LEANDIP: Lean Diabetes in Asian Indians

Type 2 diabetes (T2D) is a significant global public health issue, with the International Diabetes Federation projecting that the number ofindividuals with diabetes will rise to 783 million by 2045¹. While obesity is a leading risk factor for T2D, a notable number of individuals who areneither obese nor overweight develop a form known as lean T2D²˒³. Despite the established links between obesity and diabetes, the exactpathophysiology of lean T2D, particularly impaired pancreatic insulin secretion, remains unclear⁴˒⁵. Additionally, the prevalence of lean T2Dvaries by race and ethnicity and is more common in non-white populations, yet much of the research has focused primarily on obesity-relateddiabetes. This indicates a need for better monitoring and understanding of lean T2D for effective public health strategies⁶.

A nationally representative study using data from the Longitudinal Ageing Study in India (LASI) Wave-1 examined 58,824 individuals aged 45years and above and found the prevalence of lean diabetes (BMI <18.5 kg/m²) to be 0.8% (95% CI: 0.7-0.9). The prevalence was higher amongthe poorest quintile at 1.1% (95% CI: 0.7-1.6) compared with 0.5% (95% CI: 0.3-0.6) in the richest quintile. A negative Erreygers' concentrationindex (ECI = -0.006) indicated a pro-poor inequality in lean diabetes distribution. Decomposition analysis revealed that 72% of the inequalitywas explained by economic status, followed by place of residence (24%) and education (20%). The study highlights the urgent need to focus onlean diabetes among socially and economically marginalized populations within India's healthcare framework⁶.

Lean diabetes is heterogeneous and under-studied at the very low end of body mass index (BMI), particularly among individuals with BMI <19kg/m². The study by Mohan et al. showed that lean diabetes (BMI <18.5 kg/m²), affecting about 1.5% of T2D patients, represents one end of theT2D spectrum and is characterized by earlier onset, greater insulin deficiency, poorer glycemic control, and higher risk of microvascularcomplications compared with higher BMI groups⁷. The most detailed phenotyping to date comes from Lontchi-Yimagou et al. (2022, India), across-sectional metabolic study of 73 South Indian males including 20 participants with "low BMI diabetes." Participants had a mean BMI of 18.3± 0.2 kg/m² and underwent systematic exclusion of known diabetes forms. Next-generation sequencing of 13 MODY genes and sixlipodystrophy/insulin-resistance genes detected no pathogenic variants, and autoimmune diabetes was excluded using anti-GAD65, IA-2, andZnT8 antibody assays. Physiological profiling included mixed-meal tolerance testing with C-peptide deconvolution, hyperinsulinemiceuglycemic clamps with glucose tracers, ¹H-MRS of hepatic and myocellular lipids, and DEXA. Compared with type 2 diabetes and lean controls,individuals with lean diabetes showed markedly reduced first-phase and total insulin secretion, greater peripheral insulin sensitivity, and nosignificant hepatic insulin resistance⁸. This remains the only detailed study from India with limited sample size without metabolomicinvestigation and long-term follow-up. In lean diabetes, subcutaneous and intra-abdominal adipose tissue may accumulate while β-cell reserveprogressively declines, and skeletal muscle mass and function decrease, contributing to impaired insulin secretion and reduced glucose uptake.(figure1) Several studies characterize lean diabetes using BMI thresholds and provide context for pancreatic phenotype and genetics. In the UgandaDiabetes Phenotype (UDIP) study, Kibirige et al. analyzed 500 newly diagnosed autoantibody-negative adults (160 lean, BMI <25 kg/m²; 340non-lean). Lean participants had lower visceral fat, leptin, and uric acid, fewer metabolic syndrome features, and substantially reducedinsulinogenic index and C-peptide than non-lean participants; insulin resistance (HOMA2-IR) was also lower

. Lean-diabetes trends fromKNHANES (Kim et al., 2024, South Korea; BMI <23 kg/m²) similarly showed increasing prevalence and lower HOMA-β and HOMA-IR comparedwith overweight or obese T2D

. Genetic stratification by BMI in large consortia has identified loci enriched among lean diabetes cases. Perry et al. (2012) stratified lean andobese cases using GWAS and Metabochip approaches and reported risk-variant enrichment among lean individuals, highlighting loci such asLAMA1

