Still Showing High COVID Spike Antibodies Years Later? What It Means, and What It Does Not

If you recently saw that your COVID spike antibodies are still strikingly high, maybe two, three, or four years after your last infection or vaccine dose, you may have felt a flicker of worry. Patients bring me this question constantly, and the worry usually sounds like this: “Is something still living in me? Is my immune system stuck fighting it? And if it is distracted, is it still watching for cancer?”

Those are fair questions, and they deserve a careful answer. Let me take them in order, because the order matters. Much of what circulates online about this topic gets the sequence backward, and that backward version causes needless fear.

Why vaccination usually explains it

Start here, because for most people this is the whole answer.

Your body makes two different kinds of antibodies against this virus, and they behave very differently by design. Antibodies against the nucleocapsid (the inner core, abbreviated N) fade fairly quickly. Antibodies against the spike protein (the outer knob, abbreviated S) are built to last. In a study that followed 3679 healthcare workers, researchers tracked 349 who tested positive and collected 1163 monthly blood samples. At 200 days after symptoms, more than 95% still had detectable spike antibodies, while only about 75% still had nucleocapsid antibodies, which fell away with a half-life of roughly 60 days [7].

The reason spike antibodies last is elegant. When researchers looked inside the lymph nodes and bone marrow of vaccinated people, they found that the immune response did not simply switch off. Spike-binding cells in the lymph nodes stayed active for at least 12 weeks after the booster, and bone marrow samples taken 29 weeks later still contained spike-specific antibody-producing cells [21]. These long-lived cells in your bone marrow are a quiet factory that keeps producing spike antibodies for years. That is the design working, not a malfunction.

Now the part that matters most for interpreting your own result. Because the vaccines used in the United States encode spike only, every dose you have received has instructed your body to build and rebuild this factory. Each additional mRNA dose raises the level again and resets it at a new plateau. Follow-up work found that after boosting, the frequency of spike-specific bone marrow antibody-producing cells was several times higher than what was measured seven months after a mild infection or six months after the original two-dose series [20]. A three-year study of vaccinated healthcare workers found anti-spike still high at the three-year mark, drifting down from roughly 3600 to 2040 BAU/mL over the final year in those without recent boosters [19].

Layer an infection on top of vaccination, and you get what scientists call hybrid immunity, which produces the highest and most durable levels. In a head-to-head healthcare-worker study, the vaccinated-only group had higher levels at 3 months but lower levels at 10 months than the infected group, because vaccine-driven levels decline faster from a higher peak. The hybrid group had the highest levels at both time points and the slowest decline [9]. A paired analysis of 197 previously infected men found that adding vaccination raised their spike antibodies 3.3-fold, with more doses corresponding to higher levels [14].

So, if you are still registering a very high spike number years out, the likeliest explanations, in plain order, are hybrid immunity, multiple boosters, or a recent infection. This is the common case. It is not the exception, and it is not evidence that anything is hiding inside you.

Why the antibody number cannot tell you which branch you are on

You may have read that a persistently high spike antibody level points to a long-term “reservoir” of viral antigen hiding in your tissues, rather than simply an active, healthy immune response. I want to address this directly, because it is the claim that drives most of the anxiety, and because the evidence points in a different direction than the popular version suggests.

Here is the problem. Antibodies are the output of your immune system, not the driver of anything. A high number tells you that antibody-producing machinery is running. It does not tell you what is running it. And when researchers have actually looked, the correlation people assume exists has not held up.

Consider three findings that all point the same way. In the group where circulating spike protein genuinely was detected very late, out to roughly 709 days in some patients reporting chronic illness after vaccination, those patients tended to have lower antibody levels, not higher, largely because they had received fewer doses. In a study of 198 long COVID patients, vaccinated patients had considerably higher spike antibodies (averaging 1925 versus 481 BAU/mL), had symptoms in fewer organ systems, and had higher antibody levels that tracked with fewer affected systems [8]. And across the broader literature, it is low antibody levels during acute infection, not high ones, that have been more consistently linked to developing long COVID.

