Caloric-Restriction Mimetics

Caloric-Restriction Mimetics

A caloric-restriction mimetic is an intervention proposed to reproduce selected biological effects of caloric restriction without sustained caloric restriction itself. Most candidates affect a pathway that caloric restriction also affects. Far fewer reproduce a metabolic signature of restriction. Fewer still extend lifespan in a genetically heterogeneous mammal. Human evidence for delaying aging in otherwise healthy people is thinner still. This article applies that graded test. It is not a catalogue of compounds to take.

Compiled by South Beach Longevity · 20 August 2026 Copyright 2026 Series SBL-41 / SP-CALORIC-RESTRICTION-MIMETICS · Register A scientific article Sources peer-reviewed lifespan studies, the NIA Interventions Testing Program, human trials, primate colonies, and labelled mechanistic work · verified NCBI records Constraint This document describes published research. It is not medical advice. No human use, dose, route or schedule is recommended anywhere in this document.

How to read this document Every finding is labelled, in the sentence that reports it, by the kind of study that produced it. A genetically heterogeneous mouse lifespan trial is not a pathway diagram. A diabetes-outcome trial is not an aging trial in healthy adults. An epigenetic-clock movement is not a demonstrated delay of aging. Amounts and durations appear only as reported experimental parameters, always with the population attached. Nothing here is a recommendation for or against any drug, food, or supplement.

Three layers are kept apart throughout: pathway modulation, reproduction of metabolic or transcriptional CR signatures, and organism-level effects (lifespan, healthspan, or a pre-registered human aging endpoint). Findings are graded in place as established, strongly supported, supported, emerging, plausible, or speculative, and where a claim fails on its own evidence, as refuted. Two further labels mark careful absences rather than verdicts: not established, where the evidence is too thin to place a claim on the ladder at all — untested or insufficient, an absence of proof rather than disproof; and not supported, where the weight of evidence leans against a claim but stops short of a formal refutation. Those grades are not marketing stars.

01 The word, and the substitutions that wreck it

The phrase is older than the supplement aisle. Ingram, Zhu, Mamczarz, Zou, Lane, Roth and de Cabo, writing in Aging Cell in 2006, treated a calorie-restriction mimetic as an emerging research object: an intervention that would produce the pro-longevity physiology of restriction without a sustained cut in intake (Ingram et al., 2006). Earlier program papers from the same group had already named 2-deoxy-D-glucose as a prototype and had warned that the field would need organism-level tests, not only enzyme assays (Ingram et al., 2004; Roth et al., 2005). Madeo, Pietrocola, Eisenberg and Kroemer later offered a molecular definition organized around autophagy induction and protein-acetylation states (Madeo et al., 2014). Ingram and Roth restated the practical problem in 2015: the attractive slogan is “have your cake and eat it too”; the scientific requirement is to show that the cake was unnecessary because the organism-level effects were reproduced (Ingram and Roth, 2015).

Three substitutions wreck the word.

First, a shared pathway is treated as a shared outcome. AMPK activation, mTOR inhibition, sirtuin engagement, NAD replenishment, insulin/IGF dampening, autophagy induction, mitochondrial remodeling, and altered glucose handling all appear in CR reviews (Masoro, 2005; Fontana and Partridge, 2015). They also appear, singly, in the mechanism section of almost every candidate’s marketing file. A pathway that CR uses is not evidence that a drug has done what CR does.

Second, a disease-treatment result is treated as an aging result. Metformin’s UKPDS record in overweight adults with type 2 diabetes (UKPDS Group, 1998) and acarbose’s STOP-NIDDM diabetes-prevention record (Chiasson et al., 2002) are real. They are not demonstrations that biological aging was delayed in otherwise healthy people.

Third, a surrogate is treated as a lifespan. Circulating glucose, p62 flux, NAD+ in whole blood, and a DNA-methylation clock can all move without an organism living longer or functioning longer. The NIA Interventions Testing Program exists because those moves had already misled the field (Nadon, Strong, Miller et al., 2008).

This title is the classification problem. The sibling question—how caloric restriction itself should be implemented in a person—is not asked here. Amounts used in experiments are experimental facts. They are not instructions.