• Reviews synthesizing multi-ethnic datasets suggest that β-cell-related loci including TCF7L2, HNF1A/HNF4A, SLC30A8, and MTNR1Bcontribute more strongly to non-obese T2D than classical insulin-resistance loci
• Experimental work has also implicated KCNJ15 in impairedglucose-stimulated insulin secretion

Metabolomic studies have identified circulating metabolites associated with insulin resistance and metabolic dysfunction in lean or non-obesediabetes. Elevated branched-chain amino acids (BCAAs) such as leucine, isoleucine, and valine are linked to impaired insulin signaling andincreased diabetes risk

. Alterations in lipid metabolites including ceramides and long-chain fatty acids reflect lipotoxicity and disruptedinsulin signaling pathways

, while changes in acylcarnitines indicate mitochondrial dysfunction and incomplete β-oxidation

. Yagihashi summarized findings showing reduced β-cell mass (~35% in lean U.S. T2D and ~30-40% in Asian and European cohorts), reducedpancreatic volume, irregular borders, and increased pancreatic fat in diabetes

. Prospective data have linked fatty pancreas with higher risk ofincident T2D

• Quantitative MRI techniques including volume measurement and fat fraction estimation have been proposed as useful tools forpancreatic phenotyping in diabetes research
• Unnikrishnan and Mohan summarized fibrocalculous pancreatic diabetes (FCPD), historicallyassociated with lean individuals, and reported genetic associations including SPINK1 mutations in some populations

Emerging evidence suggests that lean individuals may have atypical etiologies including latent autoimmune diabetes in adults (LADA), maturity-onset diabetes of the young (MODY), pancreatogenic diabetes, lipodystrophy-related diabetes, and malnutrition-related diabetes mellitus(MRDM). There is ongoing debate regarding classification of MRDM as "type 5 diabetes," and we have previously published a critique on thispoorly defined entity . Nevertheless, lean individuals with diabetes represent an important group requiring comprehensive investigation.

Rationale:

Lean diabetes in Asian Indians remains poorly understood, and it is unclear whether it represents a distinct form of diabetes or lies at theextreme end of the diabetes spectrum. Current evidence is limited and largely derived from small studies, with little information on genomic,metabolomic, pancreatic imaging, and clinical phenotypes in this population. Therefore, a comprehensive integrated evaluation of these factors,distinguishing it from healthy lean without diabetes or non-lean T2D, is necessary to understand mechanisms and progression of lean diabetes.

9 10 11 12 2 13 14 15,16 17 17,18 19 20 21 22 23

Key Research question(s):

Lean diabetes in Asian Indians is an under-recognized phenotype of diabetes occurring in individuals with low body mass index (BMI <19kg/m²). Despite increasing recognition, the biological mechanisms, metabolic characteristics, and clinical progression of lean diabetes remaininadequately understood. Therefore, this study seeks to address the following key research questions:

1. Do lean Asian Indians with diabetes exhibit distinct genomic and metabolomic profiles 2. Are there unique structural and functional characteristics, including hepatic fat fraction, pancreatic fat fraction, pancreatic volume, andabdominal fat distribution, associated with lean diabetes? 3. What are the clinical, metabolic, nutritional, and body composition characteristics of lean adults with diabetes? 4. How does lean diabetes progress over time in terms of glycemic control, complication burden, and treatment requirements?

Study Objectives:

Define the objectives (maximum 4): The objectives should be SMART (Specific, Measurable, Achievable with the project's budgetand time, resourced within the project's budget, and time bound)

Study Objective :

1. Define the objectives (maximum 4): The objectives should be SMART (Specific, Measurable, Achievable with the project'sbudget and time, resourced within the project's budget, and time bound).

