Abstract

Historical descriptions of malnutrition-related diabetes mellitus (MRDM) have regained attention in contemporary discourse, with the term type 5 diabetes mellitus (T5DM) increasingly used as a harmonized research construct to describe insulin-deficient diabetes associated with early-life undernutrition. Although the International Classification of Diseases, 11th Revision (ICD-11) includes categories related to malnutrition-associated diabetes, the pathophysiological interpretation, diagnostic boundaries, and therapeutic implications of T5DM remain incompletely defined, resulting in variable clinical adoption. Consequently, many affected individuals continue to be classified as lean type 2 diabetes and managed within obesity-centric care paradigms. This narrative review synthesizes existing evidence to advance a conceptual framework that distinguishes T5DM from other lean diabetes phenotypes by integrating developmental nutritional reserve, β-cell functional capacity, and autoimmune status. Clinical observations are highlighted to indicate that body size–based descriptors alone may not adequately capture clinically relevant nutritional heterogeneity. Emerging pharmacological evidence is also reviewed, suggesting that responses to glucagon-like peptide-1 receptor agonists, sodium–glucose cotransporter-2 inhibitors, and insulin secretagogues may not be determined solely by residual β-cell functional reserve, but may also be modulated by underlying nutritional status. Building on this premise, a nutrition-integrated pharmacometabolic model is proposed as a hypothesis-generating framework, in which standardized nutritional assessment is considered alongside β-cell evaluation to inform exploratory treatment stratification rather than prescriptive clinical guidance. Finally, priorities for translational research are outlined, including the development of consensus-based indicators of nutritional reserve, improved phenotypic classification frameworks, and nutritional reserve–stratified randomized trials. Conceptualizing T5DM as a nutritionally conditioned metabolic phenotype may facilitate more targeted research agendas and support context-sensitive, individualized interventions in populations where malnutrition and diabetes intersect.

1 Introduction

Increasing recognition that diabetes comprises a spectrum of heterogeneous metabolic phenotypes, rather than a single glycemic disorder, has highlighted important limitations of traditional binary classification frameworks (1). Although advances in precision endocrinology have motivated more nuanced taxonomies, contemporary classification systems remain largely structured around obesity-associated insulin resistance (IR) and autoimmune-mediated β-cell destruction (2). In contrast, diabetes presentations associated with early-life undernutrition and adverse developmental environments remain under-represented within prevailing clinical and research paradigms (3).

The long-standing dichotomy separating type 1 diabetes mellitus (T1DM) from type 2 diabetes mellitus (T2DM) does not fully capture the metabolic diversity observed across global populations (1, 2). This limitation is particularly evident in undernourished settings, where relative β-cell insufficiency has been reported in the absence of obesity or overt autoimmunity (3, 4). Recognition of such heterogeneity has historically prompted efforts to refine classification beyond the conventional T1DM–T2DM framework, including latent autoimmune diabetes in adults (LADA), maturity-onset diabetes of the young (MODY), and diabetes secondary to systemic disease or pharmacological exposure (2).

Historical descriptions of malnutrition-related diabetes mellitus (MRDM) characterized predominantly young individuals with low BMI who presented with insulin-requiring hyperglycemia, relatively low reported propensity for ketosis, and limited evidence of autoimmunity—features that did not align neatly with classical T1DM or T2DM definitions (3, 5). Rather than introducing a novel nosological entity, the contemporary term “type 5 diabetes mellitus” (T5DM) has been proposed as a harmonized research construct intended to integrate these historical observations with emerging data on insulin-deficient diabetes in nutritionally vulnerable populations (3, 5, 6). In this manuscript, historical descriptions are referred to as malnutrition-related diabetes mellitus (MRDM), recent consensus terminology as undernutrition-associated diabetes, and the term “T5DM” is used as a harmonized research construct unless otherwise specified.

Within this conceptual framework, early-life nutritional deprivation is hypothesized to constrain β-cell developmental capacity and reduce functional reserve; however, this interpretation remains provisional and requires further empirical validation. The notion of “nutritional imprinting,” illustrated schematically in Figure 1A, is therefore presented as a heuristic model rather than a definitive mechanistic explanation. A comparative schematic contrasting autoimmune β-cell destruction (T1DM), insulin resistance–dominant metabolic overload (T2DM), and nutrition-conditioned limitation of β-cell reserve (proposed T5DM) is shown in Figure 1B, highlighting differences in upstream drivers despite convergence on hyperglycemia.