02 The ten-question intervention test

For every candidate the same ten questions are asked. A “yes” to question 1 does not promote the agent. A “yes” to questions 3–6, replicated, is what would justify the term.

The grades used in the matrices:

• STRONGLY SUPPORTED — replicated organism-level effect in a design built to test longevity or a CR-like healthspan battery; human evidence, if claimed, is the same kind of evidence.
• SUPPORTED — organism-level effect in at least one rigorous setting, with a recognized limit (sex, strain, diet, or disease population).
• EMERGING — a real signal on a narrower endpoint, or a single unrepeated lifespan result.
• NOT ESTABLISHED — credible positives and credible nulls that have not been reconciled.
• NOT SUPPORTED — the claim has been tested in a design that should have seen it, and did not, or the human claim is only a pathway or a clock.

03 Pathway classes that are not, by themselves, CR mimetics

The following classes are real biology. They are the filing system for candidates, not a list of proven mimetics.

AMPK. Energy-charge sensing that CR and some biguanides engage. Activation in a cell or a tissue is a mechanism note.

mTOR. Nutrient-sensitive kinase; chronic inhibition is the best-replicated pharmacological longevity intervention in mice (Harrison et al., 2009; Miller et al., 2011; Miller et al., 2014). Inhibition is not automatically CR. Miller and colleagues showed that rapamycin’s metabolic signature is distinct from CR even when both extend life (Miller et al., 2014). Lamming, Ye, Katajisto and colleagues separated mTORC1-linked longevity from mTORC2-linked insulin resistance in a mouse genetic study (Lamming et al., 2012).

Sirtuins and NAD metabolism. Howitz, Bitterman, Cohen, Blander, Westphal and colleagues reported that resveratrol and related molecules activated yeast Sir2 and extended replicative lifespan (Howitz et al., 2003). Chang and Guarente reviewed SIRT1 and metabolic sirtuins as CR-associated enzymes, not as proof that every sirtuin ligand is CR (Chang and Guarente, 2014). NAD+ decline and precursor feeding are a later chapter of the same story (Yoshino, Mills, Yoon and Imai, 2011; Trammell et al., 2016).

Insulin/IGF signaling. Dampened in many CR protocols. Also dampened by drugs whose clinical indication is hyperglycemia.

Autophagy. Madeo’s molecular definition puts autophagy near the center (Madeo et al., 2014). Eisenberg, Knauer, Schumacher, Fussi and colleagues showed that spermidine induced autophagy and promoted longevity in yeast, flies and worms (Eisenberg et al., 2009). Autophagy flux in a cell is not an aging trial.

Mitochondrial metabolism. CR changes substrate use and, in some protocols, oxidative-damage markers (Redman, Smith, Burton, Martin, Il’yasova and Ravussin, 2018). Uncoupling, NAD boosting, and ketogenic feeding can move overlapping readouts without being CR (Roberts et al., 2017).

Glucose handling. Acarbose and SGLT2 inhibitors change post-prandial or urinary glucose. That is their pharmacology. It becomes a CR-mimetic claim only if organism-level aging endpoints follow.

04 How this article grades a candidate

A candidate enters the reviewed record if a serious geroscience paper, an ITP publication, or a human trial has treated it as a CR-mimetic or aging-delay candidate. Marketing use of the phrase is not an entry ticket. A compound is discussed because the experimental record is strong enough to be wrong about, not because a label used the words.

The matrices in Part Three and Part Four apply the ten questions. The judgment in section 10 is the only place the phrase “CR mimetic” is awarded or withheld.

05 What CR has actually been shown to do

Masoro’s 2005 overview remains the right starting point: in laboratory rodents, a chronic reduction of energy intake that avoids malnutrition extends median and, in many colonies, maximum lifespan, delays a wide range of age-associated lesions, and changes insulin, body temperature, and fuel selection (Masoro, 2005). Barger, Walford and Weindruch, writing as the transgenic era opened, treated that rodent record as the most reproducible environmental intervention on mammalian aging then available (Barger, Walford and Weindruch, 2003). The effect is not magic and it is not uniform. May, Telford, Salo, Lapham, Han and Finch reported that dietary restriction failed to retard selected age-related neurochemical changes in mice—an early reminder that “CR does everything” was already a bad summary (May et al., 1992). López-Domínguez, Ramsey, Tran, Imai and colleagues later showed that dietary fat composition and age modulate apoptotic signaling in CR liver, which is a mechanism paper, not a lifespan paper (López-Domínguez et al., 2015).