   Study Objective :

2. Primary Objective • To characterize the genomic and metabolomic profile of lean adults with diabetes (excluding type 1,monogenic, or secondary diabetes) in Asian Indian. 2. Secondary Objectives • To assess pancreatic and liver imaging, pancreatic functionalparameters, and detailed body composition using magnetic resonance imaging. • To evaluate clinical and metabolic characteristics, includingnutrition, lifestyle, psychological profiles, complication burden, and treatment requirements Methodology

Describe the research methods that could best achieve the study objectives under the following sub-headings (as applicable):

Study Design :

Multicentric prospective longitudinal comparative study (Annexure 4, Figure 2)

Study Area or Settings :

The study will be conducted at Diabetes Foundation (India), New Delhi and Dr. Kanungo's Diabetes Centre, Bhubaneswar

Study Subjects :

Inclusion Criteria:

Group 1- Lean Diabetes:

• BMI <19 kg/m²
• Age 18-50 years
• Newly diagnosed, treatment-naive diabetes according to ADA criteria

Group 2- Healthy lean without diabetes:

• BMI <19 kg/m²
• Age 18-50 years
• No history of diabetes and normal glycemic status (FPG, HbA1c and/or OGTT as per ADA criteria)

Group 3- Non-lean T2D:

• BMI >19 kg/m²
• Age 18-50 years
• Diabetes according to ADA criteria All the participants willing to provide written informed consent

Exclusion Criteria:

Proven Type 1 diabetes Monogenic diabetes Lipodystrophies, Mitochondrial diabetes Secondary diabetes (chronic calcific pancreatitis, pancreatectomy, steroid induced diabetes) Seriously ill patient Patients with history of acute myocardial infarction, advanced renal failure and liver disease tuberculosis, carcinoma, malabsorptionsyndrome, hyperthyroidism, HIV or any severe infection.

Pregnant or lactating women

Interventions :

The study does not involve any therapeutic or lifestyle intervention. Participants will undergo clinical evaluation, biochemical testing, genomicand metabolomic analysis, body composition assessment, and pancreatic imaging.

Comparison :

The study will involve comparative analyses, such as:

Comparative analyses will be performed across the three study groups to identify clinical, metabolic, body composition,pancreatic, hepatic, genomic and metabolomic characteristics associated specifically with lean diabetes.

Longitudinal comparisons of clinical, metabolic and treatment-related parameters will additionally be performed within the leandiabetes group at baseline, 6 months and 12 months.

Primary and secondary outcome measures :

1. Primary Outcome: Genomic and Metabolomic Characterization of Lean Diabetes Identification of an integrated biological phenotype distinguishing lean type 2 diabetes from healthy lean individuals and non-lean T2D (Genomic signatures, Metabolomic signatures, β-cell function, body fat compartments)

2. Secondary Outcome: Pancreatic and Body Composition Phenotype Quantification of pancreatic fat fraction, pancreatic volume, and hepatic fat content using CSE-MRI (PDFF) protocol. Assessment of bodycomposition parameters including percent body fat, fat mass, truncal body fat, fat-free mass, and abdominal fat depot distribution Methodology

   Describe the research methods that could best achieve the study objectives under the following sub-headings (as applicable):

   Assessment of pancreatic β-cell function and insulin secretory reserve using fasting and stimulated insulin and C-peptideconcentrations obtained during a standardized mixed meal tolerance test (MMTT), together with HOMA-IR, HOMA-β, Δ insulin,Δ C-peptide and stimulated-to-fasting response ratios.

3. Tertiary Outcome: Clinical Progression and Metabolic Characteristics at 1-Year Follow-up among individuals with lean diabetes Longitudinal assessment of glycemic control (HbA1c, fasting and postprandial glucose), anthropometric changes (BMI, waistcircumference, body composition), and progression of diabetes-related complications (retinopathy, neuropathy, nephropathy,cardiovascular and cerebrovascular disease) Evaluation of treatment requirements and therapeutic response, including insulin dependency and oral antidiabetic medication needs Comparison of nutritional status, lifestyle factors (diet, physical activity), psychological profile (mental health screening), and complicationburden between baseline and 1-year follow-up.

Sample Size estimation and sampling strategy :

The sample size has been calculated for comparison among three independent groups using one-way ANOVA. Based on a medium-to-large standardized effect size (Cohen's f = 0.30), supported by the substantial between-group differences in β-cell function and insulinsensitivity observed by Lontchi-Yimagou et al. (Diabetes Care, 2022)8, the minimum required sample size was estimated to be 111participants (37 participants per group) to achieve 80% power at a two-sided significance level of 5%. To account for potentialattrition, improve the precision of estimates, and enhance the robustness of imaging, genomic, metabolomic and longitudinalanalyses, the lean diabetes group (phenotype of interest) will be intentionally oversampled, while maintaining adequately poweredcomparator groups. Accordingly, a total of 151 participants will be recruited, comprising 75 participants with lean diabetes, 38 healthylean participants without diabetes, and 38 participants with non-lean T2D. This design provides sufficient power for between-groupcomparisons while maximizing information from the lean diabetes group.

