Blood-cell mutations: when a result matters

DNA sequencing sometimes finds an acquired mutation in blood cells while looking for something else. Such a result can justify specialist assessment or follow-up of blood counts. Its meaning depends on the mutation, its frequency in the sample and the person’s health—not simply on the word “mutation.” Research that uses blood-cell changes to estimate age asks a separate question. [3] [8] [9]

Healthy Longevity Clinic contributes to aging research

Healthy Longevity Clinic is proud to have contributed samples and clinical work to the 2024 Cell Tree Rings study, helping researchers explore age estimation from relationships among blood cells. The study received sole funding from AgeCurve and acknowledged Petr Sramek of LongevityTech.Fund for infrastructure support. AgeCurve’s company page listed Sramek as an investor in September 2026. AgeCurve is also listed in Longevitytech.fund’s official portfolio. We are proud of this investment and of HLC’s direct contribution to the research. [14] [9] [10]

What can a result tell you?

• Clinical value: the mutation pattern, blood counts and medical history can help distinguish risk groups and guide appropriate investigation or monitoring. [3] [4] [11]

• Not established: a positive result alone does not diagnose cancer, measure remaining healthy years or establish which preventive drug someone should take. [3] [8] [9]

• What would strengthen aging-screening claims: evidence that using the exact test in otherwise healthy people improves a meaningful clinical decision and health outcomes, with the burdens of testing and follow-up included.

A somatic mutation is an acquired change

A somatic mutation is a DNA change acquired after conception in a body cell, rather than in an egg or sperm. It can pass to that cell’s descendants as the cell divides. A mutation confined to blood cells is not passed on to children. Some blood-test findings need further assessment before anyone can be sure the change is blood-confined rather than inherited. [1] [2] [3]

Changes can arise through DNA-copying errors and other processes, including exposure to substances that damage DNA. Many have little or no detectable effect on health. “Mutation” describes a sequence change; it does not establish harm. Counting more changes does not necessarily identify which ones matter. [1] [2]

A blood-forming stem cell can acquire a change that gives it an advantage over neighboring cells. Its descendants can then make up a growing share of blood production. This is clonal hematopoiesis, or CH. A clone is a family of cells with a shared ancestor. Expansion of that family is not, by itself, a cancer diagnosis. [3]

CHIP and CCUS mean different things

These definitions require more than a gene name. Low hemoglobin, for example, may have a cause unrelated to the clone. Sequencing cannot by itself distinguish a nutritional problem, another illness, CCUS or an established bone marrow disorder. Assessment may include other blood tests and, when indicated, a bone marrow examination. [3] [4] [11]

A 5% result does not mean a 5% cancer risk

Variant allele fraction, or VAF, is the proportion of DNA sequence readings at a particular position that contain the variant. A hypothetical VAF of 5% is not a 5% chance of cancer. It also does not mean that 5% of the body is damaged, or necessarily that exactly 5% of blood cells carry the change. [3]

The relationship between VAF and the number of cells depends partly on how many copies of that DNA region each cell carries. Detection also depends on the test’s sensitivity and error controls. The conventional 2% CHIP threshold was shaped partly by earlier sequencing technology. It is not a biological boundary between harmless and dangerous. [3]

Before interpreting a percentage, establish what was measured and whether the laboratory considers the finding credible and clinically relevant. If it might be inherited, further assessment can change the implications for relatives and genetic counseling. VAF alone does not settle that question. [3] [11]

Most people in one risk model were in its low-risk group

In 2023, researchers developed a Clonal Hematopoiesis Risk Score using exome data—sequencing of protein-coding DNA—from 438,890 UK Biobank participants. They used separate development and validation groups and also tested the score in clinical cohorts. The risk of a future myeloid malignancy—a cancer affecting blood-forming cells—depended on the mutation pattern, number of mutations, clone size, age, blood counts and red-cell characteristics. [4]

Among the 11,337 people with CHIP or CCUS in the development cohort, about 88% were classified as low risk and about 1% as high risk. These numbers describe how people were distributed among categories. They are not the chance that an individual will develop cancer. A small high-risk group differed substantially from the majority. [4]