No test in clinical practice can confirm a tissue reservoir in an individual. None. And both the FDA and the CDC have said plainly that spike antibody results should not be used to judge a person’s immunity, protection, or need for vaccination. The FDA’s 2021 safety communication stated that these results should not be used to evaluate a person’s level of immunity or protection at any time, and especially after vaccination. The CDC’s guidance says antibody testing is not recommended to assess immunity following vaccination or to assess the need for vaccination.

So the honest framing is not “high antibodies mean either a reservoir or an active response, and you need to figure out which.” It is this: the antibody number by itself cannot distinguish those two possibilities. What distinguishes them is how you feel and what your inflammation looks like when you actually measure it. The titer is the question.

Your symptoms and your inflammatory panel are the answer. That reframe is not a dodge. It turns an unanswerable worry into an answerable one.

The reservoir branch: what is real, and what is contested

That said, the reservoir concept is not internet folklore. It is a serious research hypothesis with real evidence behind it, and it deserves an honest hearing.

A landmark autopsy study found viral RNA and protein spread widely through the body, including brain tissue, well after the original illness [18]. In people with inflammatory bowel disease undergoing intestinal biopsies, most still had viral genetic material in gut tissue, and about half still had viral nucleocapsid protein roughly seven months after a mild infection; persistent antigen tracked with lingering symptoms [22]. A major immunology review pulled this together and described how such a reservoir could keep releasing viral proteins and keep the immune system engaged [16].

There is also striking imaging evidence. Using a scan that lights up activated T cells, researchers imaged 24 participants and found significantly elevated uptake compared with pre-pandemic controls across the brain stem, spinal cord, bone marrow, gut wall, and lymphoid tissue, in some people as far out as roughly 2.5 years after infection, with viral RNA found in gut tissue in a subset [12]. That is about as close as we have come to seeing lingering immune activation in living tissue.

Now the contested part, and I would rather you hear it from me than discover it later. The most-cited evidence for circulating spike protein comes from ultrasensitive blood assays. One found spike in about 60% of long COVID patients up to 12 months out, with none detected in fully recovered controls [17]. But that study measured no inflammatory markers, so it cannot by itself show that antigen drives inflammation. A 2025 analysis of 188 participants that measured both in the same people found that spike levels at 3 and 6 months correlated modestly with IL-6, IL-2, and IL-1beta, but not reliably with TNF-alpha, CRP, or interferon-gamma, and that persistent spike antigenemia was not associated with having post-COVID condition at all. It also raised a serious technical objection: a chemical step used in the earlier assays reduced detection of genuine spike protein by about 91% while increasing signal in spike-free pre-pandemic plasma, which suggests some of the “persistent spike” signal may be a measurement artifact [5]. That analysis is a preprint and has not completed peer review, so it is provisional too. But it means the foundation here is genuinely under debate.

What is well established is that people with ongoing long COVID show real, persistent inflammation. Patients showed elevated interferon signatures still high at 8 months [13], and a large study of 318 patients with a 333-patient validation cohort found a persistent triad of elevated IL-1beta, IL-6, and TNF traced to overactivated monocytes and macrophages [15]. Much of this improves by the two-year mark [11]. Importantly, these markers separate groups on average but are unreliable for diagnosing any one person.

And chronic exposure to any antigen can exhaust T cells, meaning they gradually lose their ability to multiply and kill effectively [3]. A widely cited study found long COVID patients had altered stress hormones and elevated antibodies against Epstein-Barr virus, a fingerprint of dormant viruses reactivating when immune control slips [10].

So the fair summary of this branch: persistent antigen is a leading hypothesis, the inflammation in symptomatic long COVID is real, and the specific step from circulating spike to measured inflammation in the same person is weakly supported and actively disputed.

If you have active cancer: what this actually means

This is where I want to spend the most care, because your situation differs from the general reader’s in several ways that change the advice:

First, you are more likely to make too few antibodies than too many. This surprises patients, and it is worth knowing. If you are on rituximab or another anti-CD20 therapy, your antibody response is typically blunted hard. In one study, the anti-RBD geometric mean was 5371 U/mL in healthy controls versus 69.3 U/mL in rituximab-treated patients. In lymphoma patients on rituximab, seroconversion was only 8% after the first dose and 21% after the second, compared with 93% and 100% in healthy controls. But here is the reassurance inside that: about 69% of those seronegative patients still had measurable T-cell responses. Your antibody number is not a scorecard of your protection, and a low one on these therapies is expected, not a failure.

