Evidence summary: iAge (Inflammatory Age) (Edifice Health)

iAge (Inflammatory Age) is a blood test sold by Edifice Health, a US company [3]. You collect a small blood sample at home with a Tasso+ device, pressed against the upper arm to draw blood from small vessels near the skin into a tube, and post it back [3][4]. The company says samples go to a CLIA-certified laboratory and results take about 2 weeks, or up to 30 business days [3]. The test is for adults aged 18 and over [3][8]. Researchers can also order it, with blood drawn from a vein or at home [6].

A sample report on the company's website, labelled "made-up values for illustration", shows two numbers [3][4]:

• an iAge in years, shown next to your calendar age (in the sample, "4 years below calendar age");
• an iAge Score on a scale from 0 to 10, with a "typical range (2.5 to 7.5)" and the note that a higher score means more inflammation.

It also lists five blood protein levels, each compared with the median for your age, names a "focus protein" with general diet and lifestyle suggestions ("General wellness information, not medical advice"), and charts the score across repeat tests [3][4]. We have not seen a report a customer actually received.

The laboratory measures the levels of five immune signalling proteins in your blood [3][4]. Cytokines are small proteins that immune cells use to signal to each other; chemokines are cytokines that guide immune cells to where they are needed. The five are eotaxin (also called CCL11), MIG (also called CXCL9), GRO-α (also called CXCL1), all chemokines; interferon gamma (IFN-γ), a cytokine; and TRAIL (short for TNF-related apoptosis-inducing ligand), a signalling protein that can trigger cell death [3][16]. Those five levels are the only measurements.

The company's current pages do not say whether serum or plasma (the liquid part of blood, without or with its clotting proteins) is tested; a 2020 company white paper said "from a standard serum blood draw" [11].

Everything else is calculated:

• iAge in years: the output of a formula that turns the five protein levels into an age-like number [4].
• iAge Score (0 to 10): in the sources we reviewed, the company does not explain how it is calculated, how it relates to the iAge in years, or what the "typical range" is based on.
• Comparison with your age group: the company says your inflammation level is compared with "reference values from the Stanford 1,000 Immunomes Project" [4], the research group of about 1,000 people in which iAge was developed [12].

The method comes from a 2021 peer-reviewed study by Stanford researchers, two of whom are co-founders of Edifice [12]. They measured 50 immune proteins in blood drawn from a vein in 1,001 healthy volunteers aged 8 to 96 and trained a deep-learning model to predict calendar age from them, so a person whose proteins look like those of older people gets a higher iAge [12]. The study reports a correlation of 0.78 between iAge and calendar age, and "average reconstruction errors on the test set for prediction of age" of 15.2 years [12]. A model using only five proteins, the five on the company's report, had an error "not significantly different" from the full model; the text gives no separate error value for it [12]. The study's code is public [12].

The test you buy is not simply that published model. Later studies in a French cohort, with a developer as co-author, calculated iAge with a simpler linear formula on the five proteins, calibrated against Stanford samples [14][15]. Two studies that measured iAge with Edifice's technology, one of them designed and run by Edifice, describe "Edifice's proprietary assay" and "the Edifice proprietary machine learning algorithms" [13][21]. In the sources we reviewed, the formula used for the retail test and its 0 to 10 score are not published.

iAge (Inflammatory Age) is currently rated Emerging on Compare [1]. The rating concerns the inflammatory-age result calculated from five blood proteins, meaning the iAge in years compared with your calendar age and the 0 to 10 score worked out from the same proteins, as a marker of age-related health decline such as multiple long-term conditions, frailty and loss of physical function. It does not cover the individual protein levels, which are laboratory measurements, the report's suggestions, the optional consultation, disease risk or diagnosis, or research use of iAge. This page explains that rating. It does not change it.

We rate iAge Emerging for its own inflammatory-age result: the method behind it is published and peer-reviewed, but the studies linking it to health come from its developers, from research they co-authored or from the same cohort calibrated against their samples, most of the findings are small, from a single point in time, or show no link, and in the sources we reviewed no study has checked the home sample, the company's scoring formula or how results vary between repeat tests.

