Inflammaging

Inflammaging

Chronic age-associated inflammatory signaling and immune-system remodeling — causes, consequences, adaptations, and the limits of the construct

Inflammaging is a name for a pattern, not a diagnosis. Circulating cytokines rise with age in many cohorts and fall when fat mass, infection, or senescent-cell burden is reduced. Whether the signal is a cause of aging, a consequence of damaged tissue, an adaptive defence, or some mixture of the three is the question this article grades. It is not a treatment protocol.

Compiled by South Beach Longevity · 20 August 2026 Copyright 2026 Series SBL-41 / SP-INFLAMMAGING · Register A scientific article Sources peer-reviewed human trials, cohorts, Mendelian randomisation, consensus statements, meta-analyses, and labelled animal or in-vitro 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 circulating interleukin is not tissue inflammation. A mouse knockout is not a human aging outcome. A cardiovascular event trial is not a longevity trial. Where two results conflict, both are given. Amounts and durations appear only as reported experimental parameters, always with the population attached. Nothing here is a recommendation. Findings are graded in place as established, strongly supported, emerging, plausible, or speculative. 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.

---

01 What inflammaging is, and four things it is not

Inflammaging, originally written inflamm-aging, is Claudio Franceschi’s name for a progressive, low-grade, systemic proinflammatory status that accompanies aging and is attributed to a continuous antigenic and stress load (Franceschi, Bonafè, Valensin et al., 2000). The 2000 paper is a theoretical essay, not a measurement study. It placed the macrophage at the centre of an evolutionary argument: the same innate machinery that kept a short-lived ancestor alive becomes, in a longer-lived human, a source of chronic tissue injury. That framing is established as the origin of the term. It is not, and was not claimed to be, a diagnostic criterion.

Two later reviews restated the construct in clinical language. Ferrucci and Fabbri defined inflammageing as elevated blood inflammatory markers in older people, carrying susceptibility to chronic morbidity, disability, frailty, and premature death, and listed candidate sources: genetic susceptibility, central obesity, gut permeability, microbiota change, cellular senescence, NLRP3 activation, mitochondrial oxidative stress, immune-cell dysregulation, and chronic infection (Ferrucci and Fabbri, 2018). Franceschi, Garagnani, Parini, Giuliani, and Santoro then treated inflammaging as sterile, chronic, low-grade inflammation that shares mechanisms with metaflammation, the metabolic inflammation of nutrient excess (Franceschi et al., 2018). Furman, Campisi, Verdin and colleagues widened the frame again to systemic chronic inflammation across the life span, driven by infection, inactivity, diet, toxicants, and psychological stress (Furman et al., 2019). These are syntheses. They are strongly supported as maps of a literature. They are not a licence to treat every elevated C-reactive protein (CRP) in an older adult as a single disease.

Four things follow immediately and will be enforced for the rest of this document.

First, inflammaging is not acute inflammation. A fever, an abscess, a cytokine storm, and a septic cascade are high-amplitude, time-limited innate responses. The aging literature is about a small, persistent shift in the same molecules. Collapsing the two is how a useful macrophage becomes a slogan.

Second, it is not autoimmune disease and it is not infection. Rheumatoid arthritis, lupus, and untreated HIV raise the same cytokines. They are not aging. Chronic viral carriage, especially cytomegalovirus (CMV), can look like aging of the T-cell compartment (Nikolich-Žugich, 2018). That overlap is a method problem, not a synonym.

Third, it is not localized tissue inflammation. A synovial biopsy, a coronary plaque, a senescent niche in fat, and a plasma interleukin-6 (IL-6) concentration are four different objects. Section 19 is written to keep them apart. A blood test that predicts death is not a map of the joint.

Fourth, it is not a single cytokine, a single pathway, or a single clock. IL-6, tumour necrosis factor (TNF), CRP, CXCL9, the NLRP3 inflammasome, cGAS–STING, and the senescence-associated secretory phenotype (SASP) appear in the same reviews because they co-travel, not because they are one mechanism (Ferrucci and Fabbri, 2018; Sayed et al., 2021). A composite inflammatory age that tracks multimorbidity is a useful score. It is not a cause.

The working definition used here is therefore modest. Inflammaging is the empirical observation that, in many human cohorts, circulating inflammatory markers rise with chronological age and with the diseases of later life, together with the set of innate, adaptive, metabolic, microbial, and senescent mechanisms proposed to explain that rise. The word is a research programme. It is not a diagnosis, and this document will not treat it as one.

02 History: from network theory to a hallmark

Franceschi’s group had already described immunosenescence as a remodelling, not a simple decline, before they named inflamm-aging (Franceschi, Bonafè, Valensin et al., 2000). The 2000 paper grafted that remodelling onto an evolutionary stress-response argument. The macrophage, not the T cell, was the load-bearing cell. Later reviews kept the macrophage and added debris, nutrients, and microbiota as equivalent “quasi-self” stimuli sensed by a degenerate receptor set (Franceschi et al., 2018). Fulop, Larbi, Dupuis and colleagues asked whether immunosenescence and inflamm-aging were two sides of one coin, and whether both were always foes; they recorded the possibility that some of the remodelling is adaptive (Fulop et al., 2017). That question is not decorative. It is the adversarial core of Part Five.

Geroscience then absorbed the construct. Kennedy, Berger, Brunet, Campisi and colleagues named the programme that treats aging itself as the shared risk factor for chronic disease (Kennedy et al., 2014). López-Otín, Blasco, Partridge, Serrano, and Kroemer’s 2013 hallmarks did not yet list chronic inflammation as a separate hallmark; altered intercellular communication carried part of the load (López-Otín et al., 2013). The 2023 expansion added chronic inflammation and dysbiosis as hallmarks in their own right, with the three-premise test: age-associated manifestation, experimental acceleration, and the possibility of deceleration by intervention (López-Otín et al., 2023). Inflammaging therefore entered the hallmark table as a claim that still has to earn the third premise in humans. Most of the deceleration evidence is still animal, surrogate, or disease-specific. That asymmetry is established as a description of the literature. It is not a reason to ignore the first two premises.

Clocks followed the reviews. Horvath’s DNA-methylation age is not an inflammatory clock (Horvath, 2013). Levine’s PhenoAge mixes clinical chemistry, including CRP, into a mortality predictor (Levine et al., 2018). Alpert, Pickman, Leipold and colleagues derived IMM-AGE from longitudinal cellular frequencies in 135 adults sampled over nine years and reported that the trajectory predicted all-cause mortality better than chronological age (Alpert et al., 2019). Sayed, Huang, Nguyen and colleagues trained iAge on the blood immunome of 1,001 people aged 8–96 years; the score tracked multimorbidity, immunosenescence, frailty, and cardiovascular aging, and CXCL9 was the strongest contributor (Sayed et al., 2021). These are strongly supported as descriptions of covariance. They are emerging as causal instruments. A clock that contains CXCL9 will move when endothelium ages. That does not prove CXCL9 ages the person.

03 Immunosenescence versus inflammaging

Immunosenescence is the age-related remodelling of immune cells, soluble guides, and lymphoid organs that increases vulnerability to infection and impairs vaccine response (Nikolich-Žugich, 2018). Inflammaging is the cytokine and acute-phase side of the same decades. They are correlated. They are not identical. A person can have a CMV-driven CD8 expansion with a quiet CRP, or an obese CRP with a relatively intact naive T-cell pool. Treating the words as synonyms is how a T-cell repertoire paper becomes a statin advertisement.

Nikolich-Žugich’s review is the current map of adaptive aging: thymic involution, reduced naive output, oligoclonal memory expansion, impaired lymph-node architecture, and an exposome — including CMV — that rewrites the rules of the system (Nikolich-Žugich, 2018). Fulop and colleagues insisted that some of this remodelling may be the price of a long antigenic life rather than a defect (Fulop et al., 2017). Franceschi’s original essay had already said the same thing in evolutionary language (Franceschi, Bonafè, Valensin et al., 2000). The grade for “immune aging is remodelling, not simple immunodeficiency” is strongly supported. The grade for “the remodelling is therefore beneficial” is plausible in selected functions and not established as a general claim.