Second, the dormancy research is about acute infection, not about your titer. This is the finding that legitimately deserves your attention. In 2025, a study in Nature showed that respiratory viral infections, including influenza and SARS-CoV-2, can awaken dormant breast cancer cells sleeping in the lungs. In mice, infection triggered a greater than hundredfold expansion of these cells into metastatic lesions within about two weeks, and the effect depended on IL-6. Mice lacking IL-6 were protected [4]. The paper included supporting human observational data from the UK Biobank and a large breast cancer database. Read that carefully. The mechanism runs through an acute inflammatory surge from a new infection. It does not run through your antibody level. A 2026 review of this area put the limitation plainly, noting that the human findings in that study are observational and do not establish that respiratory viral infection causes metastatic relapse in people. So this is a serious, plausible concern that justifies protecting yourself from respiratory infections. It is not a reason to fear your lab result.
Third, the immunotherapy question genuinely cuts both ways. If you are on or considering a checkpoint inhibitor, you may have heard that exhausted T cells respond poorly. That is a reasonable hypothesis, but it has not been demonstrated in humans with lingering COVID antigen. Meanwhile, a large 2025 Nature study found that patients with advanced lung cancer or melanoma who received an mRNA COVID vaccine within 100 days of starting checkpoint therapy lived substantially longer, with three-year survival of 55.7% versus 30.8% in lung cancer and 67.6% versus 44.1% in melanoma [6]. Independent researchers have urged caution, since only a randomized trial can establish causation, and one is being planned. I mention this so nothing here pushes you toward avoiding vaccination, which would be the wrong conclusion.

High spike antibodies plus high inflammatory markers: the combination that matters

Now to your specific question, and this is where the article earns its keep.

If your spike antibodies are high and you feel well and your inflammatory markers are normal, that is immune memory. No action needed.

If your spike antibodies are high and your inflammatory markers are elevated, then stop looking at the antibody number entirely. It has done its job by prompting the question. The inflammatory numbers are what carry real information about your cancer, and here is why each one matters:

hs-CRP is the workhorse, with the strongest evidence of any inflammatory marker in cancer. Elevated C-reactive protein predicts worse outcomes across many tumor types. The Glasgow Prognostic Score and its modified version, which pair CRP with albumin, are independently associated with reduced survival in colorectal, esophageal, gastric, and non-small-cell lung cancer. In one colorectal series, the separation was dramatic: median cancer-specific survival of 89.5 months for patients scoring zero versus 25.8 months for those scoring two. No antibody titer offers anything remotely like that prognostic power. If you order one inflammatory test, order this one, and note that it must be the high-sensitivity assay, not the standard CRP used for acute infections.

Albumin is the quiet partner to CRP, and it is why the Glasgow score works as well as it does. Albumin falls as inflammation rises and as nutritional reserve erodes, so the combination captures something a single number misses: not just how much inflammation is present, but what it is costing you. A normal CRP with a falling albumin, or a rising CRP with a dropping albumin, tells a different and more concerning story than either alone. Most of you already have albumin on your routine chemistry panel, so this costs nothing extra to track.

IL-6 is the cytokine upstream of CRP, and cancer research keeps pointing to it. Your liver makes CRP largely in response to IL-6 signaling, so a high CRP is usually telling you that IL-6 is active. Measuring IL-6 directly adds value when CRP is borderline or when you want to watch the response to an intervention. It is also the specific molecule implicated in the dormancy work described above, and the pathway targeted in the CANTOS trial.

Neutrophil-to-lymphocyte ratio is the best value on this list because it requires no additional test. You calculate it from the complete blood count you are almost certainly already getting. A rising ratio reflects the combination of inflammatory drive and lymphocyte depletion, and it has accumulated solid prognostic data across multiple cancers, including as a predictor of response to immunotherapy. If you are on a checkpoint inhibitor, this is a number worth watching over time.