Two facts travel together. First, the rodent effect is large enough, and old enough, that any proposed mimetic is answerable to it. Second, the rodent effect is sensitive to diet composition, age at onset, degree of restriction, and genetic background. A drug that “works like CR” must say which CR.

06 Nonhuman primates: two colonies, one argument

The Wisconsin National Primate Research Center study reported that adult-onset CR in rhesus macaques delayed age-associated disease and reduced mortality compared with controls fed a semi-purified diet ad libitum (Colman, Anderson, Johnson, Kastman, Kosmatka, Beasley et al., 2009). The NIA intramural study, using a different diet (natural ingredients), different control-feeding practice, and a different mix of ages at onset, did not find a survival benefit of CR in its 2012 report, though some health measures moved (Mattison, Roth, Beasley, Tilmont, Handy, Herbert et al., 2012). Mattison, Colman, Beasley, Allison, Kemnitz, Roth and de Cabo later published a combined analysis: when the protocols are read as two implementations rather than as one failed replication, CR improved health and survival of rhesus monkeys under the conditions that actually produced a restricted, not merely a less-obese, animal (Mattison et al., 2017).

That pair is the primate lesson for mimetics. Site, diet, and control body weight are not footnotes. A pharmacological agent tested in one inbred, obese mouse strain against an overfed control has the same problem the 2012 NIA paper exposed: the comparison may be “less sick than the sick control,” not “CR-like.”

Didier, MacLean, Mohan, Didier, Lackner and Kuroda reviewed the broader nonhuman-primate aging contribution; it is context, not a third CR colony (Didier et al., 2016).

07 Humans: CALERIE, voluntary restriction, and what was not measured

No randomized human trial has tested whether CR extends lifespan. The Comprehensive Assessment of Long-term Effects of Reducing Intake of Energy (CALERIE) program tested feasibility and predictors of health span, not survival.

Heilbronn, de Jonge, Frisard, DeLany, Larson-Meyer, Rood and colleagues reported a six-month randomized trial in overweight adults: prescribed 25% CR, CR plus exercise, a low-calorie diet, or control. Fasting insulin, body temperature, and some oxidative-stress markers moved; the study was a biomarker and feasibility trial (Heilbronn et al., 2006). Rickman, Williamson, Martin, Gilhooly, Stein, Bales and colleagues described the two-year Phase 2 design: an intended 25% restriction in non-obese adults (Rickman et al., 2011). Ravussin, Redman, Rochon, Das, Fontana, Kraus and colleagues reported that two-year CR was feasible, achieved a smaller restriction than prescribed (about 12% on average in the published feasibility analysis), reduced circulating T3 and cardiometabolic risk markers, and did not produce the catastrophic safety signal a starvation protocol would have produced (Ravussin et al., 2015). Kraus, Bhapkar, Huffman, Pieper, Das, Redman and colleagues reported exploratory two-year cardiometabolic outcomes: blood pressure, lipids, insulin sensitivity, and inflammatory markers moved in the CR arm (Kraus et al., 2019). Redman, Smith, Burton, Martin, Il’yasova and Ravussin reported metabolic slowing and reduced oxidative damage with sustained CR, framed as support for rate-of-living and oxidative-damage accounts (Redman et al., 2018). Khong’s 2020 clinical note restated the cardiometabolic improvement; it is secondary reporting, not a new trial (Khong, 2020).

Separately, Fontana, Meyer, Klein and Holloszy measured long-term voluntary CR practitioners and found a carotid and lipid profile that would be favorable in a cardiovascular clinic (Fontana et al., 2004). That is a selected cohort, not a randomized aging trial.

What CALERIE and the voluntary cohorts do not show: delayed mortality, delayed dementia, or a general “anti-aging” license for any drug that lowers glucose. They establish that moderate CR in selected, monitored, non-obese or overweight adults can move metabolic and some oxidative-stress markers. That is the human reference a mimetic must meet or beat—without the restriction.

