Genetically-Guided Medical Nutrition Therapy in Type 2 Diabetes Mellitus and Pre-diabetes: A Series of n-of-1 Superiority Trials

Type 2 diabetes mellitus (T2DM) is a heterogeneous metabolic disorder of multifactorial etiology that includes genetic and dietary influences. By addressing the latter, medical nutrition therapy (MNT) contributes to the management of T2DM or pre-diabetes toward achieving glycaemic control and improved insulin sensitivity. However, the clinical outcomes of MNT vary and may further benefit from personalized nutritional plans that take into consideration genetic variations associated with individual responses to macronutrients. The aim of the present series of n-of-1 trials was to assess the effects of genetically-guided vs. conventional MNT on patients with pre-diabetes or T2DM. A quasi-experimental, cross-over design was adopted in three Caucasian adult men with either diagnosis. Complete diet, bioclinical and anthropometric assessment was performed and a conventional MNT, based on the clinical practice guidelines was applied for 8 weeks. After a week of “wash-out,” a precision MNT was prescribed for an additional 8-week period, based on the genetic characteristics of each patient. Outcomes of interest included changes in body weight (BW), fasting plasma glucose (FPG), and blood pressure (BP). Collectively, the trials indicated improvements in BW, FPG, BP, and glycosylated hemoglobin (HbA1c) following the genetically-guided precision MNT intervention. Moreover, both patients with pre-diabetes experienced remission of the condition. We conclude that improved BW loss and glycemic control can be achieved in patients with pre-diabetes/T2DM, by coupling MNT to their genetic makeup, guiding optimal diet, macronutrient composition, exercise and oral nutrient supplementation in a personalized manner.


Intervention and comparator
The energy requirements of all participants were estimated using the Mifflin-St. Jeor prediction equations (2,3) and the level of physical activity of each patient. The appropriate energy deficit level was calculated based on the estimated energy requirements (EER) and the goal for a 5% body weight (BW) loss, as suggested in the clinical practice guidelines for the treatment of prediabetes and T2DM (4). Two interventions were administered in a cross-over, non-randomized design. The procedure of each trial is detailed in Figure 1. Each patient was initiated with 8 weeks of conventional medical nutrition therapy (MNT), followed by 1 week of washout, and then a precision MNT intervention was applied for an additional 8-week period. For the conventional dietary intervention, the Problem, Etiology, Signs and Symptoms (PES) statement was structured for each patient, based on the nutrition care process (NCP) model, the revised standards of practice and the phenotypic characteristics of participants (5,6). Moreover, recommendations were based on the clinical practice guidelines of the ADA and Academy of Nutrition and Dietetics (AND), and as the latest evidence of higher hierarchy regarding prediabetes/T2DM MNT (7)(8)(9). With regard to exercise, this remained similar according to the patients' previous schedule unless otherwise indicated.
As for all cases response to the treatment was inadequate according to the predefined NCP outcomes within 8 weeks of the initial intervention, the genetic profile of each patient was used to guide a more personalized lifestyle intervention (diet and exercise), based on the available scientific evidence. No changes were made to the pharmacological treatment of participants. Baseline and end of treatment period measurements were recorded for each participant.

Adherence to the dietary treatment
Adherence to the treatment was assessed qualitatively, through telephone interviews by experienced dietitians (M.G.G, E.L). Previous 24h diet recalls were collected through these interviews regarding the diet of the participants. This is a retrospective short-term method for assessing dietary and food intake often performed through telephone interviews. Adherence to the oral nutrient supplementation (ONS) treatment (whenever applicable) was self-reported and exercise adherence was evaluated in the telephone interviews using specific questions assessing adherence and possible barriers.

Phenotypic profiling related to diabetes and weight status
Detailed phenotypic profiling of patients was performed by a multidisciplinary team of registered dietitians (E.L. and M.G.G) and endocrinologists (D.G.G), including a complete assessment of their dietary intake, physical activity levels and diabetes risk factors.

Anthropometric characteristics
BW and height of participants were measured during morning hours, by an experienced dietitian, with the use of a digital scale (SECA 813, SECA Group, Hamburg, Germany) and a wall-mounted stadiometer (SECA 216, SECA Group, Hamburg, Germany). At least duplicate measurements were taken. Body mass index (BMI) was calculated for each patient and weight status was defined according to the World Health Organization BMI thresholds (10). Waist circumference was measured at the iliac crest, with patients on a horizontal plane, and central obesity was diagnosed in those with a perimeter exceeding 94 cm (11).

Assay of blood markers
Morning fasting blood samples were collected from all patients for blood glucose and a lipidemic profile assay, including total cholesterol (TC), triglycerides (TG), high-density lipoprotein (HDL) and low-density lipoprotein (LDL) cholesterol. Moreover, glycosylated haemoglobin (HbA 1c ) of all patients was assessed from a capillary fingerstick using a DCA 2000 analyzer (Bayer Diagnostics, Tarrytown, NY, USA). With respect to the availability of micronutrients, ferritin serum levels were assessed according to the International Committee for Standardization in Haematology (12) and 25hydroxy vitamin D (25(OH)D) concentrations were evaluated via liquid chromatography-tandem mass spectrometry (LC-MS).

Blood pressure
Blood pressure (BP) was measured with patients in a seated, relaxed position, in the arm exhibiting the highest BP value, using an oscillometric device (Microlife Exact BP, Microlif AG, Widnau, Switzerland). Duplicate measurements were taken.

Outcomes of interest
For all participants, primary outcomes involved change (Δ) in BW (kg) and fasting plasma glucose (FPG) concentrations (mg/dL). Secondary outcomes were systolic and diastolic involved blood pressure (BP) (mm Hg), depressive symptoms (whenever applicable) and possible adverse events.

Depressive symptoms
For patients exhibiting depression traits, the Beck Depression Inventory (BDI) was used to assess possible improvement (13). The scale consists of 21 questions each with four possible answers in a Likert scale. Each answer provides a score of 0-3, giving a pooled total score that can range from 0 (complete lack of depression) to 63. The scale has been officially translated and validated in the Greek language and used in a plethora of research.
Genotyping and genetic scores for obesity, habitual coffee consumption, T2DM and elevated fasting plasma glucose Buccal swabs were stored at 4 o C and processed for DNA extraction within 24 hours using the Purelink Genomic DNA extraction mini kit (Invitrogen, Thermo Fisher Scientific, Waltham, Massachusetts, USA). Single nucleotide polymorphism (SNP) status was assessed on an Open Array Quant Studio 12X flex thermocycler (Applied Biosystems, Waltham, USA). The success rate of genotyping was 99%.
The GRS for obesity was calculated as the weighted sum of risk alleles across 32 SNPs indicating the highest association with BMI and/or WHR (and affected recommendations for protein intake in proposed diet) (17)(18)(19).

Data monitoring
Three experts (A.G.E., D.G.G. and D.S.) comprised the data monitoring committee (DMC) according to the SPENT guidelines. The DMC had full access to the anonymously coded dataset of the study.

Statistical analyses
As each patient received a different, personalized treatment, this heterogeneity does not allow for a quantitative synthesis of data. Thus, for each patient, outcomes at the end of each intervention cycle were compared to the start of the intervention period (treatment-by-period interaction) (30). No missing data were apparent in the dataset.