That distinction makes the next question more useful than an average relative-risk headline: “What is my estimated absolute risk, over what period, and how uncertain is it?” A score can support that discussion. It cannot diagnose cancer or guarantee an outcome, and estimates need care when transferred to a different clinical setting. [4]

The cardiovascular evidence depends on the population

A 2017 study linked CHIP with coronary heart disease. In analyses from two groups followed over time, carriers had roughly 1.9 times the risk of noncarriers. Separate mouse experiments involving the Tet2 gene supported a mechanism connecting altered blood cells with atherosclerosis, the buildup of plaque in arteries. The human finding was observational; the mouse work did not establish a benefit from treating human carriers. [5]

A 2024 analysis of 63,700 participants from five cardiovascular trials gave a more qualified picture. Over a median 2.5 years, CHIP was not significantly associated with the overall combined cardiovascular outcome after adjustment. It was associated with first heart attacks, but not recurrent heart attacks. Nor did CHIP status identify a significantly greater treatment benefit from the therapies examined. [6]

These results answer different questions in different people. Existing disease, treatment, mutation and the outcome being counted can change the picture. A general-population association cannot be used as a universal prediction for someone already receiving cardiovascular care. A nonsignificant finding also does not prove that risk is absent. [5] [6]

Promising treatment signals need the right comparison

CANTOS enrolled people who had already had a heart attack and had elevated C-reactive protein, a marker of inflammation. An exploratory analysis suggested that participants with TET2 variants might benefit particularly from canakinumab, an anti-inflammatory drug. The TET2 subgroup contained 103 patients. Its favorable result did not translate into a statistically significant difference in treatment effects between the genetic groups: that formal comparison had P = .14. The small subgroup and exploratory design support another study, not a mutation-guided prescribing rule. [7]

A 2025 exploratory LoDoCo2 substudy examined clone growth in 854 participants with chronic coronary disease. The overall difference between colchicine and placebo was not statistically significant (P = .13), although the TET2 subgroup showed slower clone expansion with colchicine (P = .04). This is an interesting biological signal. It does not establish that choosing treatment through CHIP testing prevents cardiovascular events. [12]

A 2026 COLCHIP conference report described a randomized crossover trial in 54 people with coronary disease and CHIP. Each person was assigned to receive four weeks of colchicine and four weeks of placebo in sequence; 48 completed both periods and were analyzed. Colchicine changed inflammatory protein patterns. There was no washout interval—a gap intended to let the effects of the first treatment subside—between treatments. The short study measured inflammatory biology rather than prevention of heart attacks or strokes. These conference findings remain preliminary. [13]

The American Heart Association’s February 2026 scientific statement, summarized by the Association, concluded that no CH-specific therapy had demonstrated efficacy for preventing or treating cardiovascular disease. The inflammation and clone-growth findings do not answer that clinical-outcome question. Established cardiovascular care retains its usual indications; a clone alone does not establish a reason to start a particular drug. [8] [12] [13]

When does testing have a clinical purpose?

Persistent unexplained low blood counts, a mutation found incidentally during other testing, or particular cancer-treatment circumstances can warrant specialist assessment. Published clinical frameworks emphasize interpreting the complete finding and matching follow-up to risk. These are different situations from sequencing a symptom-free person with normal counts simply to measure aging. Neither the clinical frameworks nor the Cell Tree Rings experiment demonstrate that such routine aging screening improves health outcomes. [3] [8] [9] [11]

For someone already found to have CHIP, a useful plan may include review of blood counts and ordinary cardiovascular risk factors. Not everyone needs the same follow-up schedule or an immediate bone marrow biopsy. A hematologist, a specialist in blood disorders, can help decide whether the findings call for reassurance, monitoring, more investigation or a research study. Possible unnecessary testing and anxiety belong in that discussion too. [3] [11]

Before buying a test, ask what each possible result would change. Is it designed to assess blood-cell clones, inherited risk or a research estimate of age? Who would interpret it, and what would justify confirmation or referral? A negative result also needs an explanation of what the test could and could not detect.