08 The NIA Interventions Testing Program as the adult table

Nadon, Strong, Foster, Nelson, McVey, Richardson and colleagues described the ITP as a three-site, peer-reviewed program in genetically heterogeneous UM-HET3 mice, built because single-laboratory, single-strain positives had already failed to travel (Nadon et al., 2008). Local Firecrawl retrieval of the NIA program page (July 2026 update) confirms the standing design: Jackson Laboratory, University of Michigan, and UT Health San Antonio; up to six interventions a year; open nomination; public data. That institutional fact is why ITP results are weighted above a single C57BL/6 colony in this article.

ITP-positive lifespan results in the reviewed record:

ITP-null or ITP-negative results that matter for slogans:

Flurkey, Astle and Harrison reported life extension by diet restriction and by N-acetyl-L-cysteine in UM-HET3 mice; that paper is a reminder that CR itself still works in the stock the ITP uses (Flurkey et al., 2010).

09 Rapamycin and rapalogs

Pathway. mTORC1 inhibition is the intended geroprotective pharmacology. mTORC2 loss is the metabolic liability (Lamming et al., 2012).

Organism-level evidence. Harrison, Strong, Sharp, Nelson, Astle, Flurkey and colleagues showed that rapamycin fed late in life extended lifespan in genetically heterogeneous mice of both sexes (Harrison et al., 2009). Miller, Harrison, Astle, Baur, Boyd, de Cabo and colleagues replicated the lifespan effect from mid-life and, in the same design, found no lifespan extension from resveratrol or simvastatin (Miller et al., 2011). Miller, Harrison, Astle, Fernandez, Flurkey, Han and colleagues then showed that the lifespan increase is dose- and sex-dependent and metabolically distinct from CR (Miller et al., 2014). Bitto, Ito, Pineda, LeTexier, Huang, Sutlief and colleagues reported that a transient rapamycin course in middle-aged C57BL/6 mice increased lifespan and some healthspan measures (Bitto et al., 2016). Zhang, Bokov, Gelfond, Soto, Ikeno, Hubbard and colleagues reported lifespan and health extension in C57BL/6 mice on a different protocol (Zhang et al., 2014). Those inbred-mouse papers support generality; they do not replace the ITP.

Human evidence. Mannick, Del Giudice, Lattanzi, Valiante, Holmes, Bolognese and colleagues reported that everolimus (RAD001) improved influenza-vaccine response in older adults in a randomized trial—an immune-function endpoint, not a lifespan endpoint (Mannick et al., 2014). Mannick, Morris, Hockey, Roma, Pesko, Hanna and colleagues reported that TORC1 inhibition enhanced immune function and reduced infections in older adults in a subsequent randomized program (Mannick et al., 2018). Mannick and Lamming’s 2023 review is the current map of what mTOR inhibitors have and have not shown in aging biology; it is a review, not a new trial (Mannick and Lamming, 2023). Transplantation immunosuppression is disease-context pharmacology and is not treated here as geroprotection.

Independence from disease treatment. The mouse lifespan effects were measured in animals that were not enrolled as diabetics or transplant recipients. The human trials were measured as vaccine or infection endpoints in older adults, not as “aging delayed.”

Safety tradeoffs, as studied. Glucose intolerance / insulin resistance from mTORC2 (Lamming et al., 2012); infection risk and delayed wound healing in the labelled immunosuppressant setting; stomatitis and dyslipidemia in clinical rapalog use. Those are reasons a healthy-person extrapolation is not a small step.

Is “CR mimetic” justified? Partial, and only with a qualifier. Rapamycin is the strongest pharmacological longevity agent in the mammalian record in the reviewed record. It is not a metabolic copy of CR (Miller et al., 2014). Calling it a CR mimetic is a historical courtesy. Calling it an mTORC1 longevity drug is more accurate.