(A) Conceptual framework linking early-life undernutrition to β-cell–limited diabetes phenotypes. This schematic depicts a hypothesized sequence whereby early-life undernutrition may imprint pancreatic β-cell development, resulting in a constrained β-cell reserve and a lean, metabolically fragile phenotype. Under subsequent metabolic stressors (e.g., intercurrent illness, infection, or nutritional transition), limited β-cell reserve may manifest as insulin-requiring hyperglycemia, often with a lower propensity for ketosis in reported cohorts. This conceptual framework highlights nutrition-conditioned β-cell vulnerability and distinguishes it from autoimmune β-cell destruction in T1DM and insulin resistance–dominant mechanisms in obesity-associated T2DM. (B) Comparative schematic of metabolic pathways leading to hyperglycemia in T1DM, T2DM, and T5DM. This schematic contrasts three conceptual trajectories that converge on hyperglycemia but differ in upstream drivers. In T1DM, autoimmune β-cell attack leads to β-cell loss and severe insulin deficiency. In obesity-associated T2DM, insulin resistance predominates, accompanied by β-cell stress and progressive β-cell failure. In the proposed T5DM framework, early-life undernutrition is hypothesized to imprint β-cell developmental capacity, resulting in limited β-cell reserve and constrained insulin secretion, with relatively preserved insulin sensitivity.

Observational studies and experimental models suggest that sustained protein–energy deficiency during critical developmental windows may be associated with reduced β-cell mass and diminished insulin secretory capacity, creating a metabolically vulnerable endocrine state (6, 7). Some reports indicate that such impairments may persist despite later nutritional improvement, although the magnitude and clinical relevance of this persistence remain incompletely defined (6, 7). In contrast to obesity-associated T2DM—where IR driven by adiposity-related inflammation predominates—presentations consistent with the proposed T5DM framework are characterized by relative β-cell insufficiency with comparatively preserved insulin sensitivity (2, 8). Direct human evidence demonstrating a permanent structural β-cell deficit attributable solely to early-life undernutrition, however, remains limited.

Conventional anthropometric indices have limited sensitivity for detecting clinically meaningful nutritional vulnerability in low-BMI diabetes presentations (9–11). Accumulating evidence further indicates that impaired nutritional status is associated with increased complication burden and mortality among individuals with diabetes, underscoring the prognostic relevance of nutritional factors not captured by body size alone (12–15).

Despite growing recognition of phenotypic heterogeneity, current proposals related to T5DM remain largely conceptual, and standardized operational diagnostic criteria have not been universally established (2, 3, 5). In routine practice, classification frequently relies on readily available anthropometric descriptors, which may inadequately reflect endocrine–nutritional vulnerability in insulin-deficient, nutritionally conditioned presentations (2, 6). Consequently, such phenotypes may remain under-recognized, limiting opportunities for earlier identification of individuals at risk of accelerated β-cell functional deterioration (4–6). This review therefore synthesizes current evidence related to T5DM, integrating epidemiological observations, diagnostic challenges, and nutritional pathophysiological insights to provide a structured research-oriented framework rather than formal diagnostic or therapeutic guidance (3–5).

This manuscript is a narrative review. Literature was identified through structured searches of major biomedical databases, including PubMed and Embase, over the past decade using combinations of terms related to MRDM, undernutrition-associated diabetes, β-cell function, and diabetes classification. Reference lists of key reviews and consensus statements were manually screened for additional relevant studies. Given the heterogeneity of available evidence, formal quantitative synthesis and uniform risk-of-bias assessment were not undertaken, and findings are interpreted in a conceptual, hypothesis-generating context.

2 Definition and characteristics of T5DM2.1 Historical evolution: J-type diabetes → MRDM → ICD-11 malnutrition-associated diabetes → current type 5 construct

The phenotype currently discussed as type 5 diabetes mellitus (T5DM) is best understood as a contemporary research construct that consolidates multiple historical descriptions of insulin-deficient diabetes occurring in lean individuals with exposure to undernutrition (1, 5, 6). Early reports from the mid-20th century, including Hugh-Jones’ report of “J-type” diabetes in Jamaica (1955), which described young, lean patients presenting with marked hyperglycemia, insulin requirement, and a reportedly lower propensity for ketosis than classic autoimmune type 1 diabetes (16). Subsequent reports across low- and middle-income settings described similar phenotypes variably labelled as tropical diabetes or malnutrition-related diabetes mellitus (MRDM) (17, 18). The key historical milestones are summarized in Figure 2.