AgeCurve’s Cell Tree Rings: a separate research question

AgeCurve describes a platform under development. Its 2024 GeroScience paper, “Cell Tree Rings,” inferred blood-cell family trees from variants detected through single-cell RNA sequencing, which reads RNA from individual cells. Features of those trees were then used to predict chronological age. The study was an age-model experiment, not a validation of a clinical CHIP screening test. [9] [10]

The model was developed using samples from 18 HLC participants and tested with a public dataset from another 18 people. Researchers analyzed five overlapping subsets of 700 cells per person. Those were repeated views of the same people’s samples, not additional independent participants. [9]

The default model’s mean absolute error—the average size of the age-prediction error, ignoring its direction—was 7.6 years internally and 12.8 years in the external dataset. External predictions systematically overestimated age despite strong correlation with chronological age. A model can rank younger and older people well while giving inaccurate individual ages. [9]

The roughly four-year figure highlighted in the study abstract was the internal median error: 4.4 years. A median is the middle error in an ordered set, whereas the mean uses all error sizes. It should not be mistaken for the external mean error or for a guarantee about a particular person’s result. [9]

The study demonstrated an approach for estimating age from cell relationships. It did not establish prediction of future clinical events or selection of an effective treatment. Those uses require separate validation. A research publication and a company’s development description also do not establish a particular patient-use authorization. [9] [10]

How Healthy Longevity Clinic experts evaluate the evidence

For a person holding a report with “5% VAF,” the helpful question is what to do next; the number alone is not a cancer-risk percentage. A useful clinical discussion starts with the exact mutation, whether it is acquired, recent blood counts and the reason for sequencing. A person with a low-risk incidental finding and normal counts faces a different decision from someone with persistently unexplained low counts. The second situation may require further investigation; the first may call for a tailored monitoring discussion. [3] [4] [11]

The next distinction is between predicting risk and selecting treatment. The favorable TET2 subgroup findings justify further research, but their limits prevent using a positive CHIP result as a stand-alone drug-selection rule. Ordinary cardiovascular risk assessment and indicated care still matter. [7] [8] [12]

For an age estimate, the question is different again: how accurate is the exact method in new people, and does acting on it improve a clinical decision? The Cell Tree Rings external error is particularly relevant to someone interpreting a personal age number. Better external accuracy would strengthen measurement confidence; a study following people forward to show improved decisions and health outcomes would be needed to support a clinical aging-screening program. [9]

Three questions for a consultation

Bring the complete laboratory report and recent blood counts so the clinician can interpret the finding in context. [3] [11]

1. What gene and variant were found, what does the VAF mean in this assay, and is the change acquired or potentially inherited?

2. How do my blood counts and history change the interpretation, what is my absolute risk over a stated period, and what follow-up is justified?

3. Is this a clinical clone assessment or an age estimate, and what evidence shows that acting on this exact result would improve my care?

Common questions

Does a blood-cell mutation mean leukemia?

No. An expanded cell family can exist without an established blood malignancy. The mutation, blood counts and clinical assessment determine its significance. An acquired finding should be interpreted, rather than ignored or treated as a diagnosis by itself. [3] [11]

Could my children inherit it?

A mutation confined to blood cells is not passed on to children. But a blood report does not always settle whether a finding is acquired or inherited. Some cases need further assessment and genetic counseling. [1] [3]

Does every CHIP finding require a bone marrow biopsy?

No. The need depends on the blood counts, mutation and wider clinical picture. Published frameworks describe risk-adapted follow-up, including situations where a biopsy is unnecessary. Persistently unexplained low counts deserve their own assessment. [3] [11]

Does a cell-based age estimate tell me which treatment to take?

The Cell Tree Rings study did not test treatment selection. It predicted chronological age in small groups, with larger errors in the external dataset. Choosing a treatment requires evidence beyond that task. [9]