The practical consequence for this article is methodological. Studies that report “inflammaging” from a single IL-6 draw in a clinic cohort have not measured immunosenescence. Studies that report CD28-null T cells have not measured adipose IL-6. Both may be true in the same person. The citation must say which object was measured.

04 How inflammation is measured — a methods review

The aging-inflammation literature is a methods literature wearing a biology costume. The objects in circulation are not interchangeable.

CRP is an acute-phase pentraxin induced in hepatocytes largely by IL-6 (Gabay and Kushner, 1999; Pepys and Hirschfield, 2003). High-sensitivity assays report values in the 0.1–10 mg·L⁻¹ band that cardiovascular epidemiology uses. A clinical CRP of 40 mg·L⁻¹ is infection or injury. An epidemiologic CRP of 2.2 mg·L⁻¹ is a risk marker. Using one word for both is a category error (Pepys and Hirschfield, 2003).

IL-6 is a cytokine with classic signalling through membrane IL-6 receptor plus gp130, and trans-signalling through soluble IL-6 receptor (Hunter and Jones, 2015). It is produced by immune cells, adipocytes, myocytes, endothelial cells, and senescent cells. A plasma IL-6 of 2 pg·mL⁻¹ in a fasting older adult is not the same molecule, in the same place, as the IL-6 that rises during a marathon. Pedersen and Febbraio treated contracting muscle as a secretory organ precisely to keep those two IL-6s apart (Pedersen and Febbraio, 2012).

TNF is a transmembrane and soluble cytokine with two receptors and a large downstream death-and-NFκB repertoire. Circulating TNF in aging cohorts is often near the floor of the assay. Soluble TNF receptors are sometimes used as more stable proxies (Varadhan et al., 2014). A near-undetectable TNF is not evidence that TNF signalling is idle in a plaque.

Interferons. Type I interferon reports on nucleic-acid sensing, including cGAS–STING. It is not a routine inflammaging panel. An interferon-stimulated-gene signature in a subset of older adults is a different claim from a raised CRP (Ablasser and Chen, 2019).

Composite scores. Varadhan, Yao, Matteini and colleagues reduced a nuclear-factor-κB-related panel in InCHIANTI and the Cardiovascular Health Study to a weighted inflammatory index that predicted 10-year mortality; CRP, IL-1 receptor antagonist, IL-6, IL-18, and soluble TNF receptor-1 were independent 5-year mortality predictors in the discovery set (Varadhan et al., 2014). iAge and IMM-AGE are later, higher-dimensional versions of the same instinct (Alpert et al., 2019; Sayed et al., 2021). Composites improve prediction. They hide which pathway moved.

Confounding that will not go away. Body-mass index, visceral fat, smoking, infection on the day of the draw, assay batch, circadian time, sleep the night before, and residual kidney function all move CRP and IL-6 (Gabay and Kushner, 1999; Meier-Ewert et al., 2004; Minihane et al., 2015). A paper that reports “age-adjusted IL-6” without a fat-mass term has not isolated aging. A paper that reports “inflammation” from a biobank CRP without excluding acute illness has not isolated inflammaging. Reverse causation is ordinary: people who are already ill have higher cytokines, and the cytokines then “predict” the illness that produced them. Mendelian randomisation and randomised trials exist because the cohort association cannot settle that (IL6R Genetics Consortium and Emerging Risk Factors Collaboration, 2012; Ridker et al., 2017).

What a significant p-value is not. In a cohort of several thousand, a 0.3 pg·mL⁻¹ IL-6 difference will be statistically significant. It may still be smaller than the intra-individual day-to-day variance. Effect sizes, not stars, are what this document reports.

05 Innate immunity remodeled

The innate side of immune aging is a change in set-point, not a loss of the system. Myeloid cells remain numerous. Their output becomes noisier. Shaw, Goldstein and the reviews that followed describe impaired pattern-recognition responses to new infection together with a higher tonic production of inflammatory mediators — the combination that makes an older adult both more inflamed and more infection-prone (Nikolich-Žugich, 2018; Ferrucci and Fabbri, 2018). That combination is strongly supported as a description. It is not a single molecular lesion.

Macrophages sit where Franceschi put them (Franceschi, Bonafè, Valensin et al., 2000). In adipose tissue they accumulate with obesity and switch phenotype (Weisberg et al., 2003). In plaques they foam and secrete. In senescent niches they respond to SASP chemokines (Coppé et al., 2010). In CHIP, a mutant myeloid clone can over-produce IL-1β through NLRP3 (Fuster et al., 2017). These are different macrophages. “The macrophage” in an inflammaging sentence is usually a mixture of those jobs.

Neutrophils and natural killer cells change with age in ways this document will not pretend to have harvested as primary data. They belong in a complete immunology of aging. They are not load-bearing for the biomarker-and-outcome claims that follow.

Innate sensors that are load-bearing here are NLRP3 (section 12), cGAS–STING (section 11), and the Toll-like recognition of bacterial lipopolysaccharide that Cani and colleagues used to define metabolic endotoxemia in mice (Cani et al., 2007). The last is labelled animal until a human experiment measures the same loop.

06 Adaptive immunity remodeled

Thymic involution reduces naive T-cell output. The peripheral repertoire becomes oligoclonal, skewed toward memory, and, in CMV-seropositive people, occupied by large virus-specific expansions (Nikolich-Žugich, 2018). B-cell diversity and the quality of antibody responses to new antigens decline; this is why influenza and SARS-CoV-2 vaccine papers in older adults are an immunosenescence literature before they are a product literature. These statements are strongly supported as directional descriptions of a large human observational and vaccine-response record. They are not a claim that every older adult is immunosuppressed.

The adaptive–innate joint is where inflammaging talk becomes sloppy. Senescent-like T cells can secrete inflammatory cytokines. That does not make every raised IL-6 a T-cell product. Most circulating IL-6 in an older adult with central obesity is still more economically explained by fat and endothelium than by a CD8 clone (Hotamisligil, 2006; Franceschi et al., 2018). CMV can deepen the clone. Obesity can raise the IL-6. A study that measures only one will attribute the other to “aging.”

Duggal, Pollock, Lazarus, Harridge, and Lord reported that several features of immunesenescence, including reduced thymic output, were ameliorated in older adults with high levels of physical activity (Duggal et al., 2018). That is a cross-sectional comparison of highly active older adults with less active peers, not a randomised reassignment of the thymus. It is emerging as evidence that the adaptive phenotype is not a pure calendar function. It is not a training prescription.

---

07 Cytokines as a class

A cytokine is a short-range or endocrine peptide signal. In aging papers it is usually a plasma concentration. Those are different claims. Hunter and Jones described IL-6 as a keystone cytokine with context-dependent pro- and anti-inflammatory properties (Hunter and Jones, 2015). The same caution applies to the class. A raised cytokine can mean production, reduced clearance, assay artefact, or a spill from a local site. It does not, by itself, mean the person is “inflamed” in every tissue.

The aging panel is historically IL-6, TNF, and CRP, sometimes IL-1β, IL-18, IL-1 receptor antagonist, and soluble TNF receptors (Harris et al., 1999; Varadhan et al., 2014). CXCL9 entered with iAge (Sayed et al., 2021). Chemokines that recruit myeloid cells into fat or plaque are mechanistically important and poorly represented in routine biobanks. The panel is a convenience sample of what was easy to measure in the 1990s, not a complete innate census.

08 IL-6

IL-6 is the load-bearing cytokine of inflammaging epidemiology and the most genetically interrogated. Hepatocytes make CRP when they see it (Gabay and Kushner, 1999). Adipocytes and stromal cells make it in obesity (Hotamisligil, 2006). Myocytes make it during contraction, which is why an exercise bout raises IL-6 without being a disease (Pedersen and Febbraio, 2012). Senescent cells make it as part of the SASP (Coppé et al., 2008; Coppé et al., 2010). Classic signalling and trans-signalling do not have the same cellular targets (Hunter and Jones, 2015). A plasma number does not say which route was used.

