Beyond Glucose and body Weight: Standards of Care Revaluation and A Citrate-Centred Model of Cellular Metabolic Failure in Type 2 Diabetes  

Summary  The American Diabetes Association (ADA) Standards of Care in Diabetes has been revised annually since 1989 and now exists as a continuously updated living guideline. Its underlying treatment logic has shifted twice: from glucose-lowering as an end in itself (1989–2008), to cardiorenal outcomes and person-centred care (2009–2019), to adiposity as the principal modifiable lever (2020–2026). Each shift followed a reconstruction of the evidence base rather than a change in fashion. 

To review the evidence underlying these shifts, to identify what they have not solved, and to develop a mechanistic hypothesis that links dietary energy surplus, mitochondrial redox stress, insulin resistance and inflammaging through a single node — the oxidative inactivation of mitochondrial aconitase (ACO2) and the resulting export of citrate. 

We propose that chronic caloric and substrate surplus raises the mitochondrial NADH/NAD⁺ ratio, generating reductive stress and electron leak; that the solvent-exposed [4Fe-4S] cluster of ACO2 is among the earliest and most kinetically favoured targets of the resulting superoxide and peroxynitrite; that ACO2 inactivation causes matrix citrate accumulation and SLC25A1-mediated export; and that cytosolic citrate is then cleaved by ATP-citrate lyase into an acetyl-CoA pool that feeds two pathogenic branches — de novo lipogenesis with diacylglycerol–PKCθ-mediated blockade of insulin signalling, and histone acetylation at senescence-associated secretory phenotype (SASP) loci. The second branch is supported by direct experimental evidence published in 2026 showing that the mitochondrial pyruvate–citrate–acetyl-CoA axis licenses SASP gene transcription and that SLC25A1 inhibition suppresses inflammation and improves healthspan in aged mice. 

If a common upstream node exists, the logical next step for guideline development is not another drug class but a stratification layer: assessment of mitochondrial oxidative capacity, redox burden, membrane integrity and residual β-cell reserve before and during pharmacotherapy. We outline a candidate five-domain phenotyping panel, grade the analytical and clinical maturity of each marker honestly, propose a four-group stratification that is additive to — not a replacement for — current algorithms, and set out the falsifiable predictions and clinical trials that would be required to move any of this from hypothesis to standard of care. 

 Keywords: type 2 diabetes; aconitase; SLC25A1; ATP-citrate lyase; acetyl-CoA; insulin resistance; senescence-associated secretory phenotype; metabolic phenotyping; Standards of Care 

 1. Introduction  Clinical guidelines are usually read as instructions. They are more informative when read as archaeology: each recommendation is a sediment layer deposited by a trial that changed what the field believed. Read that way, the ADA Standards of Care in Diabetes is an unusually complete record of how a disease concept has migrated over thirty-seven years — from a disorder of blood glucose, to a disorder of cardiorenal risk, to a disorder of energy storage.  Three observations motivate this review. First, each migration was driven by a trial whose result contradicted the prevailing model, not by incremental refinement of it. Second, the current adiposity-centred model, for all its therapeutic success, describes what to reduce without specifying what is broken inside the cells whose failure produces the phenotype. Third, a substantial minority of patients — those with normal body weight, those who do not respond to incretin therapy, those whose glycaemia deteriorates despite achieved weight targets — are poorly served by an algorithm indexed principally to body mass index (BMI) and glycated haemoglobin (HbA1c).  This paper has three parts. Part I reviews the evidentiary history of the ADA standards and the residual problems it leaves. Part II proposes a mechanistic hypothesis centred on oxidative inactivation of mitochondrial aconitase and citrate export, and examines the supporting and contradicting evidence with equal seriousness. Part III considers what a clinically deployable "cellular metabolic assessment" would actually require, which of its components are ready and which are not, and what trials would test it. Nothing in this paper should be used to defer or withhold guideline-directed therapy.  