10 Acarbose, canagliflozin, 17-alpha-estradiol, glycine

Acarbose. An alpha-glucosidase inhibitor that blunts post-prandial glucose. Strong, Miller, Antebi, Astle, Bogue, Denzel and colleagues reported longer lifespan in male UM-HET3 mice treated with acarbose, 17-alpha-estradiol, or NDGA, with a weaker female acarbose effect (Strong et al., 2016). Harrison, Strong, Alavez, Astle, DiGiovanni, Fernandez and colleagues reported that acarbose improved health and lifespan in aging HET3 mice (Harrison et al., 2019). Human evidence is disease-prevention and vascular: STOP-NIDDM reduced incident type 2 diabetes (Chiasson et al., 2002); Hanefeld, Cagatay, Petrowitsch, Neuser, Petzinna and Rupp reported vascular-disease analyses in abnormal glucose tolerance (Hanefeld et al., 2008). Those are not healthy-aging trials.

Canagliflozin. Miller, Harrison, Allison, Bogue, Debarba, Diaz and colleagues reported that canagliflozin extended lifespan in genetically heterogeneous male but not female mice (Miller et al., 2020). Snyder, Casey, Galecki, MacCoss, Rabinovitch, Kaeberlein and colleagues later reported retardation of age-related lesions in heart, kidney, liver and adrenal in the ITP stock (Snyder et al., 2023). Jayarathne, Stocker, Shetty, Simmons, Sung and colleagues reported a hypothalamic sex-specific metabolic shift—mechanism, not a second lifespan trial (Jayarathne et al., 2024). Human evidence is disease-outcome: CANVAS showed cardiovascular and renal event effects in type 2 diabetes (Neal et al., 2017). That is not a CR-mimetic demonstration in healthy adults.

17-alpha-estradiol. Male-specific lifespan extension at ITP conditions (Strong et al., 2016), including when started late (Harrison et al., 2021). Sadagurski, Cady, D’Alessio, Huffman, Miller and colleagues reported sex-specific reduction of hypothalamic inflammation by several anti-aging drugs, including this class (Sadagurski et al., 2017). There is no adequate human aging trial in the reviewed record. The sex restriction is the scientific finding, not a footnote.

Glycine. Miller, Harrison, Astle, Bogue, Brind, Fernandez and colleagues reported lifespan extension in male and female UM-HET3 mice with dietary glycine at the studied 8% level (Miller et al., 2019). Human aging-trial evidence is absent here. An amino-acid diet at that experimental fraction is not a capsule narrative.

CR-mimetic status. Acarbose and canagliflozin are glucose-handling longevity agents with ITP organism-level evidence, male-biased, and human disease-drug records. They are closer to “CR-adjacent pharmacology” than to proven human mimetics. 17-alpha-estradiol is a male-mouse ITP fact. Glycine is a modest two-sex ITP fact. None has human evidence that aging was delayed independently of disease treatment.

11 Metformin

The slogan. “Metformin is an anti-aging drug.”

What the mouse record actually says. Martin-Montalvo, Mercken, Mitchell, Colman, Lambert, Malloy and colleagues reported that 0.1% metformin in diet improved healthspan and modestly increased lifespan in male C57BL/6 mice; a higher 1% condition was harmful (Martin-Montalvo et al., 2013). The ITP, using UM-HET3 mice and 0.1% metformin, did not replicate a lifespan extension (Strong et al., 2016). That pair is the strain-and-design lesson. Novelle, Ali, Diéguez, Bernier and de Cabo reviewed the hope and the limits (Novelle et al., 2016).

What the human record actually says. UKPDS 34 found that intensive glucose control with metformin in overweight patients with type 2 diabetes reduced diabetes-related endpoints and all-cause mortality compared with conventional policy (UKPDS Group, 1998). Bannister, Holden, Jenkins-Jones, Morgan, Halcox, Schernthaner and colleagues reported an observational comparison in which people with type 2 diabetes initiating metformin had lower mortality than matched people without diabetes; the design cannot convert that into “metformin makes healthy people live longer” (Bannister et al., 2014). Barzilai, Crandall, Kritchevsky and Espeland described TAME—Targeting Aging with Metformin—as a proposed trial to test aging-related endpoints rather than a single disease (Barzilai et al., 2016). A proposal is not a result. Prescrire’s 2014 monotherapy note is included only as a labelled secondary caution on the strength of first-line efficacy evidence; it is not a geroscience paper (Prescrire, 2014).