Harris, Ferrucci, Tracy and colleagues followed 1,293 nondisabled participants in the Iowa 65+ Rural Health Study for a mean of 4.6 years. The highest IL-6 quartile (≥3.19 pg·mL⁻¹) had a relative risk of death of 1.9 (95% CI 1.2–3.1) versus the lowest; CRP ≥2.78 mg·L⁻¹ had RR 1.6 (1.0–2.6); elevation of both had RR 2.6 (Harris et al., 1999). That is established as an association in that cohort. It is not a demonstration that IL-6 killed them.

The IL6R Genetics Consortium and Emerging Risk Factors Collaboration studied the functional Asp358Ala variant (rs2228145) in 125,222 participants for biomarkers and in 51,441 coronary-heart-disease cases versus 136,226 controls. Each inherited 358Ala copy raised mean IL-6 receptor by 34.3% and IL-6 by 14.6%, lowered CRP by 7.5% and fibrinogen by 1.0%, and reduced coronary-heart-disease risk by 3.4% (1.8–5.0), without moving lipids, blood pressure, adiposity, dysglycaemia, or smoking (IL6R Genetics Consortium and Emerging Risk Factors Collaboration, 2012). That pattern is strongly supported as evidence that IL-6 receptor signalling is causal for coronary disease, not merely a passenger with obesity. It is not evidence that lowering plasma IL-6 by any means, in any population, extends life.

IL-6 blockade is a licensed strategy in rheumatoid disease. That clinical literature is not an aging literature and is not imported here as a longevity result.

09 TNF

TNF is older than the inflammaging coinage and weaker as a circulating aging biomarker. It is easier to measure in tissue, synovial fluid, and experimental systems than in the picogram plasma of a community-dwelling older adult. Where soluble TNF receptors enter composite scores, they predict mortality as part of a panel, not as a solo diagnostic (Varadhan et al., 2014). Anti-TNF drugs changed rheumatoid arthritis and inflammatory bowel disease. They are not evidence that TNF is the clock of aging, and they carry infection risk that any aging discussion has to keep in the same sentence.

Mechanistically, TNF sits upstream of NFκB programmes that also appear in SASP and in adipose macrophages (Coppé et al., 2010; Hotamisligil, 2006). Thevaranjan, Puchta, Schulz and colleagues, in mice, found that germ-free animals did not show the usual age-related rise in circulating proinflammatory cytokines, and that TNF-deficient mice were protected from age-associated inflammation and from the microbiota shift that accompanied it (Thevaranjan et al., 2017). That is a labelled animal loop — microbiota, permeability, macrophage function, TNF — not a human TNF-level finding. Grade for the mouse loop: strongly supported in that design. Grade for “TNF causes human inflammaging”: plausible, not established.

10 CRP and the acute-phase

CRP is a downstream hepatic readout. It is not an upstream cause of atherosclerosis in the genetic record that matters here. Pepys and Hirschfield’s critical update remains the biochemical map (Pepys and Hirschfield, 2003). Gabay and Kushner remain the acute-phase map (Gabay and Kushner, 1999). Epidemiology then asked the risk question.

The Emerging Risk Factors Collaboration meta-analysed individual records of 160,309 people without prior vascular disease. Per 1-SD higher log CRP (a threefold higher concentration), the age-and-sex-adjusted risk ratio for coronary heart disease was 1.63 (1.51–1.76) and 1.37 (1.27–1.48) after conventional risk-factor adjustment; further adjustment for fibrinogen moved the coronary ratio to 1.23 (1.07–1.42). Similar attenuations were seen for ischaemic stroke and vascular death. Non-vascular mortality remained associated (Emerging Risk Factors Collaboration, 2010). The interpretation in that paper is the one this document keeps: CRP has continuous associations with several deaths; the vascular associations depend considerably on conventional risk factors and other inflammatory markers. Grade: established as a risk marker; not established as a causal vascular target.

JUPITER randomised 17,802 apparently healthy men and women with LDL cholesterol <130 mg·dL⁻¹ and hs-CRP ≥2.0 mg·L⁻¹ to rosuvastatin 20 mg daily or placebo. The trial stopped after a median 1.9 years. Rosuvastatin reduced LDL cholesterol by 50% and hs-CRP by 37%. The primary end-point rates were 0.77 versus 1.36 per 100 person-years (HR 0.56, 95% CI 0.46–0.69) (Ridker et al., 2008). JUPITER is established as evidence that a statin reduces events in people selected on CRP as well as on LDL. It is not a clean test of CRP lowering, because LDL fell with it. Ridker later wrote the “upstream” essay that moved the field from CRP to IL-6 to IL-1 (Ridker, 2016). That essay is a programme, not a trial.

11 Interferon signalling and cGAS–STING

Cyclic GMP–AMP synthase (cGAS) senses cytosolic DNA and produces cGAMP, which activates STING and a type I interferon programme (Sun, Wu, Du, Chen, and Chen, 2013; Ablasser and Chen, 2019). Glück, Guey, Gulen and colleagues showed that innate sensing of cytosolic chromatin fragments through cGAS promotes senescence (Glück et al., 2017). West and Shadel reviewed mitochondrial DNA as an innate ligand (West and Shadel, 2017). Together these papers give a labelled mechanism by which damaged nuclei and leaking mitochondria can look like a viral infection to the cell that owns them.

In human aging this pathway is emerging. Interferon-stimulated signatures appear in some senescent and some older-adult datasets. They are not the routine CRP literature. A person can have inflammaging by IL-6/CRP without a measurable type I interferon pulse. Treating cGAS–STING as “the” inflammaging pathway is a category error of the opposite kind from ignoring it.

12 Inflammasomes and NLRP3

The NLRP3 inflammasome assembles in response to a wide set of damage-associated signals and activates caspase-1 to mature IL-1β and IL-18 (Swanson, Deng, and Ting, 2019). Vandanmagsar, Youm, Ravussin and colleagues reported that the NLRP3 inflammasome instigates obesity-induced inflammation and insulin resistance, and that weight loss in obese people with type 2 diabetes reduced adipose NLRP3 expression together with inflammation (Vandanmagsar et al., 2011). Youm, Grant, McCabe and colleagues then showed, in mice, that Nlrp3 ablation protected against age-related innate activation, improved glycemic control, and attenuated bone loss and thymic demise; IL-1 accounted for only part of the cognitive benefit (Youm et al., 2013). These are strongly supported animal and human-adipose findings. They are not a demonstration that NLRP3 is the clock of human aging.

CANTOS later inhibited IL-1β, a principal NLRP3 product, in people with prior myocardial infarction (Ridker et al., 2017). That trial is discussed in section 32. It is a disease-event trial, not an NLRP3-aging trial. CHIP-associated TET2-deficient macrophages over-secrete IL-1β through NLRP3 in mice, and an NLRP3 inhibitor was more atheroprotective in that chimera (Fuster et al., 2017). That is a labelled mechanistic bridge from a human clone to an inflammasome, still one step short of a human aging intervention.

13 Senescent cells and the SASP

Cellular senescence is a stable arrest. The SASP is the inflammatory and matrix-remodelling secretome that makes a senescent cell a neighbourhood event (Coppé et al., 2008; Coppé et al., 2010). Coppé, Patil, Rodier and colleagues showed that human cells induced to senesce by genotoxic stress secrete factors associated with inflammation and malignancy, that the SASP develops slowly, and that a similar secretome appears in vivo after DNA-damaging chemotherapy (Coppé et al., 2008). The 2010 review named the dark side of tumour suppression: the same arrest that blocks cancer can promote it through the secretome (Coppé et al., 2010). Basisty, Kale, Jeon and colleagues later built a proteomic atlas of SASPs for biomarker work (Basisty et al., 2020). Gorgoulis, Adams, Alimonti and colleagues attempted a consensus path forward for the field (Gorgoulis et al., 2019). Grade for “senescent cells secrete inflammatory mediators”: established in cell systems and supported in tissues. Grade for “the SASP is the main source of human plasma IL-6”: plausible, not measured at that scale.