Data Collection methods :

The study design shall include pre-screening interview, screening tests followed visits at Diabetes Foundation (India) and Dr. Kanungo'sDiabetes Centre, Bhubaneswar. Pre-screening shall comprise of explaining subjects about the study procedure, obtaining written consent.

Visit 1 at Baseline:

1. Sociodemographic and Lifestyle Assessment using Questionnaire

1. Age, sex, duration of diabetes, family history, intercurrent illness, socio-economic stratum (Annexure 1) 2. Birth weight (Annexure 1) 3. History of malnutrition in childhood (Annexure 1) 4. Nutrition profile in detail (Annexure 1) 5. Physical activity (Annexure 1) 6. Use of GLIM (Global Leadership Initiative on Malnutrition) criteria if suspected to be having malnutrition (Annexure 2) 7. Mental Health screening using a Patient health questionnaire (Annexure 3) 2.

General physical examination and anthropometric measurements:

1. Presence of acanthosis nigricans, arcus, xanthelasma, hepatomegaly. 2. Blood pressure (mmHg), 3. Height (cm), weight (kg), BMI (kg/m2), waist circumference (cm), Hip circumference calculates waist-height ratio.

4. Other circumferences (cm): Mid-arm circumference, mid-thigh circumference (cm), seven skinfold measurements - biceps, triceps,thigh, calf, subscapular, suprailiac (vertical and diagonal) and thoracic (measured in mm).

3. Biochemical Assessments

1. Glycemic status

: Fasting blood glucose, 2-hour postprandial blood glucose, and glycosylated hemoglobin (HbA1C) 2. Hematological assessments

: Complete blood count (CBC/Hemogram) 3. Lipid profile

• Total cholesterol, triglycerides, HDL, LDL and VLDL 4. Liver Function Test (LFT)
• SGOT (AST), SGPT (ALT), GGTP, serum albumin, and total protein 5. Kidney Function Test (KFT)
• Serum creatinine, Microalbuminuria, ACR. 4. Autoantibodies and tests related to insulin

C-peptide (fasting, stimulated) 2. Serum insulin (fasting, stimulated) 3. Anti-GAD65, IA-2, ZnT8 antibodies 5. Insulin Resistance and β-cell Function Indices (calculated)

: A standardized mixed meal tolerance test (MMTT) will beperformed in all three study groups. Blood samples will be collected in the fasting state and 2 hours after administration of astandardized mixed meal for measurement of plasma glucose, serum insulin and C-peptide. HOMA-IR and HOMA-β will becalculated from fasting measurements. Dynamic β-cell responses will be assessed using stimulated insulin and C-peptideconcentrations, Δ insulin, Δ C-peptide and stimulated-to-fasting response ratios.

6.

Complication Assessment:

1. Fundus Evaluation for Assessment of Diabetic Retinopathy For the assessment, an ophthalmologist performed dilated fundusexamination which is the gold standard. It can be categorized into mild Non-Proliferative Diabetic Retinopathy (NPDR) (early-stage Methodology

Describe the research methods that could best achieve the study objectives under the following sub-headings (as applicable):

damage to retinal blood vessels and few microaneurysms without major vision impairment), moderate NPDR (more significantdamage with intraretinal microvascular abnormalities, venous beading in two quadrants, or intraretinal hemorrhages in all fourquadrants causing potential for vision loss), Proliferative Diabetic Retinopathy (PDR) (Vitreous hemorrhage or neovascularisation) orMacular Edema (ME) (severe form of diabetic retinopathy that causes swelling in the macula, leading to impaired vision).

2.