Historical classification trajectory of undernutrition-associated diabetes. The timeline highlights major milestones from the initial description of “J-type” diabetes (1955), through WHO classification discussions of MRDM (1980s–1990s) and its subsequent de-emphasis (1998/1999), to ICD-11 coding (2019) and the 2025 international consensus defining the contemporary Type 5 construct.

In WHO classification discussions from the mid-1980s to the late 1990s, MRDM was incorporated as a proposed category; however, it remained controversial because operational boundaries were inconsistent and overlap with other entities (autoimmune diabetes, pancreatogenic diabetes, and lean presentations of type 2 diabetes) was substantial (19). This reappraisal coincided with increasing availability of diagnostic tools for autoimmunity, pancreatic disease, and monogenic diabetes, and MRDM was progressively deprioritized in later classification frameworks, reflecting concerns about reproducibility rather than a definitive refutation of the underlying phenotype (20).

In parallel, the International Classification of Diseases, 11th Revision (ICD-11) introduced coding categories relevant to malnutrition-associated diabetes, providing a pragmatic administrative scaffold but not resolving the biological or clinical boundaries of the condition (21). More recently (late 2010s–2025), renewed attention has been driven by contemporary cohorts describing atypical, lean, insulin-deficient diabetes and by international consensus efforts proposing undernutrition-associated diabetes as a distinct research focus (1, 4). In this context, the term “T5DM” functions as a unifying research label of phenotyping (nutritional reserve, β-cell reserve, and autoimmunity) and to reduce misclassification as lean type 2 diabetes (1, 5).

2.2 Contemporary definition and diagnostic boundaries

Although the historical label of malnutrition-related diabetes mellitus (MRDM) has been repeatedly invoked across decades, its diagnostic reproducibility has remained limited. A major challenge is the substantial phenotypic overlap with other forms of lean diabetes, including latent autoimmune diabetes in adults (LADA), pancreatogenic diabetes (type 3c), and insulin-deficient presentations of type 2 diabetes, particularly in settings where autoantibody testing, pancreatic imaging, and genetic evaluation are not routinely available (3–5).

Importantly, reliance on low BMI as a defining feature lacks specificity, as constitutional leanness, sarcopenia, and catabolic weight loss during uncontrolled diabetes may produce similar anthropometric profiles. Accordingly, anthropometry-based criteria alone are insufficient to distinguish nutritional imprinting from non-nutritional etiologies of insulin deficiency (3, 11).

Additional challenges arose from substantial phenotypic overlap with other lean diabetes entities, including LADA, pancreatogenic diabetes, and lean presentations of T2DM, particularly in the absence of validated markers capable of reliably differentiating nutritional reserve and pancreatic β-cell functional capacity (3–5).

Contemporary proposals framing T5DM as a provisional research construct therefore emphasize a multidimensional definition incorporating evidence of undernutrition exposure or reduced nutritional reserve, impaired β-cell secretory capacity, and exclusion of autoimmune diabetes. However, validated biomarkers and consensus operational criteria remain incomplete, reinforcing that T5DM should currently be regarded as a hypothesis-generating phenotype rather than an established clinical diagnosis (1, 5, 6).

2.3 Clinical manifestations

Within current conceptual frameworks, T5DM is described as a low-BMI, insulin-deficient presentation frequently associated with impaired endogenous insulin secretory capacity in the absence of islet autoimmunity (3, 4). Compared with classical T1DM, ketosis may be less frequent in some reported cohorts; however, this feature appears variable and should not be considered a defining characteristic (5, 6).

Nutritional depletion has been reported in individuals meeting proposed T5DM characteristics and may manifest as reduced muscle mass, unintentional weight loss, and micronutrient inadequacy, contributing to metabolic instability (9–11). Pancreatic exocrine insufficiency and gastrointestinal manifestations—such as early satiety, bloating, nausea, and delayed gastric emptying—have also been documented in subsets of patients with diabetes and may further impair nutrient intake and absorption, potentially exacerbating glycemic variability and protein–energy deficiency (22–26). In this context, structured nutritional assessment tools, including the Patient-Generated Subjective Global Assessment (PG-SGA), have been used to characterize nutritional status, with evidence suggesting that nutritional deficits may worsen alongside increasing metabolic stress and reduced physiological reserve (27).