Ferritin and ESR round out the picture with the caveat that both are nonspecific. Ferritin rises with inflammation but also with iron overload and liver issues, and ESR is slow and easily confounded. I use them as supporting context rather than as decision-makers, and I would not act on either in isolation.

A practical note on interpretation. Any of these can be transiently elevated by a recent infection, a recent procedure, a flare of an unrelated condition, or your cancer treatment itself. One elevated reading is a prompt to repeat it, not a conclusion. What matters is a persistent pattern over time, which is exactly why these belong on a panel you track rather than a test you run once.

So, what does the combination mean? If you have active cancer and your hs-CRP and fibrinogen are genuinely elevated, you are carrying a systemic inflammatory burden, and that burden has well-documented cancer relevance regardless of what caused it. The IL-6 and STAT3 pathway is aberrantly activated in many cancers and drives proliferation, survival, invasion, and metastasis while suppressing the anti-tumor immune response [1]. Persistent signaling through STAT3 and NF-kB promotes tumor-supporting inflammation, expands immune-suppressing myeloid cells, and contributes to the T-cell exhaustion that undermines immune surveillance [23]. And the strongest human evidence that this is actionable comes from CANTOS, a randomized trial of more than 10,000 patients in which blocking IL-1beta lowered hs-CRP by 26-41% and IL-6 by 25-43%, and was associated with significantly less incident lung cancer (hazard ratio 0.61 at the 150 mg dose and 0.33 at 300 mg) and lower lung cancer mortality at the highest dose [17a]. The investigators themselves called this hypothesis-generating rather than definitive, and follow-up trials were mixed, but it remains the best randomized signal we have that inflammation is not merely a marker.

Notice what I did not say. I did not say your spike antibodies caused your inflammation. There is no good evidence for that, and in a vaccinated person the antibodies are almost certainly from the vaccine. What I said is that if inflammation is present, it matters on its own terms, and it matters more for your cancer than any antibody number ever will.

What I would actually do

If antibodies are high and you feel well, nothing. Be reassured.

If antibodies are high and you have fatigue, brain fog, or joint pain, get hs-CRP and albumin, a complete blood count so the neutrophil-to-lymphocyte ratio can be calculated, and, where appropriate, IL-6, with ferritin and ESR as supporting context. That panel tells you something. The antibody did not.

If you want to know whether you were ever actually infected rather than only vaccinated, the nucleocapsid antibody test is the tool. Just know its limits: nucleocapsid antibodies wane, and sensitivity drops over time, after mild infection, and in vaccinated or immunosuppressed people. A negative result does not rule out past infection.

If you have active cancer, put your energy where the evidence is strongest. Avoid respiratory infections, because that is the mechanism the dormancy research actually implicates. Discuss vaccination timing with your oncologist rather than avoiding it. Address elevated inflammatory markers directly through the things that reliably move them. Regular exercise measurably lowers IL-6 and TNF-alpha in cancer survivors, with meta-analyses in breast cancer survivors showing significant reductions, though several individual trials found no change, so the benefit is real on average and not guaranteed for everyone. Correct nutrient deficiencies. Treat the metabolic drivers. These are the levers that move.

And please be cautious with anyone selling a “spike detox.” Three randomized trials of antiviral treatment in people with existing long COVID, STOP-PASC [7a], PAX LC [17b], and RECOVER-VITAL, all failed to show meaningful benefit. That does not disprove the reservoir hypothesis, but it does mean nobody currently has a proven way to clear it, and your resources are better spent on measurable targets.

The bigger picture

There is something quietly reassuring in the biology here. Your immune system was designed with a memory that lasts for years, maintained by cells in your bone marrow you will never feel working. When a number comes back high, our job is not to panic at the data point. It is stewardship: find out whether there is actual inflammation, reduce the load your body is carrying, and treat the person, not the printout.

If your results have you worried, bring them in and let us look at the whole picture together.

References

Still Showing High COVID Spike Antibodies Years Later? What It Means, and What It Does Not

Dr. Daniel Thomas, DO, MS