Causality in mice is stronger than in people. Baker, Wijshake, Tchkonia and colleagues used INK-ATTAC to clear p16^Ink4a-positive cells and delayed ageing-associated disorders (Baker et al., 2011). Baker, Childs, Durik and colleagues then showed that clearing naturally occurring p16^Ink4a-positive cells from one year of age extended median lifespan in male and female mice of two backgrounds (Baker et al., 2016). Xu, Pirtskhalava, Farr and colleagues transplanted senescent cells into mice, spread senescence, impaired physical function, and shortened survival; dasatinib plus quercetin reduced senescent-cell burden (Xu et al., 2018). Demaria, Ohtani, Youssef and colleagues showed the other face: senescent fibroblasts and endothelial cells appear early in a cutaneous wound and accelerate closure through PDGF-AA; eliminating them delayed healing (Demaria et al., 2014). Senescence is therefore established in mice as both a source of pathology and a requirement for optimal repair. Human senolytic trials remain small, disease-selected, and surrogate-heavy. They are not an aging outcome literature.

14 Mitochondria and DAMPs

Damaged mitochondria release DNA and formyl peptides that innate sensors treat as infection (West and Shadel, 2017). That is a damage-associated molecular pattern (DAMP) argument. It connects mitochondrial dysfunction — a hallmark in both López-Otín lists (López-Otín et al., 2013; López-Otín et al., 2023) — to cGAS–STING and, in some models, to NLRP3 (Swanson, Deng, and Ting, 2019). In human plasma the DAMP is rarely the analyte. The analyte is still CRP. The mechanistic story is plausible and, in defined experimental systems, strongly supported. The epidemiologic story is downstream and indirect.

---

15 Adipose tissue and metabolic dysfunction

Obesity is an inflammatory organ state. Weisberg, McCann, Desai and colleagues showed that adipose tissue in obesity accumulates macrophages (Weisberg et al., 2003). Hotamisligil reviewed the broader coupling of metabolic disease to inflammation (Hotamisligil, 2006). Franceschi’s later review treated metaflammation and inflammaging as mechanistically related responses to nutrient excess and to aging debris (Franceschi et al., 2018). This is established as human histopathology plus a large biomarker literature: fat mass moves IL-6 and CRP.

The wrong equation is “age-associated IL-6 rise = aging.” A substantial fraction of that rise in industrial populations is the adipose fraction. Minihane, Vinoy, Russell and colleagues reviewed diet, low-grade inflammation, and health and treated body composition as a first-order term (Minihane et al., 2015). A study that “adjusts for BMI” has not fully captured visceral fat, fatty liver, or fitness. Residual confounding is not a footnote. It is a rival hypothesis, and it is often the better one.

Vandanmagsar and colleagues connected adipose NLRP3 to human insulin resistance and to the fall in adipose NLRP3 after diet-and-exercise weight loss in type 2 diabetes (Vandanmagsar et al., 2011). That is one of the cleaner human-tissue measurements in this document. It still does not make NLRP3 an aging clock.

16 Gut permeability and the microbiome

Two literatures share a border. One is human descriptive: the gut microbiota of older adults differs from that of younger adults and tracks diet and frailty. Biagi, Nylund, Candela and colleagues described microbiota and inflammatory status across seniors and centenarians (Biagi et al., 2010). Claesson, Jeffery, Conde and colleagues reported that gut microbiota composition correlates with diet and health in the elderly (Claesson et al., 2012). Ghosh, Rampelli, Jeffery and colleagues, in the NU-AGE trial, reported that a Mediterranean-diet intervention in older people altered the gut microbiome and was associated with reduced frailty and improved health status over one year (Ghosh et al., 2020). These are strongly supported as descriptions of covariance and, for NU-AGE, as a diet-microbiome-frailty package. They are not a proof that a named taxon causes inflammaging.

The other literature is mechanistic and mostly animal. Cani, Amar, Iglesias and colleagues defined metabolic endotoxemia in mice: a high-fat diet raised plasma lipopolysaccharide two- to threefold, and infusing LPS reproduced fasting glycaemia, insulinaemia, and weight-gain features of the diet (Cani et al., 2007). Thevaranjan and colleagues showed that germ-free mice did not develop the usual age-related rise in circulating proinflammatory cytokines, that co-housing with old conventional mice transferred the inflammatory signal, and that TNF and the microbiota were mutually dependent (Thevaranjan et al., 2017). Grade for the mouse permeability–LPS–TNF loop: strongly supported in those designs. Grade for “leaky gut causes human inflammaging”: plausible, under-measured as permeability in the same people who provide the CRP.

17 Chronic and latent infection

A continuous antigenic load was already in the 2000 definition (Franceschi, Bonafè, Valensin et al., 2000). CMV is the best-studied latent passenger. It expands memory CD8 clones and is part of the immunosenescence phenotype (Nikolich-Žugich, 2018). Periodontal disease, chronic viral hepatitis, untreated HIV, and recurrent urinary infection are ordinary sources of systemic inflammatory markers that have nothing conceptually to do with aging and everything to do with the age-associated rise in those conditions. A cohort that does not serotype CMV or ask about teeth will attribute their variance to birthdays.

This is a methods point with a clinical face. Infection control and vaccination (section 31) are not “anti-inflammaging drugs.” They are ways of reducing the antigenic load that the original essay named.

18 Clonal hematopoiesis

Clonal hematopoiesis of indeterminate potential (CHIP) is an expanded somatic blood-cell clone in a person without a hematologic diagnosis. Jaiswal, Fontanillas, Flannick and colleagues analysed whole-exome sequences from 17,182 persons. Clonal mutations were rare before 40 and rose to 9.5% at 70–79, 11.7% at 80–89, and 18.4% at 90–108, mostly in DNMT3A, TET2, and ASXL1. Mutation presence was associated with hematologic cancer (HR 11.1, 95% CI 3.9–32.6), all-cause mortality (HR 1.4, 1.1–1.8), incident coronary heart disease (HR 2.0, 1.2–3.4), and ischemic stroke (HR 2.6, 1.4–4.8) (Jaiswal et al., 2014). Jaiswal, Natarajan, Silver and colleagues then showed, in 4,726 coronary-disease cases and 3,529 controls, that CHIP carriers had 1.9 times the coronary-heart-disease risk of noncarriers (95% CI 1.4–2.7) in prospective nested analyses, and 4.0 times the risk of early-onset myocardial infarction (2.4–6.7); DNMT3A, TET2, ASXL1, and JAK2 were each associated (Jaiswal et al., 2017). Fuster, MacLauchlan, Zuriaga and colleagues showed that TET2-deficient hematopoietic cells accelerate atherosclerosis in Ldlr^−/− mice and that the macrophages over-produce NLRP3-dependent IL-1β (Fuster et al., 2017).

CHIP is established as an age-associated, inflammation-relevant clone with a human cardiovascular association and a mouse causal bridge. It is one of the few inflammaging sources that is not obesity in disguise. It is still not “aging.” It is a somatic mutation whose prevalence rises with age.

19 Systemic biomarkers versus localized tissue inflammation

This distinction is the methodological spine of the article.

A plasma CRP is a hepatic echo of IL-6 and of other acute-phase drivers (Gabay and Kushner, 1999). It integrates fat, infection, sleep loss, smoking, and plaque activity into one number (Meier-Ewert et al., 2004; Emerging Risk Factors Collaboration, 2010). A coronary plaque is a tissue with macrophages, lipids, and a necrotic core (Libby, 2021). A SASP is a local secretome (Coppé et al., 2010). Neuroinflammation in Alzheimer disease is a CNS innate state that may be almost invisible in plasma (Heneka et al., 2025). An osteoarthritic synovium can be hot while hs-CRP is unremarkable.