CR-mimetic status. Not justified on present organism-level and healthy-human evidence. Metformin is a first-line diabetes drug with a plausible AMPK/mitochondrial mechanism, a C57BL/6 healthspan paper, an ITP lifespan miss, and a human record that is disease treatment plus observational controversy. TAME, if completed to its registered endpoints, would change the human sentence. It would still not automatically make metformin a CR mimetic.

12 Resveratrol and the sirtuin narrative

Howitz et al. (2003) is the yeast origin. Baur, Pearson, Price, Jamieson, Lerin, Kalra and colleagues reported that resveratrol improved health and survival of mice on a high-calorie diet—an obese-mouse survival paper, not a lean-CR equivalent (Baur et al., 2006). Pearson, Baur, Lewis, Peshkin, Price, Labinskyy and colleagues reported that resveratrol delayed age-related deterioration and mimicked transcriptional aspects of dietary restriction without extending lifespan on a standard diet in the way CR does (Pearson et al., 2008). Miller et al. (2011) and Strong et al. (2013) then tested resveratrol in UM-HET3 mice and did not extend life.

Timmers, Konings, Bilet, Houtkooper, van de Weijer, Goossens and colleagues reported calorie-restriction-like effects on energy metabolism in eleven obese men after 30 days of resveratrol at the studied experimental amount—a small, short, disease-adjacent metabolic trial (Timmers et al., 2011). It is not a lifespan study and it is not a healthy-aging program.

CR-mimetic status. NOT SUPPORTED as an organism-level mammalian CR mimetic. The honest remainder is: a yeast activator, an obese-mouse helper, a transcriptional resemblance in some tissues, a human metabolic hint in a tiny obese sample, and a clear ITP miss.

13 NAD-related interventions

Yoshino, Mills, Yoon and Imai reported that NMN treated pathophysiology of diet- and age-induced diabetes in mice (Yoshino et al., 2011). Trammell, Schmidt, Weidemann, Redpath, Jaksch, Dellinger and colleagues showed that nicotinamide riboside is orally bioavailable in mice and humans and raises NAD+ (Trammell et al., 2016). Martens, Denman, Mazzo, Armstrong, Reisdorph, McQueen and colleagues reported that chronic NR in healthy middle-aged and older adults was well tolerated and elevated NAD+; blood-pressure and aortic-stiffness signals were exploratory (Martens et al., 2018). Yoshino, Yoshino, Kayser, Patti, Franczyk, Mills and colleagues reported that NMN increased muscle insulin sensitivity in prediabetic women in a randomized trial—a metabolic endpoint in a disease-risk population (Yoshino et al., 2021). Irie, Inagaki, Fujita, Nakaya, Mitsuishi, Yamaguchi and colleagues reported oral NMN clinical-parameter and nicotinamide-metabolite data in a small Japanese sample (Irie et al., 2020). Yi, Ma, Zhang, Wang, Shan, Li and colleagues reported efficacy and safety of beta-NMN in healthy middle-aged adults on selected laboratory and NAD-related readouts (Yi et al., 2023).

The ITP sentence is shorter: nicotinamide riboside at 1000 ppm did not extend UM-HET3 lifespan in the report that did extend male life with late 17-alpha-estradiol (Harrison et al., 2021).

CR-mimetic status. NOT SUPPORTED. Raising NAD+ is a pharmacokinetic success. Reproducing CR’s organism-level effects is a different claim and has not been met in the heterogeneous-mouse lifespan assay in the reviewed record. Human trials remain small, short, and endpoint-narrow.

14 Spermidine

Eisenberg, Knauer, Schumacher, Fussi, Butschal, Carmona-Gutierrez and colleagues reported that spermidine induced autophagy and promoted longevity in yeast, flies and worms (Eisenberg et al., 2009). Eisenberg, Abdellatif, Schroeder, Primesnig, Stekovic, Pendl and colleagues later reported cardioprotection and lifespan extension in mice (Eisenberg et al., 2016). Kiechl, Pechlaner, Willeit, Notdurfter, Paulweber, Willeit and colleagues reported that higher dietary spermidine intake was linked to lower mortality in a prospective population study (Kiechl et al., 2018). Yu, Jiang, Zhang, Liu, Chen and colleagues reported serum spermidine and prognosis after myocardial infarction—a disease-cohort association, not a CR trial (Yu et al., 2022).