Peripheral Neuropathy:

Neuropathy refers to damage to the nerves often leading to symptoms such as pain, tingling, or loss ofsensation, identified via vibration perception threshold (VPT) using a biothesiometer or monofilament loss, a quantitative andreproducible method to assess peripheral sensory loss and identify patients at risk for foot complications. Neuropathy can rangefrom mild (early symptoms, such as occasional numbness or tingling, reduced sensation at 1-2 sites) to moderate (more persistentpain or numbness, reduced sensation at 3-4 sites, >15 volts) to severe (loss of sensation, muscle weakness, or significantimpairments in daily activities, reduced sensation at >4 sites, >25 volts).

3.

Nephropathy:

Nephropathy refers to kidney damage that often progresses to End-Stage Renal Disease (ESRD) if not managedappropriately. The markers of chronic kidney disease (CKD) in this study included micro/macroalbuminuria assessed by spot urine(elevated urine albumin levels, >30mg/g) or Albumin-to-Creatinine Ratio (ACR) using standard immunoassay-based laboratorymethods (presence of albumin in urine relative to creatinine levels, >30mg/g), increased serum creatinine (decrease in kidneyfunction, men >1.3 mg/dL and women >1.1 mg/d), decreased eGFR using the CKD-EPI equation (reduction in glomerular filtrationrate, <60 mL/min/1.73 m² for ≥3 months).

4.

Peripheral Artery Disease (PAD):

As indicated by an Ankle-Brachial Index (ABI) with values <0.9 signifying arterial obstruction.

5.

Cardiovascular Diseases (CVD):

Two-dimensional transthoracic echocardiography will be performed using the Siemens ACUSONX300 ultrasound system with a phased-array transducer. The examination will be conducted in the left lateral decubitus position bya trained operator, obtaining standard views to assess cardiac structure and function including chamber dimensions, ejectionfraction, and valvular morphology.

7. Handgrip Strength, Body Composition and Imaging

:

1.

1. Handgrip strength will be measured using Jamar handgrip dynamometer 2. Body composition will be done by dual-energy X-ray absorptiometry (DEXA) scan using LUNAR Prodigy (Advance model noPA+350752, GE Healthcare) for Fat mass, Fat free mass and Muscle mass.

3.

MRI based quantification:

MRI-based detailed body fat analysis, including pancreatic fat, hepatic fat and abdominal fatcompartments, will be performed in all participants with lean diabetes and 50% of participants in each comparatorgroup. Participants selected for MRI in the comparator groups will be selected using computer-generated randomsampling after enrolment, independent of demographic or clinical characteristics.

Fat fractions in the liver and pancreas were quantified as proton density fat fraction (PDFF) using multi-echo chemical-shift encodedmagnetic resonance imaging (CSE-MRI): Single-breath-hold multipoint Dixon-based acquisition, IDEAL IQ sequences will be performed by3-T MRI (GE Healthcare). Nine regions of interests (6 in the right lobe and 3 in the left lobe) for liver and three ROIs for pancreas (oneeach in head, body and tail regions) Pancreatic volume was derived by three-dimensional liver acquisition with volume acceleration (3D LAVA) sequence. A pancreas volumeindex, defined as the pancreatic volume (cm3)/ body surface area divided by the body surface area (m2) calculated by the method ofDubois and Dubois will be used to examine possible effects of different anthropometry on observed measurements. The ratio of thepancreas volume to the waist circumference (in cm3/cm) will also be calculated to evaluate the data corrected for potential excess visceralfat in subjects We quantified cross sectional areas of following body fat compartments by MRI at lumbar vertebra 3-4; abdominal adipose tissue (AT),subcutaneous AT (SCAT), intra-abdominal AT (IAAT), superficial abdominal SCAT, deep abdominal SCAT, anterior abdominal SCAT,posterior abdominal SCAT, intra-peritoneal AT and retroperitoneal AT (Annexure 4, Figure 3). A single radiologist who is trained in these methods of MRI analysis will perform the measurements. The radiologist will be blinded to clinicaldata 1. Genomic Analysis 1. MODY (to be clinically excluded, genetic analysis to be done if clinically indicated) : GCK (glucose sensor), HNF1α/HNF4A/HNF1B (β-cell transcription), PDX1, ABCC8/KCNJ11 (KATP channel), NEUROD1 2.