• The developers' own findings are small or indirect. In the 2021 study, iAge was linked with having more long-term conditions in 285 adults over 60, at one point in time, in the group the model was built in [12]. iAge measured in 2010 was linked with a frailty score in 2017 in 29 older adults [12]; the company itself calls this "a research finding, not a seven-year forecast for every person who takes the test" [5]. A finding that centenarians tended to have low scores used a different nine-protein model, measured in plasma on a different platform, in 37 people [12]. The heart and blood-vessel findings concerned one protein, CXCL9, not the iAge score, in 97 people [12]. Death from any cause was tested only with a gene-activity stand-in for iAge, because "there were no sufficient proteomics data to directly estimate iAge in the cohort" [12].
• The one large later cohort found weak or no links. In INSPIRE-T, about 1,000 adults in Toulouse, France, studied with a developer among the authors, iAge was not linked with a combined measure of physical and mental capacity at the start, and its weaker, age-dependent link over three years "should therefore be interpreted with caution", the authors say [15]. Higher iAge went with poorer physical performance mainly at older ages, and with better performance on one test at ages 45 to 66 [14]. DNA-methylation clocks, GrimAge in particular, were more strongly linked with function in the same people [14][15].
• Other analyses showed mostly no link. Other INSPIRE-T papers found no consistent link with diet [16], appetite loss [18] or heart health and risk scores [20], and one with less elastic skin [17]. In 226 prefrail or frail adults with an average age of 82.5, iAge was not linked with daily functioning or capacity, at the start or over about two years [19]. Three of these papers have no developer author, but they analyse the same INSPIRE-T iAge data, calibrated against the developers' Stanford samples [15][16][18], so we do not count them as independent checks.
• No independent replication. In the sources we reviewed, no group without ties to the developers has calculated the iAge score in its own data and confirmed the reported links. A group without disclosed Edifice ties used 15 proteins taken from iAge only as one comparison for telling people with rheumatoid arthritis from healthy people [25]. Studies without Edifice ties do link higher CXCL9, the protein that contributes most to iAge, with older age and, in older adults, with death [26][27][28]. That supports one ingredient of the score, not the score itself.
• No check of the home test. The studies we reviewed used serum from blood taken in a research setting, drawn from a vein where stated [12][13][14][15]. We found no published comparison of the home sample with vein blood, validation of the retail scoring, or repeat-test data.

Our methodology rates a commercial test "on the evidence for that product’s own output, not for the scientific method behind it" [2]. The test fits our Emerging tier, which describes "pilot human studies", "Mechanistic plausibility but limited clinical data", "No independent replication in humans yet" and evidence that "May be based on small sample sizes or vendor-only datasets" [2]. It does not reach Promising, which asks for "At least one independent replication or large-cohort validation" [2]: we found no independent replication, and the large INSPIRE-T cohort found weak or no links and did not test long-term conditions or death. Two criteria in our Insufficient tier also fit, at least in part: "Proprietary models without methodological transparency", since the retail formula is unpublished although the method and its five proteins are published, and "Marketing language exceeds published evidence" (see the next section) [2]. "Vendor-only claims without published validation" and "No peer-reviewed studies or only vendor-funded studies without independent oversight" [2] do not fit a peer-reviewed method tested in a publicly funded academic cohort [12][15]. The Insufficient tier's last criterion, "No independent replication or external validation" [2], overlaps with the Emerging description "No independent replication in humans yet", so we do not treat it as decisive. Our confidence in Emerging is moderate.

An independent study confirming the five-protein iAge's links with long-term conditions, frailty or survival, a published retail formula, or data comparing the home sample with vein blood and repeat samples could change this rating.

• "Validated in independent cohorts". The home page answers "Is it accurate?" with "Yes", adding that "a 2026 Nature Medicine paper describes it as validated in independent cohorts" [3]. That paper is an expert Perspective, not a new study: a framework for choosing immune-aging measures for clinical trials and the XPRIZE Healthspan competition [7][23]. Its competing-interests statement says "D.F. is a founder of Edifice Health, a company that commercializes the iAge test" [23]. We could not read its main text. Its summary does not name iAge, says "there is no consensus on how to quantify immune aging", and reports that "inflammaging scores and functional assays performed best against our selection criteria" [23]. Among studies that use iAge, its reference list cites the 2021 development study, one INSPIRE-T paper, the oat and plasma-exchange studies and the comparison study described on this page [12][13][14][21][25]; none of them is an independent replication of iAge's reported links. A news item from an organisation whose two chief scientific officers are among its authors, one of whom also co-authored the 2021 iAge study [12], says iAge and another measure "emerge as the field’s strongest candidates so far" [24]. We could not check the vendor's wording against the full text, and in the primary studies we reviewed we found no independent cohort that confirmed iAge's reported links.
• Linked with "all-cause mortality". The company's XPRIZE page says iAge is "associated with multimorbidity, frailty, immunosenescence and all-cause mortality" [6]. As set out above, the protein-based score has not been tested against death in the sources we reviewed [12].
• "And it responds". The same page says that in a trial of an oat product, "iAge fell within two weeks in participants with elevated baseline scores" [6]. That study, funded by PepsiCo and designed and run by Edifice, found no significant overall change; the fall was in a subgroup whose cut-off was chosen afterwards from the results [13]. A larger Edifice trial of dietary supplements in 781 people, with the change in iAge as its main outcome, was stopped; its registry entry says "There was no significant difference between placebo and control" and posts no results, and we found no publication of it [22].
• "Clinical-grade" and "actionable insights". The product page calls the test a "clinical-grade analysis" whose results "help identify targeted interventions for optimizing healthspan" [8]. In the sources we reviewed, no study has validated the retail test, or shown that acting on its results improves health.
• A laboratory certificate is not evidence for the score. The "Yes" rests on samples being "analyzed in a CLIA-certified laboratory" [3]. CLIA, the US federal standard for clinical laboratories, concerns laboratory quality, not whether a score says anything about your health.
• Centenarians "40 years lower". A 2020 company white paper, still linked from the publications page, says that in centenarians "iAge averaged 40 years lower than their corresponding chronological age" [7][11]. The peer-reviewed 2021 study does not report that figure; it reports that 13 of 19 centenarians fell into a low-score group, using a different nine-protein model [12].