CR-mimetic status. EMERGING as a longevity candidate; not justified as a demonstrated human CR mimetic. Autophagy and a C57BL/6 cardioprotection/lifespan paper are scientifically interesting. A food-frequency mortality association is not an intervention. There is no ITP lifespan result in the reviewed record.

15 Alpha-ketoglutarate and 2-deoxyglucose

Chin, Fu, Pai, Vergnes, Hwang, Deng and colleagues reported that alpha-ketoglutarate extended C. elegans lifespan by inhibiting ATP synthase and TOR (Chin et al., 2014). Asadi Shahmirzadi, Edgar, Liao, Hsu, Lucanic, Asadi Shahmirzadi and colleagues reported that calcium alpha-ketoglutarate extended lifespan and compressed morbidity in aging mice, with a stronger morbidity-compression signal in females (Asadi Shahmirzadi et al., 2020). Demidenko, Barardo, Budovskii, Finnemore, Palmer, van Bruggen and colleagues reported Rejuvant, a commercial Ca-AKG plus vitamin formulation, and DNA-methylation-clock movement in a human observational / open-label setting (Demidenko et al., 2021). A clock movement in a marketed formulation is not a CR-mimetic demonstration.

2-Deoxy-D-glucose was the prototype CR mimetic in the Ingram–Roth–Lane program (Ingram et al., 2004, 2006). Minor, Smith, Csiszar, Kaushik, Khodr, Spangler and colleagues then reported that chronic 2-DG ingestion induced cardiac vacuolization and increased mortality in rats (Minor et al., 2010). The prototype failed the organism-level safety test that the definition requires.

CR-mimetic status. AKG is EMERGING (worm plus one mouse paper; human clock study is not an aging trial). 2-DG is NOT SUPPORTED as a usable mimetic; it is a historical warning.

16 Objects that appear beside this literature and are not CR mimetics

Fahy, Brooke, Watson, Good, Vasanawala, Maecker and colleagues reported reversal of epigenetic-aging and immunosenescent trends in a small, uncontrolled human protocol using growth hormone, DHEA and metformin (Fahy et al., 2019). That is the TRIIM report. It is not a CR-mimetic study. It is cited so it cannot be silently imported as one.

Roberts, Wallace, Tomilov, Zhou, Koenig, Cazares and colleagues reported that a ketogenic diet extended longevity and healthspan in adult mice (Roberts et al., 2017). A diet that changes calories’ form is not a CR mimetic; it is a different dietary intervention.

Mau, Yung, and colleagues reported that several life-span-extending drug interventions affect adipose inflammation in aging (Mau et al., 2020). That is a downstream tissue paper, not a new candidate class.

17 Table A. Candidate / mechanism / organism-level matrix

18 Table B. Species lifespan (selected, as published)

19 Table C. Healthspan evidence (selected)

20 Table D. Human trial matrix (disease treatment kept separate)

21 Table E. Safety and tradeoff (studied populations only)

22 Table F. CR versus proposed mimetics

23 Claim / evidence ledger

24 Five-lens interrogation

Geroscience advocate. The strongest case is narrow and should be stated without apology. Chronic energy restriction is the most reproducible environmental intervention on rodent aging (Masoro, 2005). A three-site, genetically heterogeneous mouse program has now shown that selected drugs—rapamycin most clearly, acarbose and canagliflozin with sex structure, 17-alpha-estradiol in males—can extend life without a prescribed calorie cut (Harrison et al., 2009; Miller et al., 2011, 2014, 2020; Strong et al., 2016; Harrison et al., 2019, 2021). That is what “pharmacologic CR-adjacent geroprotection” looks like when it is real. Mannick’s older-adult immune trials show that mTORC1 can be touched in humans without immediately collapsing the program (Mannick et al., 2014, 2018). CALERIE shows that humans tolerate moderate CR well enough to measure biology (Ravussin et al., 2015). The advocate’s best sentence is not “we have CR in a pill.” It is “organism-level pharmacology of aging is no longer empty.”