2. Part I — Three eras of the Standards of Care 

2.1 The glucocentric era (1989–2008) The first ADA standards appeared in 1989 and were built on a simple and, at the time, well-supported premise: hyperglycaemia causes the complications of diabetes, so lowering glucose should prevent them. The Diabetes Control and Complications Trial established this for type 1 diabetes, with intensive therapy reducing the risk of retinopathy progression by roughly three-quarters [1]. The United Kingdom Prospective Diabetes Study extended the logic to type 2 diabetes, reporting a 25% reduction in microvascular endpoints with intensive glycaemic control, and — in the overweight subgroup allocated to metformin — a reduction in diabetes-related death and all-cause mortality [2,3]. Ten-year post-trial follow-up subsequently demonstrated a "legacy effect": microvascular benefit, and eventually macrovascular benefit, persisted long after between-group HbA1c differences had disappeared [4].  The premise was then tested at its limit and failed. ACCORD randomised 10,251 participants at high cardiovascular risk to a target HbA1c below 6.0% or to standard therapy. The intensive arm was terminated at 3.5 years because of excess mortality — 257 deaths versus 203 — despite better glycaemic control [5]. ADVANCE found no mortality benefit from intensive control, with reduced nephropathy [6]; VADT found no significant cardiovascular benefit [7]. Taken together, these trials did not refute the glucose hypothesis, but they demolished its unqualified form. Glucose lowering is not a universal good independent of how it is achieved, in whom, and at what cost in hypoglycaemia and weight gain.  

2.2 The outcomes era (2009–2019)  The regulatory response to the rosiglitazone controversy — mandatory cardiovascular outcome trials for new glucose-lowering agents — converted a compliance exercise into the most productive period of evidence generation in the field's history.  EMPA-REG OUTCOME reported that empagliflozin reduced the composite of cardiovascular death, non-fatal myocardial infarction and non-fatal stroke by 14%, with a 38% relative reduction in cardiovascular death and a 35% reduction in hospitalisation for heart failure [8]. LEADER showed a 13% reduction in major adverse cardiovascular events with liraglutide [9], and SUSTAIN-6 a 26% reduction with semaglutide [10]. Crucially, the magnitude and time course of benefit were incompatible with glucose lowering as the mechanism: the HbA1c separations were small, and the heart failure benefit in EMPA-REG appeared within weeks.  Renal outcome trials followed the same pattern. CREDENCE, DAPA-CKD and EMPA-KIDNEY established sodium–glucose cotransporter-2 (SGLT2) inhibition as kidney-protective across populations with and without diabetes [11–13], and DAPA-HF and EMPEROR-Preserved extended the class into heart failure irrespective of diabetes status [14,15]. A drug class developed to lower glucose had become a cardiorenal therapy that happens to lower glucose.  The guideline absorbed this. Metabolic surgery entered the recommendations in 2009; glucagon-like peptide-1 receptor agonists (GLP-1 RA) in 2012; a dedicated obesity management section in 2016; and in 2017 the terminology shifted from "bariatric" to "metabolic" surgery, an explicit acknowledgement that the operation treats metabolism rather than mass. The 2018–2019 editions restructured pharmacological recommendations around comorbidity — established atherosclerotic disease, heart failure, chronic kidney disease — rather than around HbA1c alone [16].  