Established: systemic markers associate with distant outcomes (Harris et al., 1999; Emerging Risk Factors Collaboration, 2010). Established: tissue inflammation can exist without a matching blood signal, and a blood signal can exist without a matching tissue biopsy. Not established: that lowering the blood marker is equivalent to treating the tissue. CANTOS lowered hs-CRP and reduced coronary events without lowering lipids (Ridker et al., 2017). CIRT did not lower IL-1β, IL-6, or CRP and did not reduce events (Ridker et al., 2019). Those two trials are the cleanest available demonstration that the pathway, not the word “inflammation,” is what matters — and that the pathway was tested in atherosclerotic disease, not in aging as such.

---

20 Cardiovascular disease

Atherosclerosis is an inflammatory disease in the modern sense: innate and adaptive cells, not merely cholesterol, build and complicate the plaque (Libby, 2021). Inflammaging enters as the claim that the age-associated rise in blood inflammatory markers is part of why age is the dominant cardiovascular risk factor (Ferrucci and Fabbri, 2018).

The observational association is established. Harris et al. (1999) and the Emerging Risk Factors Collaboration (2010) are the prototype blood-marker papers. Jaiswal et al. (2014, 2017) add a clonal myeloid source. The genetic association for IL-6 receptor signalling is strongly supported (IL6R Genetics Consortium and Emerging Risk Factors Collaboration, 2012). The randomised evidence that inhibiting selected inflammatory pathways reduces events in selected coronary populations is established for canakinumab 150 mg in CANTOS and for colchicine 0.5 mg in COLCOT and LoDoCo2, and negative for low-dose methotrexate in CIRT (Ridker et al., 2017; Tardif et al., 2019; Nidorf et al., 2020; Ridker et al., 2019). JUPITER remains a statin trial selected on CRP (Ridker et al., 2008).

What this is not: a demonstration that “treating inflammaging” prevents first coronary disease in unselected older adults, or that CRP-guided anti-inflammatory therapy is a gerotherapeutic. Ferrucci and Fabbri were careful: clinical trials suggest the cardiovascular association is causal; whether modulating inflammation helps non-CVD aging outcomes is controversial (Ferrucci and Fabbri, 2018). That sentence is still the right one.

21 Type 2 diabetes

Pradhan, Manson, Rifai, Buring, and Ridker, in a nested case-control study inside the Women’s Health Study, found that baseline IL-6 and CRP associated with incident type 2 diabetes over four years among 188 cases and 362 controls drawn from 27,628 women (Pradhan et al., 2001). Donath and Shoelson’s later review treated type 2 diabetes as an inflammatory disease (the article is in the reviewed record as a conceptual map, not as a new trial). Hotamisligil (2006) and Vandanmagsar et al. (2011) supply the adipose and NLRP3 mechanisms. Grade for “inflammatory markers predict incident diabetes”: strongly supported. Grade for “inflammation is the cause of type 2 diabetes”: not established. Insulin resistance, β-cell failure, and fat distribution remain the physiology. Inflammation is in the room. It is not the landlord.

Knowler, Barrett-Connor, Fowler and colleagues showed that a lifestyle intervention, or metformin, reduced incident diabetes in the Diabetes Prevention Program (Knowler et al., 2002). That trial moved weight, diet, and activity. It did not isolate an anti-cytokine mechanism.

22 Frailty and sarcopenia

Fried, Tangen, Walston and colleagues operationalised a frailty phenotype in 5,317 Cardiovascular Health Study participants: shrinking, weakness, exhaustion, slowness, and low activity (Fried et al., 2001). Cruz-Jentoft, Bahat, Bauer and colleagues revised the European sarcopenia consensus around low muscle strength as the key characteristic (Cruz-Jentoft et al., 2019). Inflammatory markers associate with both constructs in multiple cohorts (Ferrucci and Fabbri, 2018). That association is strongly supported. Causality is not.

Sarcopenia is a muscle disease with neural, nutritional, hormonal, and disuse terms (Cruz-Jentoft et al., 2019). IL-6 can be catabolic in chronic excess and myogenic in an exercise pulse (Pedersen and Febbraio, 2012). Using the same word for both is how a training adaptation becomes a wasting diagnosis. Sayed et al. (2021) reported that iAge tracked frailty. A clock that tracks frailty is doing its job as a clock. It has not shown that the chemokines in the clock produced the frailty.

23 Neurodegeneration

The Lancet Commission’s 2020 dementia report treats midlife and later-life risk factors — including the cardiometabolic cluster — as the preventable fraction of dementia (Livingston et al., 2020). Inflammation appears in that cluster as a mediator and as a neuropathology. Heneka, van der Flier, Jessen and colleagues reviewed neuroinflammation in Alzheimer disease as a central innate process in brain (Heneka et al., 2025). Plasma CRP is a weak window on that process. People who collapse the two will conclude that a statin or an IL-1 antibody is a dementia drug because it moved hs-CRP. The evidence does not say that.

Grade for “innate inflammation is part of Alzheimer neuropathology”: strongly supported as neuropathology and imaging. Grade for “systemic inflammaging causes Alzheimer disease”: plausible, confounded by shared vascular risk, and not established. Grade for “anti-inflammatory drugs prevent dementia”: not established, and historically disappointing in trials this review did not need to re-litigate in detail to state the negative.

24 Cancer

Coussens and Werb, and Balkwill and Mantovani, restated Virchow’s claim in modern molecular language: inflammation can initiate, promote, and accompany cancer (Balkwill and Mantovani, 2001; Coussens and Werb, 2002). Grivennikov, Greten, and Karin mapped immunity, inflammation, and cancer as a crosstalk inside the tumour (Grivennikov, Greten, and Karin, 2010). Hanahan and Weinberg put tumour-promoting inflammation among the next-generation hallmarks (Hanahan and Weinberg, 2011). The SASP is one labelled mechanism by which a tumour-suppressive arrest becomes tumour-promoting (Coppé et al., 2010).

CANTOS reported a lower incidence of lung cancer as a secondary observation in an IL-1β-inhibition trial designed for coronary events (the lung-cancer paper is adjacent to the primary CANTOS report; the primary report itself is the event trial, Ridker et al., 2017). That is emerging as a cancer signal, not a cancer-prevention registration. Chronic infection and organ-specific inflammation (hepatitis, H. pylori, colitis) remain the strongest human inflammatory-cancer stories. They are not inflammaging. They are inflammation in a place, for a long time.

25 Mortality

Harris et al. (1999) and Varadhan et al. (2014) are the prototype mortality papers in older adults. The Emerging Risk Factors Collaboration (2010) showed that CRP’s mortality associations are not confined to vascular deaths. Alpert et al. (2019) and Sayed et al. (2021) moved the predictor from one or two markers to an immune-age trajectory. CHIP adds a clonal mortality association (Jaiswal et al., 2014). Grade: established that inflammatory markers and immune-age scores predict death. Grade: not established that the markers are the mechanism of dying, as opposed to an integrated readout of the diseases and behaviours that kill.

CANTOS did not reduce all-cause mortality (HR 0.94, 95% CI 0.83–1.06) and increased fatal infection (Ridker et al., 2017). LoDoCo2 reported a higher incidence of noncardiovascular death with colchicine that did not exclude a harm (HR 1.51, 95% CI 0.99–2.31) (Nidorf et al., 2020). Any sentence that says “suppress inflammation, live longer” has to survive those two numbers. It does not.

26 Cause, consequence, adaptation, or mixture

The owner question is the right question. The answer is a mixture, and the mixture is not the same for every pathway.

Cause. IL-6 receptor genetics and CANTOS make a causal case for selected innate signalling in coronary events (IL6R Genetics Consortium and Emerging Risk Factors Collaboration, 2012; Ridker et al., 2017). CHIP plus the TET2 mouse makes a causal case for a mutant myeloid clone in atherosclerosis (Jaiswal et al., 2017; Fuster et al., 2017). Senescent-cell clearance in mice makes a causal case for p16^Ink4a-positive cells in selected aging phenotypes (Baker et al., 2011; Baker et al., 2016; Xu et al., 2018). These are different causes. None is “inflammaging” as a single lever.

