet intervention is a mix of a DSNF and one meal of whole food each day, which may improve adherence during the intervention phase. Second, there is support for the use of time-restricted eating to improve glycemic control and reduce body fat, blood pressure, and intrahepatic fat (47,48). Third, the intervention emphasizes maintenance of muscle by increasing protein intake (1–1.5 g/kg body weight) and daily strength-based physical activity. Maintenance of muscle mass is known to promote weight maintenance after weight loss and to enhance peripheral glucose uptake (45,49,50). Loss of muscle mass is a major risk during a hypocaloric diet and is associated with reduced energy expenditure, function, and quality of life (51–53). Fourth, CGM adds valuable information to help guide lifestyle changes, diet, and exercise (54).

In 2022, the joint ADA/EASD treatment guidelines for the management of hyperglycemia in type 2 diabetes identified substantial (>10%) weight loss and weight loss early in the course of type 2 diabetes as a means to increase the chance of diabetes remission (22,23). Consequently, at least three major professional organizations have endorsed diabetes remission using a low-energy diet as a viable therapeutic goal in selected individuals with recently diagnosed type 2 diabetes and overweight or obesity (24,55,56).

Because international treatment guidelines now support diabetes remission as a treatment option, national groups have provided details on implementation of this strategy in clinical practice. Diabetes Canada recommends a very-low-calorie diet (∼800–850 kcal/day) using meal replacements for 3–5 months to attain a body weight loss of >15 kg, followed by structured food reintroduction and increased physical activity. Individuals considered eligible are adults with a BMI of 27–45 kg/m², with type 2 diabetes of <6 years’ duration, with an A1C <12%, and who are not using insulin (24). Diabetes Canada has also produced a user’s guide containing practical information and resources to help clinicians implement diabetes remission strategies in their practices (57).

In recently updated guidelines, the EASD Diabetes and Nutrition Study Group has recommended that diabetes remission can be attained through sustained weight loss in people with overweight or obesity (56). To attain sustained weight loss, these guidelines recommend a low-energy total diet replacement program provided by trained professionals for 12–20 weeks.

To date, the largest real-world implementation of a diabetes remission program has been in the United Kingdom, where the National Health Service has implemented its Type 2 Diabetes Path to Remission Programme in partnership with Diabetes UK (56). The program is implemented at specific practices in the United Kingdom. Eligible individuals are similar to those enrolled in DiRECT, and the intervention entails the use of total diet replacement for 12 weeks, followed by healthy lifestyle support.

Two-year data from DiRECT were used to analyze the effectiveness of the intervention in terms of quality- adjusted life years (QALYs) gained and cost-effectiveness (58). Although the DiRECT intervention group had higher costs than the control group during the study period, the analysis concluded that the intervention was more effective than the control treatment in terms of QALYs gained, and that it is expected to be a cost savings within 6 years. The cost savings are expected to arise from reduced lifetime health care costs resulting from diabetes remission and reduced complications. The model predicted mean total lifetime cost savings of about £1337 (∼$1,700 USD in 2020) favoring the intervention group.

It is important to remember that these projected benefits apply to individuals similar to those in DiRECT; in other patient groups less likely to attain weight loss, reduced cardiometabolic risk, or diabetes remission, such long-term effectiveness or cost savings may not be realized. Furthermore, additional research and real-world experience are needed to understand costs and cost-effectiveness in diverse economic settings and reimbursement models, including uninsured and underinsured populations. It is important that future studies assess both the economic and clinical benefits of interventions intended to induce diabetes remission.

Diabetes remission is a feasible treatment goal in some people with type 2 diabetes and overweight or obesity, especially those who have been diagnosed within the preceding 6 years, have an A1C <12%, and are not using insulin. Current evidence indicates that diabetes remission is possible in one-third to one-half of such people through a comprehensive program starting with a period (∼12 weeks) of very-low-calorie diet attained through total or partial meal replacement supplying 820–850 kcal/day. Total diet replacement products studied for diabetes remission were acceptable to study participants and provided essential nutrients. Glycemic control improved on average, although there is limited information about the adequacy of glycemic control in study participants who did not achieve diabetes remission. Among individuals with obesity at baseline, diabetes remission is significantly more likely among those who lose at least 10–15 kg of body weight. This induction phase is followed by gradual reintroduction of conventional foods and a weight maintenance phase, which requires continued behavioral and nutritional support to prevent excessive weight regain.

Currently available evidence supports the effectiveness of diabetes remission using total diet replacement in terms of QALYs gained and the potential for long-term cost-effectiveness. The approach used by the DROP program, which involves time-restricted eating and partial meal replacement, has shown promising early results and may be easily adapted to individuals’ cultural preferences.

Although these strategies for diabetes remission appear feasible for primary care practices, there are limited studies of real-world implementation. More research is also needed to explore the feasibility of these strategies in geographically and ethnically diverse populations, as well as their feasibility in different conditions related to food security and reimbursement for meal replacement products. There is also a need for more long-term studies of clinical effectiveness, cost-effectiveness, and cardiorenal protective properties related to diabetes remission, as well as the potential advantages of diabetes-specific nutrition formulas for total diet replacement and the DROP approach.

Medical writing support was provided by Ken Scholz, PhD, who is affiliated with Innovative Biopharma, LLC.

Financial support for the expert roundtable and medical writing support was provided by Abbott.

M.T. is a principal investigator for a research study funded by Abbott Nutrition and has received honoraria from Abbott Nutrition for speaking and advising. H.B. has had research funding paid to his institution by Abbott, Amgen, AstraZeneca, Boehringer Ingelheim, Ceapro, Eli Lilly, Gilead, Ionis/Janssen, Kowa Pharmaceuticals, Madrigal Pharmaceuticals, Merck, Novartis, Novo Nordisk, Pfizer, Sanofi, and Tricida. J.I.M. has received speaking honoraria from Abbott Nutrition and has served on advisory boards for Abbott Nutrition, Aveta.Life, and Twin Health. O.H. has received research support from Eli Lilly and Novo Nordisk and consultation fees from Abbott Nutrition.

M.T., J.I.M., and O.H. presented and discussed data at the scientific roundtable, wrote sections of the manuscript, and reviewed and edited the manuscript. H.B. presented and discussed data at the scientific roundtable and reviewed and edited the manuscript. Each author approved the final draft for submission. O.H. is the guarantor of this work and, as such, had full access to all the data and takes responsibility for the integrity of the data and the accuracy of the data analysis.