Collectively, T5DM may be conceptualized as an integrated clinical phenotype in which lean body habitus alone is insufficient for accurate classification (1, 2). Recognition therefore often benefits from consideration of nutritional status, β-cell functional reserve, and careful differentiation from autoimmune or monogenic forms of diabetes, as reliance on glycemic indices or body size–based descriptors alone may increase the risk of misclassification as lean T2DM or other atypical presentations (1–4). These features, summarized across historical and contemporary cohorts, are provided in Table 1, highlighting the consistent clinical pattern of T5DM over time. Notably, systematic longitudinal outcome data were not reported in these cohorts, highlighting a persistent evidence gap across decades.

Study [Ref]/country/yearAge at diagnosisBMIInsulin dependence at presentationKetosisC-peptideAutoantibodiesExocrineOutcomeHugh-Jones (16); Jamaica; 1955YoungLeanRequiredLow/rareNANANRNo longitudinal dataAbu-Bakare (17); Tropics; 1986YoungLeanVariableVariableNRNRNRNo longitudinal dataChattopadhyay (18); India; 1995YoungLeanRequiredLow/rareReducedNegativeMild EPINo longitudinal outcomesLontchi-Yimagou (4); Multi; 2022Young–midLeanRequiredLow/rareReducedNegativeNRCross-sectionalSiddiqui (90); India; 2022YoungLean/normalVariableNRReducedNegativeNRCross-sectionalKibirige (91); SSA; 2024AdultUnderweightNRNRReducedNegativeNRCross-sectional

Clinical phenotype associated with T5DM across historical and contemporary cohorts (1955–2024).

NA, not available due to historical or technical limitations; NR, not reported in the original publication; EPI, exocrine pancreatic insufficiency. C-peptide values are shown where reported; early MRDM cohorts predated standardized assays and are therefore indicated as NA rather than NR. Ketosis reflects qualitative descriptions in the original reports, as quantitative rates were infrequently provided. Longitudinal outcome data were not systematically reported in historical MRDM cohorts, and most contemporary phenotype-consistent studies were cross-sectional in design. Historical MRDM cohorts and contemporary phenotype-consistent insulin-deficient, non-autoimmune, low-BMI diabetes populations are included as representative of the clinical phenotype associated with the proposed construct of T5DM.

3 Diagnostic challenges3.1 Ambiguity of diagnostic criteria

Although T5DM has increasingly been discussed as a nutritionally conditioned research phenotype, its clinical identification remains constrained by the absence of validated and operational diagnostic criteria that can be consistently applied across settings (3, 5, 6). This ambiguity partly reflects the historical origins of the concept: early descriptions of MRDM were largely derived from region-specific clinical observations rather than prospectively defined diagnostic frameworks, contributing to heterogeneous interpretations of insulin requirement, ketosis propensity, and disease boundaries (3, 5, 6).

While these descriptive features facilitated early recognition, subsequent clinical and epidemiological experience has demonstrated limited diagnostic specificity in practice (3, 5). Substantial phenotypic overlap with other low-BMI or insulin-deficient diabetes presentations—including atypical T2DM—continues to challenge phenotype-based classification and supports development of more operational approaches integrating nutritional status and β-cell functional assessment (3, 4).

Continued reliance on overly simplified anthropometric proxies has limited clinical applicability in nutritionally heterogeneous populations, particularly in settings where constitutional leanness is common and historical MRDM criteria are less transferable to contemporary presentations.

The absence of standardized operational criteria and validated biomarkers also limits consistent characterization of T5DM (3). Pancreatic exocrine dysfunction, a plausible accompaniment of chronic undernutrition, has been documented using heterogeneous methods in studies of diabetes and nutritional deficiency, but has not yet been systematically integrated into contemporary diabetes phenotyping frameworks (7, 25, 26). Non-uniform thresholds also contribute to variability in prevalence estimates and risk stratification across studies (9, 28–30).

These limitations indicate that descriptive or anthropometry-based criteria are insufficient, motivating reproducible, multidimensional diagnostic approaches that more accurately reflect contemporary understanding of diabetes pathophysiology and nutritional–metabolic heterogeneity (1, 3, 5, 11, 15, 31).

Given the persistent heterogeneity and historical diagnostic ambiguity surrounding T5DM, the absence of reproducible operational criteria has limited cross-study comparability and cohort harmonization. To facilitate standardized case identification in observational and interventional research, provisional operational research criteria for suspected T5DM are outlined in Box 1. These criteria are intended to support structured phenotyping and hypothesis-driven investigation and should not be interpreted as formal diagnostic standards.

3.1.1 Box 1. provisional operational research framework for suspected T5DM

Scope: This framework is intended for research cohort identification and exploratory phenotyping. It does not constitute formal diagnostic criteria and should not replace established diabetes classification systems.