Consequence. Damaged tissue, fat mass, infection, sleep loss, and residual disease raise the same markers (Meier-Ewert et al., 2004; Hotamisligil, 2006; Minihane et al., 2015). In that direction the cytokine is a reporter. Treating the reporter without treating the source is how CIRT could be both an anti-inflammatory trial and a null trial: methotrexate did not move IL-1β, IL-6, or CRP (Ridker et al., 2019).

Adaptation. Demaria et al. (2014) made the repair case for senescent cells in wound healing. Fulop et al. (2017) and Franceschi’s original evolutionary essay made the case that a proinflammatory set-point can be the residue of a successful defence (Franceschi, Bonafè, Valensin et al., 2000). Acute IL-6 from muscle is an adaptive myokine, not a disease (Pedersen and Febbraio, 2012). An older immune system that spends clone space on CMV has adapted to CMV. The cost is repertoire. The adaptation is real.

Mixture. Most human aging is the mixture. A CHIP clone (cause) in an obese, CMV-positive, sleep-restricted person (consequences and exposures) whose senescent cells are also trying to close a wound (adaptation) will produce one CRP. The blood test cannot unmix that. The rest of this document exists so the reader does not have to pretend it can.

---

27 Exercise

Regular exercise has anti-inflammatory effects that Gleeson, Bishop, Stensel and colleagues attributed to reduced visceral fat and to an anti-inflammatory environment after each bout — and they noted that the same biology can suppress immunity in elite athletes (Gleeson et al., 2011). Pedersen and Febbraio mapped muscle as a secretory organ (Pedersen and Febbraio, 2012). Duggal et al. (2018) associated high lifelong activity with a milder immunesenescence phenotype. Grade: strongly supported that habitual activity associates with lower inflammatory markers and a less aged immune profile; established that an acute bout raises myokine IL-6. This is not a training plan. The exercise-intervention article in this series owns prescription-shaped evidence. Here the point is mechanistic and epidemiologic only.

28 Weight loss

Esposito, Pontillo, Di Palo and colleagues randomised 120 premenopausal obese women to a lifestyle programme aimed at ≥10% weight loss or to a control group; the intervention improved vascular inflammatory markers (Esposito et al., 2003). Vandanmagsar et al. (2011) reported that calorie restriction and exercise-mediated weight loss in obese people with type 2 diabetes reduced adipose NLRP3 and inflammation. Fontana, Meyer, Klein, and Holloszy reported that long-term calorie restriction in humans improved atherosclerosis risk-factor profiles (Fontana et al., 2004). Grade: strongly supported that fat-mass loss lowers systemic inflammatory markers. That fact is the strongest available human evidence that a large share of “inflammaging” is adipose. It is not a diet prescription.

29 Diet

Minihane et al. (2015) remains the cautious map of diet and low-grade inflammation. PREDIMED was retracted and republished: Estruch, Ros, Salas-Salvadó and colleagues assigned 7,447 high-risk Spaniards to a Mediterranean diet with extra-virgin olive oil, a Mediterranean diet with nuts, or a reduced-fat control; after a median 4.8 years the republished primary analysis remained a reduction in major cardiovascular events (Estruch et al., 2018). Ghosh et al. (2020) linked a Mediterranean pattern in NU-AGE to microbiome and frailty changes. Grade for “dietary pattern can move cardiovascular events and, in older adults, microbiome-frailty packages”: strongly supported in those trials. Grade for “an anti-inflammatory diet treats inflammaging”: not established as a specific cytokine-target claim. Food is not an IL-1 antibody.

CALERIE-adjacent work in the reviewed record includes oxidative-status outputs after two years of calorie restriction (Il’yasova et al., 2018). Those are surrogates. They are not a human lifespan result.

30 Sleep

Meier-Ewert, Ridker, Rifai and colleagues showed that acute total sleep deprivation (88 hours awake) and ten days of partial deprivation (4.2 versus 8.2 hours) raised high-sensitivity CRP in healthy adults (Meier-Ewert et al., 2004). Irwin reviewed the partnership of sleep and inflammation (Irwin, 2019). Grade: strongly supported that experimental sleep loss moves CRP. Grade for “chronic short sleep is a major source of population inflammaging”: plausible and epidemiologically entangled with obesity, depression, and sleep apnoea. A night of lost sleep is not aging. A decade of it may contribute to the biomarker. The experiment cannot last a decade.

31 Vaccination and infection control

If antigenic load is part of the original definition, then reducing preventable infection is part of the biology (Franceschi, Bonafè, Valensin et al., 2000; Nikolich-Žugich, 2018). Vaccine-response impairment in older adults is immunosenescence, not a reason to skip the vaccine literature. This document does not write a schedule. It records that infection is a reversible contributor to the same blood markers that aging papers treat as fate.

32 Anti-inflammatory pharmacology — scientific discussion, not a protocol

Four trials define the current human pharmacology of “inflammation” in vascular disease.

CANTOS. Ridker, Everett, Thuren and colleagues randomised 10,061 patients with prior myocardial infarction and hs-CRP ≥2 mg·L⁻¹ to canakinumab 50, 150, or 300 mg subcutaneously every three months or placebo. At 48 months, median hs-CRP reductions versus placebo were 26, 37, and 41 percentage points; lipids did not fall. At median 3.7 years, primary-end-point rates were 4.50 (placebo), 4.11 (50 mg; HR 0.93, 0.80–1.07), 3.86 (150 mg; HR 0.85, 0.74–0.98), and 3.90 (300 mg; HR 0.86, 0.75–0.99) per 100 person-years. Only 150 mg met the multiplicity-adjusted threshold. Fatal infection rose. All-cause mortality did not fall (HR 0.94, 0.83–1.06) (Ridker et al., 2017). Established: IL-1β inhibition can reduce recurrent coronary events independent of lipids. Established: that reduction is not a longevity result and is purchased with infection.

CIRT. Ridker, Everett, Pradhan and colleagues randomised 4,786 patients with prior infarction or multivessel disease plus diabetes or metabolic syndrome to low-dose methotrexate (target 15–20 mg weekly) or placebo. The trial stopped at median 2.3 years. Methotrexate did not lower IL-1β, IL-6, or CRP. The final primary end point occurred at 4.13 versus 4.31 per 100 person-years (HR 0.96, 0.79–1.16) (Ridker et al., 2019). Established: not every anti-inflammatory is an anti-atherosclerotic, and a drug that does not move the IL-1/IL-6/CRP axis does not reproduce CANTOS.

COLCOT. Tardif, Kouz, Waters and colleagues randomised 4,745 patients within 30 days after myocardial infarction to colchicine 0.5 mg daily or placebo. Median follow-up 22.6 months. Primary end point 5.5% versus 7.1% (HR 0.77, 0.61–0.96). Pneumonia as a serious adverse event was 0.9% versus 0.4% (Tardif et al., 2019).

LoDoCo2. Nidorf, Fiolet, Mosterd and colleagues randomised 5,522 patients with chronic coronary disease to colchicine 0.5 mg daily or placebo. Median follow-up 28.6 months. Primary end point 6.8% versus 9.6% (HR 0.69, 0.57–0.83). Noncardiovascular death 0.7 versus 0.5 per 100 person-years (HR 1.51, 0.99–2.31) (Nidorf et al., 2020).

JUPITER remains the statin-plus-CRP-selection trial (Ridker et al., 2008). It does not isolate inflammation.

Senolytics and metformin. Xu et al. (2018) is mouse. Barzilai, Crandall, Kritchevsky, and Espeland described TAME as a planned metformin aging trial (Barzilai et al., 2016). TAME is not a result. Human senolytic studies in the wider literature remain early. They are not graded here as aging-outcome evidence.

No dose, schedule, or candidate for personal use is recommended. The scientific point is narrower: pathway choice, population, and end point decide whether “anti-inflammatory” does anything, and none of these trials is a test of aging.