2.3 The adiposity era (2020–2026)  The third shift followed the demonstration that pharmacological weight reduction of surgical magnitude was achievable. STEP 1 reported a mean weight change of −14.9% with semaglutide 2.4 mg at 68 weeks versus −2.4% with placebo [17]; SURMOUNT-1 reported −20.9% with tirzepatide 15 mg at 72 weeks [18]. SELECT then showed that semaglutide 2.4 mg reduced major adverse cardiovascular events by 20% in people with overweight or obesity and established cardiovascular disease without diabetes — separating the benefit of the drug from its glucose-lowering role entirely [19].  Indication creep followed biology. FLOW demonstrated a 24% reduction in kidney disease progression and cardiovascular death with semaglutide in type 2 diabetes with chronic kidney disease [20]. ESSENCE showed histological resolution of metabolic dysfunction-associated steatohepatitis [21]. SUMMIT showed benefit in heart failure with preserved ejection fraction and obesity [22]. Oral semaglutide demonstrated cardiovascular benefit in SOUL [23], and tirzepatide met its cardiovascular outcome endpoint against an active comparator in SURPASS-CVOT [24]. In parallel, DiRECT established that a structured low-energy dietary intervention could produce remission of type 2 diabetes in 46% of participants at 12 months and 36% at 24 months, with remission tightly dependent on the magnitude of weight loss sustained [25].  The 2025 and 2026 editions codify the result. The 2026 Standards moved to a living-guideline model updated continuously online, recommends considering continuous glucose monitoring at diagnosis, supports GLP-1 therapy in type 1 diabetes for the first time, and sets a 5–7% weight-loss target, with annual screening for overweight and obesity using BMI plus an additional body-fat measure where available, and automated insulin delivery made available to all adults requiring insulin [26,27]. The pipeline continues to accelerate: the oral small-molecule GLP-1 receptor agonist orforglipron was approved for chronic weight management in April 2026, and the triple GIP/GLP-1/glucagon receptor agonist retatrutide has completed pivotal phase 3 obesity trials [28,29]. The guideline has moved from pushing the glucose down to manage the adiposity and protect the heart and kidney.  

3. What the adiposity paradigm does not solve  The therapeutic gains are real and should not be minimised. Here problems nevertheless remain, and they share a common structure: they are all failures of individualisation in a framework built on population averages. Even in the most successful outcome trials, relative risk reductions of 14–26% leave the majority of events unprevented. The absolute event rates in the treatment arms of EMPA-REG, LEADER and SELECT remain substantial [8,9,19]. Weight and glycaemic responses to incretin therapy vary widely across individuals receiving identical doses, and a meaningful minority are non-responders. Current guidelines offer no mechanistic account of why, and no pretreatment test that predicts it. A substantial proportion of type 2 diabetes — particularly, though not exclusively, in East and South Asian populations — occurs at a BMI below conventional obesity thresholds, with impaired insulin secretion rather than gross adiposity as the dominant defect [30,31]. Data-driven cluster analyses have repeatedly resolved adult-onset diabetes into subgroups including severe insulin-deficient diabetes (SIDD) and mild age-related diabetes (MARD) alongside severe insulin-resistant diabetes (SIRD), with markedly different complication profiles [31]. An algorithm whose central lever is weight reduction has little to offer SIDD, and weight-directed therapy in a sarcopenic older adult can be actively harmful — a concern the 2026 Standards acknowledge in their emphasis on muscle mass and bone density monitoring in older adults [26]. Withdrawal of semaglutide in the STEP 1 extension was followed by regain of approximately two-thirds of the lost weight within a year, with cardiometabolic variables reverting toward baseline [32]. This is not a criticism of the drug — it is the expected behaviour of a therapy that suppresses appetite rather than repairing the tissue defect. But it defines the therapeutic ceiling of the current paradigm: continuous pharmacological suppression of a phenotype whose cellular substrate is unaddressed. If weight reduction improves glycaemia, cardiovascular outcomes, renal outcomes and hepatic histology simultaneously, what single cellular process is being relieved? The remainder of this paper proposes a candidate. 

 4. Part II — A citrate-centred hypothesis 

4.1 Statement of the hypothesis  Chronic caloric and substrate surplus raises the mitochondrial NADH/NAD⁺ ratio beyond the capacity of the electron transport chain to dissipate it (reductive stress). The resulting electron leak generates superoxide and, with nitric oxide, peroxynitrite. Among matrix enzymes, mitochondrial aconitase (ACO2) is kinetically the most vulnerable target because of its solvent-exposed [4Fe-4S] cluster. ACO2 inactivation stalls the tricarboxylic acid (TCA) cycle at citrate, raising matrix citrate and driving SLC25A1-mediated export to the cytosol, where ATP-citrate lyase (ACLY) regenerates acetyl-CoA. This ectopic acetyl-CoA pool feeds two pathogenic branches — lipogenic (insulin resistance) and epigenetic (inflammaging) — which reinforce each other through a feed-forward loop.  