The company invites you to "Try changes to your routine, then retest to see how your score changes", with "Strategies and tips" on "diet, lifestyle and nutrients" [3]. Anything that moves the five protein levels, including a passing infection, can move the score; that is not evidence that a lower score means better health.

In the sources we reviewed, no published study shows that any diet, supplement or lifestyle change lowers iAge in a way that improves health. Besides the oat and supplement trials above [13][22], a plasma-exchange trial in 42 people, with a developer as co-author, found that changes in iAge "were not correlated with changes in any other omics", the other biological measures taken [21]. This page is not a recommendation to make any particular change.

The company says "Many people test quarterly (every 3 months) to track changes" and sells two- and four-test bundles [3]. In the sources we reviewed, there are no published data on how much an iAge result varies between two samples from the same person, or between two runs of the same sample, and no threshold for a meaningful change. Without those, a change between two results cannot be told apart from normal variation.

Three things make repeat results harder to read:

• Short-term changes. INSPIRE-T researchers note that measuring cytokines once "might have led to overestimation in cases of transient stress such as recent participation in strenuous physical exercise or infection" [14].
• Very low levels. In an independent study using a different laboratory test on plasma, TRAIL, GRO-α and IFN-γ were "undetected in more than 65% of the samples" [29]. It did not use Edifice's laboratory.
• Home sampling. An independent study of a different, dried upper-arm device found cytokine results "were not always consistently and reliably measured" against blood from a vein [30]. It did not test the Tasso+ device used for iAge or the iAge proteins as a set.

The tracking chart shows your score at each test. It records a level each time, not how fast you are aging.

An iAge result is one estimate, from five proteins in one home sample, of how your immune-protein pattern compares with people of different ages in the Stanford group: a gap at one point in time. The gap also depends on your age: in INSPIRE-T, iAge values ranged only from about 30.5 to 82.3 years in adults aged 20 to 102, a narrower spread than calendar age [14]. The INSPIRE-T authors describe iAge as "a metric for systemic chronic sterile inflammation that may reflect one biologically meaningful dimension of the aging process, rather than as a comprehensive surrogate of aging trajectories" [15].

In the sources we reviewed, iAge has not been shown to predict an individual's risk of a specific disease or of death. The company says iAge "has not been cleared or approved by the FDA, is not intended to diagnose, treat, or prevent disease" [3]. The reviewed studies do not show whether using this test to guide decisions improves health.

The 2021 study states that two of its senior authors, David Furman and Mark M. Davis, "are co-founders of Edifice Health, a company that utilizes iAge" [12]; the company names both as co-founders [3][10]. Edifice holds a US patent, granted in January 2026, on a method of scoring inflammation from these proteins; Furman and Davis are among its named inventors [31]. Furman is a co-author of the main INSPIRE-T papers and the frail-adults study [14][15][19][20]. One INSPIRE-T paper discloses his role at Edifice [15]; another, with him as a co-author, states "The authors declare no conflicts of interest" [14]. The oat study was funded by a food company and carried out by Edifice, with four Edifice employees among the authors [13]. The 2026 expert framework is co-authored by Furman, whose Edifice role it discloses, and Edifice is the official testing partner for the competition it discusses [6][7][23]. The company's publications page itself says "Wider literature is included for context and is not presented as validation of the iAge® test" [7].

iAge was built in healthy volunteers recruited at Stanford, and INSPIRE-T participants were "predominantly White", which the authors say "limits generalizability" [12][15]. The company's terms allow samples to be shared with "third party certified laboratories" [9].

For current price and availability, see Compare. If you spot an error, please tell us.