3.1.1.1 Core criteria (all required)

Confirmed diabetes mellitus according to ADA or WHO glycemic thresholds.

Low or low-normal BMI (ethnicity-adjusted): Asian populations <23 kg/m²; non-Asian populations <25 kg/m². A stricter threshold <21 kg/m² may be considered in high-undernutrition settings.

Evidence of nutritional vulnerability (≥1): documented childhood/adolescent undernutrition; nutritional risk by validated tools (e.g., CONUT ≥ mild risk; PG-SGA category B/C); or biochemical indicators of reduced nutritional reserve interpreted outside acute inflammatory states.

3.1.1.2 Supportive metabolic features (≥2 recommended)

Low–moderate fasting C-peptide (C-P) and/or reduced stimulated C-P response.

Negative islet autoantibodies (GAD65, IA-2, ZnT8).

No overt clinical features suggestive of insulin resistance, including marked central adiposity, acanthosis nigricans, or clinically apparent non-alcoholic fatty liver disease (NAFLD).

Low-BMI presentation without prior obesity history.

3.1.1.3 Exclusion criteria (any excludes classification)

Chronic pancreatitis or fibrocalculous pancreatic diabetes.

Established pancreatogenic (type 3c) diabetes.

Suspected monogenic diabetes (e.g., early onset with autosomal dominant inheritance pattern).

Chronic systemic glucocorticoid exposure.

Classical autoimmune T1DM.

3.1.2 Operational research classification

Individuals meeting all core criteria, ≥2 supportive metabolic features, and no exclusion criteria may be operationally classified as ‘suspected T5DM (research phenotype)’ for cohort enrichment and mechanistic investigation.

3.1.2.1 Implementation considerations

Interpret nutritional indices in clinical context, particularly in inflammatory, renal, or hepatic conditions.

Stimulated C-P testing is preferred where feasible.

Threshold calibration may require adaptation across populations undergoing nutritional transition.

3.2 Preliminary evaluation framework for suspected T5DM

Although a formal diagnostic algorithm for T5DM has not been established, a pragmatic, conceptually oriented evaluation framework may be cautiously inferred from existing evidence and expert discussions to support exploratory case identification in research settings (1, 3). This framework is not intended for routine diagnosis or clinical management and does not replace formal diagnostic criteria; rather, it aims to reduce misclassification (e.g., as low-BMI T2DM) and to facilitate earlier phenotype-oriented stratification. Figure 3 outlines this exploratory identification framework, illustrating a stepwise approach integrating nutritional assessment, autoantibody evaluation, and β-cell functional testing.

Exploratory identification framework for T5DM. Patients presenting with diabetes in a lean or atypical phenotype undergo stepwise evaluation to distinguish nutritional vulnerability from constitutional leanness. Following identification of nutritional impairment, negative islet autoantibody testing helps exclude autoimmune β-cell destruction characteristic of T1DM or LADA. Subsequent assessment of β-cell reserve and IR facilitates differentiation from IR-dominant lean T2DM. Supportive indicators—including pancreatic exocrine dysfunction, pancreatic imaging features, and family history with targeted screening for MODY—may assist in refining phenotypic classification. Individuals demonstrating nutritional impairment, evidence of reduced β-cell reserve, and absence of marked IR are considered consistent with the proposed T5DM research framework, underscoring the importance of nutrition-integrated diagnostic approaches.

In individuals with diabetes and low-to-normal BMI, structured nutritional assessment represents an important evaluative domain and may be explored using validated tools such as the Controlling Nutritional Status (CONUT) score, the PG-SGA, or selected biochemical markers (e.g., serum prealbumin), rather than reliance on screening based solely on body size or weight metrics (11, 27, 32, 33).

Assessment of autoimmune markers, including antibodies to glutamic acid decarboxylase 65, insulinoma-associated antigen 2, and zinc transporter 8, may further assist in differentiating T5DM from LADA or late-onset T1DM (2, 3, 30). Key operational discriminators between proposed T5DM and major alternative etiologies of low-BMI or insulin-deficient diabetes presentations are summarized in Table 2. Among islet autoantibody-negative individuals, β-cell functional reserve may be assessed using fasting and/or stimulated C-P measurements, providing an indication of endogenous insulin secretory capacity and supporting differentiation from alternative low-BMI diabetes presentations when interpreted alongside clinical features as needed (3, 8, 31, 34).