4.2 Reductive stress as the initiating lesion  The proposal that substrate surplus creates a reductive rather than primarily oxidative lesion is not new. Williamson and colleagues described the "pseudohypoxia" of hyperglycaemia — an elevated cytosolic NADH/NAD⁺ ratio driven by polyol-pathway flux, mimicking the redox signature of hypoxia in normally oxygenated tissue [33,34]. Brownlee subsequently unified the polyol, hexosamine, protein kinase C and advanced glycation pathways under a single upstream event: overproduction of superoxide by the mitochondrial electron transport chain under substrate excess [35,36].  The modern framing is that excess reducing equivalents, not merely excess glucose, constitute the proximate stress. When NADH delivery to complex I exceeds the rate at which the respiratory chain can oxidise it — a state favoured by simultaneous carbohydrate and lipid surplus with low energy demand — the ubiquinone pool becomes over-reduced, reverse electron transport at complex I becomes thermodynamically favourable, and superoxide production rises steeply. Reductive stress is now recognised as a distinct pathological state with its own signature and consequences, not simply the mirror image of oxidative stress [37]. Incomplete fatty acid oxidation under mitochondrial overload, with accumulation of acylcarnitine intermediates, provides a complementary route to the same end [38].  

4.3 Aconitase as the sentinel target  Mitochondrial aconitase catalyses the isomerisation of citrate to isocitrate via cis-aconitate. Its catalytic centre contains a cubane [4Fe-4S] cluster in which three iron atoms are coordinated by cysteine thiolates while the fourth, Feα, is solvent-exposed and binds the substrate directly. This architecture — a redox-active metal centre deliberately left open to the matrix — is what makes the enzyme catalytically efficient and, simultaneously, chemically fragile.  The kinetic evidence is quantitative and specific. Castro, Radi and colleagues have reviewed aconitases as principal targets of reactive oxygen and nitrogen species, noting that the reaction of the Fe–S cluster with superoxide proceeds at approximately 10⁷ M⁻¹s⁻¹, is relatively specific and is reversible in vivo — properties that have made aconitase activity itself a widely used cellular probe for superoxide formation [39]. Direct measurement on purified human ACO2 subsequently gave second-order rate constants of approximately 10⁸ M⁻¹s⁻¹ for superoxide and carbonate radical, 10⁵ M⁻¹s⁻¹ for peroxynitrite and 10² M⁻¹s⁻¹ for hydrogen peroxide [40]. Peroxynitrite-mediated inactivation converts the active [4Fe-4S]²⁺ cluster to the catalytically inactive [3Fe-4S]⁺ form with release of Feα, which is itself Fenton-reactive and can propagate hydroxyl radical generation locally [41,42]. Frataxin-dependent cluster repair determines whether the lesion is transient or persistent [40]. Two properties make ACO2 attractive as a pathophysiological node. It fails early, at ROS fluxes well below those required to damage most matrix proteins. And when it fails, it fails into a specific metabolite — citrate — rather than into generalised dysfunction.  

4.4 Citrate export and the acetyl-CoA branch point  Matrix citrate is exported by SLC25A1, the mitochondrial citrate carrier, in exchange for malate, and cleaved in the cytosol by ACLY into acetyl-CoA and oxaloacetate. Under physiological conditions this is the principal route by which mitochondrial carbon supplies cytosolic and nuclear acetyl-CoA for lipogenesis and for protein acetylation. Under the conditions proposed here, the same route becomes a pathological overflow valve.  Citrate is also, independently, a recognised inflammatory signalling metabolite: in activated macrophages, citrate export supports production of nitric oxide, reactive oxygen species and prostaglandins [43,44], while diversion of cis-aconitate to itaconate constitutes an endogenous brake on this program through inhibition of succinate dehydrogenase and activation of NRF2 [45,46]. The TCA cycle in inflammatory cells is thus already understood as a signalling hub rather than merely a catabolic pathway — precedent for the claim advanced here.

To be continued  

Note: This article was completed by Dr. Jin Haining and Dr. Xu Bin with AI assistance. Due to the limitations of our professional knowledge, there must be inaccuracies in this paper. Please feel free to point them out and give your corrections. Thank you!

支持单位：