33 Five adversarial questions

Is “inflammaging” too nonspecific? Yes, if it is asked to be a diagnosis, a pathway, or a drug target. No, if it is asked to be a name for a recurring empirical cluster: age, blood inflammatory markers, and the diseases of later life, with a short list of plausible sources. Franceschi’s term earned its keep by forcing immunology into geroscience. It loses its keep when a reviewer uses it to avoid saying which molecule, which tissue, and which confounder. This document keeps the word and refuses the vagueness.

Are elevated cytokines causes or disease markers? Both, and the mix is pathway-specific. CRP is mostly a marker (Emerging Risk Factors Collaboration, 2010). IL-6 receptor signalling has a genetic causal claim for coronary disease (IL6R Genetics Consortium and Emerging Risk Factors Collaboration, 2012). IL-1β has a pharmacological causal claim for recurrent coronary events (Ridker et al., 2017). IL-6 as a number in clinic is usually a marker of fat, infection, or illness. A sentence that does not name the design — cohort, Mendelian randomisation, or trial — has not answered the question.

Are inflammatory pathways sometimes adaptive or protective? Yes. Wound healing requires senescent cells and PDGF-AA in the Demaria mouse (Demaria et al., 2014). Acute infection requires the entire cascade. Muscle IL-6 is a myokine (Pedersen and Febbraio, 2012). CANTOS’s fatal-infection signal is the human cost of blunting IL-1β (Ridker et al., 2017). Fulop et al. (2017) were right to refuse a purely negative reading of immunosenescence. Adaptive is not synonymous with harmless. It means the pathway has a job.

Does suppressing inflammation improve aging outcomes? Not as a general statement. Suppressing IL-1β or using colchicine improved selected coronary end points in selected secondary-prevention populations (Ridker et al., 2017; Tardif et al., 2019; Nidorf et al., 2020). Methotrexate did not (Ridker et al., 2019). All-cause mortality did not fall in CANTOS. Noncardiovascular death was not reassuring in LoDoCo2. Mouse senolysis improved function and lifespan (Baker et al., 2016; Xu et al., 2018). Human aging outcomes — frailty reversal, dementia prevention, life extension — are not established for any anti-inflammatory in the reviewed literature. Ferrucci and Fabbri already said the non-CVD half was controversial (Ferrucci and Fabbri, 2018). The trials since 2018 have confirmed the cardiovascular half and have not settled the rest.

How much is obesity and comorbidity rather than aging? Enough that no paper which skips fat mass, smoking, infection, kidney function, and CHIP should be allowed to say “aging” in the result sentence. Weisberg, Hotamisligil, Minihane, Esposito, and Vandanmagsar are the adipose half of the literature (Weisberg et al., 2003; Hotamisligil, 2006; Minihane et al., 2015; Esposito et al., 2003; Vandanmagsar et al., 2011). Jaiswal is the clonal half (Jaiswal et al., 2014; Jaiswal et al., 2017). Meier-Ewert is the sleep half (Meier-Ewert et al., 2004). What remains after those subtractions is not zero — thymic involution, senescent-cell accumulation, and the IL-6 receptor genetic effect do not require obesity — but it is smaller than the word “inflammaging” is often asked to carry. The honest residual is the research programme. The dishonest residual is a birthday that explains a waist.

---

This document describes published research. It does not recommend human use of any compound, food, vaccine, device, or method, and it specifies no dose, route, or schedule for any person. It is not medical advice.

---

Table A. Inflammatory pathway map

Source	Sensor / cell	Mediator	Typical human readout	Best evidence in the reviewed literature	Not this
Senescent cell / SASP	The senescent cell itself	IL-6, IL-8, chemokines, proteases	Tissue secretome; plasma poorly specific	Coppé 2008, 2010; Basisty 2020; Baker 2011, 2016	“All plasma IL-6 is SASP”
Mitochondrial / nuclear DNA	cGAS–STING	Type I interferon, senescence reinforcement	Rarely a biobank CRP	Sun 2013; Glück 2017; West and Shadel 2017	Routine inflammaging panel
Metabolic debris, crystals	NLRP3	IL-1β, IL-18	Adipose expression; plasma IL-1β often low	Vandanmagsar 2011; Youm 2013; Swanson 2019	A licensed aging drug
Adipose macrophages	TLR, NLRP3, NFκB	IL-6, TNF, adipokines	CRP, IL-6 with fat mass	Weisberg 2003; Hotamisligil 2006	Aging independent of waist
Gut LPS / permeability	TLR4 (mouse)	Systemic cytokines	Poorly standardised human permeability	Cani 2007; Thevaranjan 2017 (animal); Biagi 2010; Claesson 2012; Ghosh 2020 (human descriptive / diet)	A single probiotic claim
CHIP	Mutant myeloid / NLRP3	IL-1β	Clone + events, not a cytokine screen	Jaiswal 2014, 2017; Fuster 2017	“Normal” aging marrow
Latent CMV / antigen	Adaptive clones + innate tone	T-cell cytokines; indirect CRP	Serology + repertoire	Nikolich-Žugich 2018; Franceschi 2000	A CMV-negative “pure aging” sample assumed
Sleep loss, acute illness	Hepatic acute-phase	CRP	hs-CRP	Meier-Ewert 2004; Gabay and Kushner 1999	A stable trait

Table B. Biomarker and outcome matrix

Marker	What it is	Outcome association in the reviewed literature	Causal status	Confounders that move it
hs-CRP	Hepatic acute-phase protein	CHD, stroke, vascular and non-vascular death (ERFC 2010); death in Iowa 65+ (Harris 1999)	Marker; not a demonstrated vascular cause	Fat, infection, smoking, sleep, kidney, IL-6
IL-6	Context-dependent cytokine	Death (Harris 1999); incident diabetes (Pradhan 2001); composite indices (Varadhan 2014)	IL6R genetics support CHD causality; plasma level is mixed	Fat, exercise bout, SASP, infection
TNF / sTNFR	Cytokine / stable proxy	Panel mortality (Varadhan 2014)	Human aging causality unproven	Assay floor; illness
CXCL9	Chemokine; top iAge weight	iAge tracks multimorbidity, frailty, CV aging (Sayed 2021)	Emerging; endothelial experiments in that paper	Vascular age itself
IL-1β	Inflammasome product	CANTOS event reduction when inhibited (Ridker 2017)	Causal for recurrent CHD events in that population	Often undetectable in plasma
Immune-age / iAge / IMM-AGE	Composite	Mortality or multimorbidity tracking (Alpert 2019; Sayed 2021)	Predictive; causal unproven	Everything in the panel
CHIP clone	Somatic VAF in blood	Cancer, death, CHD, stroke (Jaiswal 2014, 2017)	Strong observational; mouse TET2 causal for plaque	Sequencing depth; smoking