Lipodystrophy (to be clinically excluded, genetic analysis to be done if clinically indicated):

AGPAT2, BSCL2 (CGL), LMNA,PPARG (FPLD), INSR, ZMPSTE24 3. Mitochondrial Diabetes (MIDD ) (to be clinically excluded, genetic analysis to be done if clinically indicated): A3243G 4. Genetic analysis i n Unclassified People with Diabetes and BMI <19 kg/m2 (main interest of this investigation ): 1. Candidate genes: KCNJ11, PPARG, TCF7L2, KCNQ1, CDKAL1, SLC30A8, LAMA1 gene encoding laminin-1, KCNJ15, ZRANB3,CDKN2BAS, CDC123/CAMK1D, HHEX, APOC3 promoter variants, Gly972Arg polymorphism of IRS1 gene (Genome wideanalysis) 2.

Whole exome sequencing (WES):

It will be performed to analyze the protein-coding regions of clinically relevant genes.Genomic DNA will be extracted from patient specimens, and the targeted regions will be captured using hybridization probes.Libraries will be prepared with the Twist Exome 2.0 kit and sequenced on the Illumina NovaSeq 6000 platform using 2×150 bppaired-end reads. Quality control will ensure >95% of target bases at >20× coverage with a mean coverage of 80-100X. Thesequence data will be processed using the in-house MMG_WES_Ver_2.0 pipeline, which will include quality checks with FastQC(v0.12.1) and Trimmomatic (v0.39), alignment to the hg19/GRCh37 reference genome using BWA (v0.7.17), variant calling withPicard (v3.0.0) and GATK (v4.4.0.0), and annotation with VEP against Ensembl release 109, following Broad Institute's Best Methodology

Describe the research methods that could best achieve the study objectives under the following sub-headings (as applicable):

Practices. CNVs will be detected with ExomeDepth. Variants will be interpreted using ClinVar, OMIM, HGMD, Orphanet, ACMGguidelines, and family history, with low-quality calls validated orthogonally (Annexure 4, Figure 4) 2. Metabolomics Analysis: It will be performed in all participants with lean diabetes and 50% of participants in each comparatorgroup. Participants selected for metabolomic analysis in the comparator groups will be selected using computer-generatedrandom sampling after enrolment, independent of demographic and clinical characteristics.

1. Branched-chain amino acids (BCAAs): Leucine, Isoleucine, Valine 2. Ceramides (c16:0, c18:0, c24:1) 3. Acylcarnitines (c0, c2, c3) Targeted and untargeted metabolomic profiling (as appropriate) will be performed using fasting plasma samples collected in EDTA tubes,centrifuged within 30 minutes and stored at -80 °C until analysis. Branched-chain amino acids (BCAAs), including leucine, isoleucine, and valine,which are polar metabolites, will be assessed from fasting plasma following methanol-based protein precipitation and analyzed using liquidchromatography-tandem mass spectrometry (LC-MS/MS) with a hydrophilic interaction chromatography (HILIC) column to evaluate theirassociation with insulin secretion defects, altered amino-acid metabolism, and early metabolic dysregulation. Ceramide species including c16:0(palmitoyl-ceramide), c18:0 (stearoyl-ceramide), and c24:1 (nervonoyl-ceramide), representing non-polar lipids, will be measured from theplasma lipid fraction and quantified using LC-MS/MS-based lipidomics with a reverse-phase C18 column, reflecting lipotoxicity, impaired insulinsignaling, and β-cell dysfunction. Additionally, short-chain acylcarnitines including c0 (free carnitine), c2 (acetylcarnitine), and c3(propionylcarnitine), classified as semi-polar metabolites, will be extracted using methanol precipitation with stable isotope internal standardsand analyzed using LC-MS/MS to assess carbohydrate metabolism and mitochondrial acetyl-CoA production, thereby providing insights intomitochondrial metabolic function in lean diabetes.

Visit 2 (6 months follow-up in lean diabetes):

A. General physical examination, anthropometric measurements and biochemical assessments (as appropriate) 1.

Evaluation of treatment (if any):

Requirements and therapeutic response, including insulin dependency and oral antidiabetic medicationneeds

Visit 3 (1 year follow-up in lean diabetes):

The following assessments will be performed at this visit, as described above, as appropriate.