Translational skeptic. UM-HET3 is still a mouse. Control diets, husbandry, and the absence of human infection, medication, and lifelong atherosclerotic burden are not small differences. The primate CR literature already showed that two excellent laboratories can disagree until diet and control weight are faced (Colman et al., 2009; Mattison et al., 2012, 2017). Rapamycin’s human file is vaccine and infection, not a decade of healthy-person survival. Sex-restricted ITP hits (canagliflozin, 17-alpha-estradiol, NDGA, aspirin) are warnings that “the mouse” is already two experiments. A healthy 50-year-old is not a 20-month UM-HET3 male.

Clinical pharmacologist. Dose–response in the ITP is an animal-diet ppm fact, not a human regimen. Rapamycin and rapalogs have labelled immunosuppressant toxicities; Lamming et al. (2012) mechanistically tied glucose intolerance to mTORC2. Metformin, acarbose and canagliflozin have diabetes labels and diabetes-trial adverse-event tables. Combining them, pulsing them, or giving them to euglycemic adults is not an evidence-based aging protocol in the reviewed record; it is a different experiment that has not been reported as such. Drug–drug interaction space (metformin plus rapamycin plus an SGLT2 inhibitor plus an NAD capsule) is a clinic, not a stack.

Aging biologist. Does pathway modulation reproduce CR? For rapamycin, Miller et al. (2014) already answered: lifespan can rise while the metabolic signature diverges. For resveratrol, Pearson et al. (2008) answered: transcription can resemble CR while lifespan does not. For NAD precursors, the cofactor rises and the ITP NR arm does not live longer (Harrison et al., 2021). For 2-DG, the prototype pathway (glycolytic inhibition) produced a heart lesion (Minor et al., 2010). CR is a coordinated organismal state. A single node is not that state.

Methodologist. Lifespan in ITP is a pre-specified, multi-site endpoint and deserves weight. Healthspan batteries are harder: they are often secondary, sex-split, and lesion-scored after the fact (Snyder et al., 2023 is better than most). Human “aging” endpoints in the reviewed record are almost all surrogates (NAD+, HOMA, clocks, vaccine titers). Multiplicity is unaddressed in small NAD and resveratrol trials. Observational metformin (Bannister et al., 2014) and dietary spermidine (Kiechl et al., 2018) cannot carry causal aging claims. CALERIE’s achieved restriction was less than prescribed (Ravussin et al., 2015); reporting the target as the exposure would be a methods error.

25 Judgment

Scientifically interesting, with organism-level mammalian evidence: rapamycin (both sexes, replicated ITP); acarbose and canagliflozin (ITP, sex-asymmetric); 17-alpha-estradiol (ITP, males); glycine (ITP, modest, both sexes); NDGA and aspirin (early ITP, males). These are longevity-pharmacology facts in mice. They are not demonstrated human CR mimetics.

Meaningful human evidence exists for disease treatment (metformin, acarbose, canagliflozin, rapalogs in transplant and selected older-adult immune trials) and for CR itself as a metabolic intervention (CALERIE; voluntary CR cohorts). Meaningful human evidence that a proposed mimetic delayed biological aging in otherwise healthy people does not exist in the reviewed record.

Primarily longevity-industry hypotheses, on present organism-level and human-outcome evidence: resveratrol as a CR mimetic; NAD boosters as aging reversers; spermidine capsules as CR mimetics; commercial AKG plus a clock; metformin as an anti-aging drug; any stack that treats a pathway diagram as an outcome.

The term “caloric-restriction mimetic” is justified for an agent only when organism-level CR effects are reproduced without sustained CR. By that definition, the field has candidates and cousins, not a completed human mimetic.

26 Evidence handling

Findings are labelled by study type in the reporting sentence. Animal and in-vitro results are not phrased as human outcomes. Conflicting evidence is left in conflict when a later null (ITP metformin; ITP resveratrol; NIA 2012 primate survival) does not vanish because an earlier positive exists. Amounts are experimental parameters. Project-internal corpus gates, scan counts and pipeline names are filed in project notes, not in this apparatus.

Non-PubMed sources used for institutional facts only: the NIA Interventions Testing Program public page; FDA prescribing-information pages for metformin, sirolimus and acarbose, as current labels for risk class, not as aging-trial results.

References