Table C. Human cohort and trial table

Study	Design	N / population	What was measured	What can be concluded
Iowa 65+ (Harris et al., 1999)	Prospective cohort	1,293 nondisabled older adults; 4.6 y	IL-6, CRP, death	Markers predict death; not why
Women’s Health Study nested (Pradhan et al., 2001)	Nested case-control	188 incident DM / 362 controls	IL-6, CRP	Markers predict diabetes in women
InCHIANTI → CHS index (Varadhan et al., 2014)	Two-cohort composite	1,155 then 5,600	5-marker index, 10-y death	A panel beats a single cytokine for prediction
ERFC (2010)	IPD meta-analysis	160,309; 54 studies	CRP vs vascular and non-vascular events	Association; attenuates with risk factors and fibrinogen
IL6R MR (2012)	Genetic meta-analysis	125,222 biomarkers; 51,441 / 136,226 CHD	Asp358Ala	IL6R signalling causal for CHD
Jaiswal CHIP 2014	Exome cohort	17,182	Clonal mutations, death, CHD, stroke	CHIP is common in late life and adverse
Jaiswal CHIP 2017	Case-control + mouse	4,726 / 3,529	CHIP vs CHD	~1.9× CHD; mouse TET2 accelerates plaque
JUPITER (Ridker et al., 2008)	RCT	17,802; LDL <130, hs-CRP ≥2	Rosuvastatin 20 mg	Events fall; LDL and CRP both fall
CANTOS (Ridker et al., 2017)	RCT	10,061 post-MI, hs-CRP ≥2	Canakinumab	150 mg reduces events; infection; no mortality win
CIRT (Ridker et al., 2019)	RCT	4,786	Low-dose methotrexate	No cytokine move; no event move
COLCOT (Tardif et al., 2019)	RCT	4,745 post-MI	Colchicine 0.5 mg	Events 5.5% vs 7.1%; more pneumonia
LoDoCo2 (Nidorf et al., 2020)	RCT	5,522 chronic CAD	Colchicine 0.5 mg	Events 6.8% vs 9.6%; non-CV death not reassuring
DPP (Knowler et al., 2002)	RCT	Diabetes-prevention cohort	Lifestyle or metformin	Incident diabetes falls; not an IL-6 trial
Esposito 2003	RCT	120 obese premenopausal women	Weight-loss lifestyle	Vascular inflammatory markers fall
PREDIMED republished (Estruch et al., 2018)	RCT	7,447 high CV risk	Mediterranean patterns	Events fall; not a cytokine-target trial
NU-AGE (Ghosh et al., 2020)	Diet intervention	Older adults, 1 y	Microbiome, frailty	Pattern moves microbiome and frailty package
Sayed iAge 2021	Deep-learning immunome	1,001, ages 8–96	iAge, CXCL9	Score tracks multimorbidity; not a cause proof
Alpert IMM-AGE 2019	Longitudinal omics	135 adults, 9 y	Immune trajectory	Trajectory beats calendar age for status

Table D. Intervention evidence table

Lever	Human evidence in the reviewed literature	What moved	Aging outcome?	Grade
Exercise / high activity	Gleeson 2011; Pedersen 2012; Duggal 2018	Markers; immune phenotype (cross-section)	Not a lifespan trial	Strongly supported for markers
Weight loss	Esposito 2003; Vandanmagsar 2011; Fontana 2004	CRP/IL-6; adipose NLRP3	Not a lifespan trial	Strongly supported for markers
Mediterranean pattern	Estruch 2018; Ghosh 2020	Events; microbiome/frailty	Events, not aging-as-such	Strongly supported for CV events
Sleep restoration	Meier-Ewert 2004 (deprivation, not restoration RCT)	CRP rises with loss	No	Strongly supported for the deprivation effect
Vaccination / infection control	Mechanism from Franceschi 2000; Nikolich-Žugich 2018	Antigenic load	Indirect	Plausible as a contributor, not a gerotherapy
Statin in high CRP	JUPITER 2008	LDL + CRP; events	No	Established for events; mixed mechanism
IL-1β antibody	CANTOS 2017	CRP; recurrent CHD events	Mortality null	Established for events in that population
Low-dose methotrexate	CIRT 2019	Nothing on the axis	No	Established negative
Colchicine	COLCOT 2019; LoDoCo2 2020	Recurrent CHD events	Non-CV death open in LoDoCo2	Established for events in those populations
Senolytics	Xu 2018 mouse; human early, not load-bearing here	Senescent-cell burden in mice	Mouse function/lifespan	Animal; human insufficient
Metformin as gerotherapy	Barzilai 2016 (TAME plan); DPP is diabetes	TAME not reported	No	Speculative as aging therapy

Quantities are trial parameters, not recommendations.

Table E. Causality ledger

Claim	Design that could support it	What we have	Verdict
Age raises mean CRP/IL-6	Cohort	Multiple, residual confounding	Association established; pure aging effect unresolved
IL-6 signalling causes CHD	MR + pharmacology	IL6R Asp358Ala; no direct IL-6 mAb aging trial in the reviewed literature	Causal for CHD at the receptor; not for aging
IL-1β causes recurrent CHD events	RCT	CANTOS 150 mg	Causal in that population
CRP causes CHD	MR / drug that hits only CRP	ERFC attenuation; no CRP-only drug	Not established
NLRP3 causes organismal aging	Human genetics / RCT	Mouse Youm 2013; adipose Vandanmagsar 2011	Animal / tissue; human aging unproven
SASP causes organismal aging	Human senolysis outcome	Mouse Baker/Xu; Demaria opposite face	Animal; human insufficient
CHIP causes CHD	Human assoc. + mouse	Jaiswal + Fuster	Strongly supported hybrid
Microbiota cause human inflammaging	Human RCT of microbiota with cytokine primary	Mouse Thevaranjan/Cani; human descriptive	Not established in people
Fat mass causes the biomarker	RCT of weight loss	Esposito; Vandanmagsar	Strongly supported
Sleep loss causes CRP rise	Human experiment	Meier-Ewert 2004	Established acutely
Suppressing “inflammation” extends life	RCT with mortality/healthspan	CANTOS mortality null; LoDoCo2 non-CV death	Not established; some signals adverse

Evidence handling

Study type is named in the reporting sentence. Animal and in-vitro results are not phrased as human outcomes. Consensus and narrative reviews are secondary maps. Meta-analyses inherit the bias of their inputs. Conflict is left as conflict: CANTOS versus CIRT; senescent-cell harm versus senescent-cell repair; CRP as marker versus IL-6R as cause; inflammaging as biology versus inflammaging as the waistline.

Evidence grades used in the prose:

Grade	Meaning here
Established	Repeated human experiments or field measurements, stable direction
Strongly supported	Consistent human trials, genetics, or high-quality syntheses, residual caveats of population or size
Emerging	Coherent human signal, thin or new
Plausible	Mechanism plus weak or indirect human data
Speculative	Pathway cartoon or wrong-population leap

Adversarial resolution

The evidence and argument were tested against six critical perspectives, applied to the peer-reviewed record verified against NCBI. General reference and news databases were consulted only to locate literature, not as evidence.

Lens	Load-bearing objection	Resolution in this text
Construct	The word names everything and therefore nothing	Section 01 and Q1: keep the name; refuse the diagnosis
Marker vs cause	Cytokines are passengers	Section 08, 10, 26, 33: CRP passenger; IL6R/IL-1β not
Adaptation	Inflammation is defence	Section 13 (Demaria), 27 (myokine), CANTOS infection
Pharmacology	Anti-inflammatory drugs will treat aging	Section 32: pathway-specific event trials; mortality unwon
Confounding	It is obesity	Sections 15, 28, 33: a large share is; CHIP and IL6R are not only that
Methods	Blood ≠ tissue	Section 04, 19, Table B

Unresolved science

U1. The residual age effect on IL-6/CRP after visceral fat, fitness, CMV, sleep, kidney function, and CHIP are measured in the same people. U2. Human senolytic outcomes on function, frailty, and survival. U3. Whether interferon / cGAS–STING subsets define a clinically distinct inflammaging endotype. U4. Human gut-permeability measurement adequate to test the Thevaranjan loop. U5. TAME and any true aging-primary anti-inflammatory trial. U6. Whether colchicine’s noncardiovascular-death signal in LoDoCo2 is chance or cost. U7. A pre-registered anti-inflammatory trial with a primary aging endpoint rather than a cardiovascular one.

Glossary

Term	Meaning in this document
Inflammaging	Age-associated low-grade systemic inflammatory signalling; a construct, not a diagnosis
Immunosenescence	Age-related immune remodelling, especially adaptive
SASP	Senescence-associated secretory phenotype
CHIP	Clonal hematopoiesis of indeterminate potential
DAMP	Damage-associated molecular pattern
Metaflammation	Metabolic inflammation of nutrient excess
Classic / trans IL-6	Membrane-receptor versus soluble-receptor IL-6 signalling
hs-CRP	High-sensitivity C-reactive protein

References

The bibliography is verified against NCBI records; in-prose citations are author–year.