1. General physical examination and anthropometric measurements 2. Biochemical Assessments 3. Tests related to insulin 4. Insulin Resistance and β-cell Function Indices 5. Complication Assessment 6. Handgrip Strengh, Body Composition and Imaging 7. Metabolomic Analysis 8. Evaluation of treatment (if any) The study schedule for the measurements, assessments, and investigations to be performed at the baseline visit, 6-months follow-up, and 1-year follow-up is tabulated (Annexure 4, Table 2)

Data quality assurance methods :

Data quality assurance methods will be uniformly implemented at both participating study sites.

Quality Assurance and Data Management Strategy Maintaining high standards of data quality and reliability is essential for this study, which involves clinical phenotyping, biochemicalinvestigations, genomic and metabolomic analyses, and advanced imaging assessments.

Standardized Clinical Assessments All clinical and anthropometric measurements will be conducted according to uniform standard operating procedures (SOPs). Measurementsincluding body weight, height, waist circumference, mid-arm circumference, and neck circumference will be obtained using calibratedequipment. Blood pressure will be measured using validated automated sphygmomanometers, and duplicate readings will be recorded. Fastingblood samples will be collected following standardized phlebotomy protocols after overnight fasting.

Imaging Quality Control Magnetic resonance imaging (MRI) will be used to evaluate hepatic, pancreatic fat fraction, pancreatic volume and abdominal adipose tissuedistribution. Imaging will be performed using standardized acquisition protocols, and analysis will be carried out by trained radiologists. Asubset of imaging studies will undergo repeat analysis to ensure consistency and intra-observer reliability.

Metabolomic Analysis Quality Assurance Methodology

Describe the research methods that could best achieve the study objectives under the following sub-headings (as applicable):

Targeted metabolomic profiling, including assessment of branched-chain amino acids, ceramides, and acylcarnitines, will be performed usingliquid chromatography-tandem mass spectrometry (LC-MS/MS) with validated laboratory procedures. Internal standards and calibration curveswill be used to ensure accurate quantification. Pooled quality control samples will be analyzed intermittently during analytical runs to monitoranalytical stability. A coefficient of variation (CV) of less than 15% across quality control samples will be considered acceptable.

Laboratory Quality Management Biochemical analyses will be performed in ISO-certified laboratories following established laboratory quality assurance protocols. Internalquality control samples will be run regularly, and participation in external quality assurance programs will be ensured where applicable. Allbiological samples will be processed within predefined time frames and stored at -80°C until analysis to maintain sample stability.

Study Monitoring and Data Oversight All study procedures will be conducted by trained research personnel under the supervision of the principal investigator. Periodic monitoringand review of study activities will be performed to ensure adherence to protocols and to maintain consistency in data collection procedures.

Data Validation and Verification Data will be entered into a secure electronic database using standardized data entry formats. Critical variables will undergo double data entryto minimize transcription errors. Routine data checks will be performed to identify missing values, outliers, and inconsistencies, anddiscrepancies will be resolved by cross-verification with original source documents.

Data management plan :

A structured data management system will be implemented to ensure accurate collection, storage, security, and analysis of study data. The datamanagement procedures described below will be uniformly applied at both participating study sites to maintain consistency and reliability.

Data Collection All study data, including clinical assessments, laboratory investigations, imaging findings, genomic and metabolomic results, will be recordedusing case record proforma. Trained research staff will collect data according to predefined protocols to ensure completeness and consistency.

Data Entry and Storage Data from proforma will be entered into a secure electronic database with restricted access. Each participant will be assigned a unique studyidentification number, and personal identifiers will be stored separately from research data to maintain confidentiality. Electronic data will bestored on password-protected systems with regular backups to prevent data loss.

Data Quality Control To ensure accuracy, double data entry and periodic data validation checks will be conducted for key variables. Automated validation rules willbe applied to identify missing data, inconsistencies, and outliers. Any discrepancies will be verified against original source documents andcorrected as necessary.

Data Confidentiality and Security Participant confidentiality will be strictly maintained throughout the study. Only authorized members of the research team will have access toidentifiable data. All electronic files will be stored on secure servers, and physical documents will be kept in locked cabinets within restricted-access areas.

Data Analysis and Reporting Cleaned and validated datasets will be used for statistical analysis using appropriate software. Data will be analyzed in accordance with thepredefined statistical analysis plan, and results will be reported in aggregated form without revealing participant identities.