PhenotypeAge of onsetKetosis tendencyIslet autoantibodiesC-PIR featuresKey cluesProposed T5DMAdolescence–early adulthoodRareNegativeLow–moderate
(limited reserve)Often relatively preserved insulin sensitivityNutritional vulnerability;
Islet autoantibodies−;
C-P low–moderate;
exclude alternative etiologiesLean T2DMAdulthoodRareNegativePreserved early
(normal/high)IR despite low BMI
(NAFLD, acanthosis)Central adiposity/NAFLD;
Islet autoantibodies−;
C-P preserved earlyT1DMChildhood–adolescenceFrequentPositiveLowNot primary driverAutoimmune markers; higher DKA riskLADAAdulthoodOccasionalPositiveGradually decliningVariableAutoimmune markers; delayed insulin requirementMonogenic (MODY)Adolescence–early adulthoodRareNegativePreservedNot primary driverYoung onset;
AD family history;
Islet autoantibodies−;
consider geneticsPancreatogenic (type 3c/FCPC)VariableVariableNegativeReducedNot primary driverPancreatitis/FCPC;
steatorrhea;
low fecal elastase;
abnormal imaging

High-yield differential diagnosis for lean diabetes phenotypes relevant to the proposed T5DM framework.

C-P, C-peptide; IR, insulin resistance; NAFLD, non-alcoholic fatty liver disease; AD, autosomal dominant; FCPC, fibrocalculous pancreatic diabetes; LADA, latent autoimmune diabetes in adults; T1DM, type 1 diabetes mellitus; T2DM, type 2 diabetes mellitus. C-P categories reflect relative endogenous insulin secretory capacity (stimulated testing preferred where feasible), and IR features denote clinical surrogates such as central adiposity, acanthosis nigricans, or NAFLD. This table supports exploratory differential evaluation of lean or insulin-deficient diabetes phenotypes; the proposed T5DM row represents a provisional research construct and does not constitute formal diagnostic criteria.

In selected clinical contexts, such as early disease onset, strong familial clustering, or atypical clinical features, targeted genetic testing may provide additional information to support evaluation for MODY (2, 34). Within this research-oriented framework, priority is given to assessment of nutritional status, evaluation of β-cell functional reserve, and exclusion of alternative etiologies; the co-occurrence of nutritional impairment and reduced β-cell reserve in the absence of islet autoimmunity represents a pattern central to the proposed conceptualization of T5DM.

Nutritional impairment in this framework is not defined by a single biomarker. Screening tools (e.g., CONUT, PG-SGA) and selected laboratory indices should be interpreted in clinical context, particularly under conditions of inflammation, infection, hepatic dysfunction, or renal disease. In low-resource settings, pragmatic combinations of clinical history (dietary scarcity, unintentional weight loss), feasible anthropometry/body-composition surrogates, and basic laboratories may be the most implementable approach.

3.3 Limitations of low BMI

BMI has historically served as a pragmatic epidemiological surrogate in contexts where overt undernutrition was prevalent; however, its discriminatory value for identifying T5DM is limited in contemporary settings. Low BMI does not reliably distinguish nutrition-conditioned β-cell vulnerability from constitutional leanness and may also reflect genetic background, socioeconomic factors, or secondary catabolic stress during disease progression, thereby reducing its utility as an antecedent or defining marker (3, 4, 35–37).

Beyond its limited phenotypic specificity, BMI provides little insight into β-cell functional reserve, as reflected by C-P secretion, and fails to capture key dimensions of nutritional vulnerability, including skeletal muscle mass and overall nutritional risk profiles (4, 10). In contrast, selected biochemical indices (e.g., serum albumin and lymphocyte count) and measures of body composition and fat distribution have shown associations with nutritional risk and adverse outcomes in diabetes populations, offering complementary prognostic information beyond anthropometric assessment alone (13, 32, 33, 38).

Available evidence supports multidimensional nutritional phenotyping—integrating body composition measures with selected biochemical and inflammatory indices—as a more informative approach for characterizing nutrition-conditioned β-cell vulnerability in T5DM (11, 15).

4 Epidemiological characteristics4.1 Global epidemiology

The conceptualization of T5DM has been informed by clinical phenotypes historically described as MRDM and has predominantly been reported among individuals with a history of early-life nutritional deprivation, particularly in resource-limited settings with constrained access to structured endocrine care (3, 5).

The global burden of diabetes continues to rise, affecting an estimated 529 million individuals worldwide in 2021, approximately 96% of whom