Long Covid is still commonly approached as though it were a single condition. Patients are grouped under one label, trials recruit a single cohort, and treatments are tested as if one underlying biological mechanism were responsible.
That framing was perhaps unavoidable early on. It is no longer adequate.
At Attomarker, our work using the COVID Antibody Spectrum Test has consistently reinforced a central conclusion: Long Covid is not one immune state. It comprises distinct immune endotypes, with different biological drivers and, therefore, different implications for treatment response, safety, and trial design.
Recognising those endotypes is not a semantic exercise. It is foundational to whether the field can move beyond repeated ambiguity and towards meaningful progress.
What an endotype is, and why it changes the problem
An endotype is a biologically defined subtype of a condition, characterised by an underlying mechanism rather than symptoms alone.
Two patients may present with similar symptoms, yet be biologically very different. Conversely, patients with quite different symptom profiles may share a common immune mechanism.
Long Covid exhibits exactly this pattern.
This matters because treating biologically distinct patients as a single cohort does not merely reduce precision. It actively undermines interpretation, safety, and learning.
More about endotypes: https://attomarker.com/disease-areas/long-covid/understanding-endotypes/
What Attomarker’s work shows in practice
Using the COVID Antibody Spectrum Test, Attomarker measures the breadth, strength, and quality of antibody responses across multiple SARS-CoV-2 variants. This allows immune responses to be profiled rather than treated as binary.
Across the Long Covid populations analysed to date, Attomarker’s published material describes three broad immune endotypes:
- Hypo-immune endotype
Characterised by weak or absent antibody responses to one or more variants, consistent with immune compromise or an “immune gap”. This pattern has been observed in approximately 60–65% of Long Covid patients studied. - Hyper-immune endotype
Characterised by strong, potentially over-reactive antibody binding, described by Attomarker as consistent with autoimmune disease patterns. This endotype has been observed in around 10% of patients. - Broad or universal response endotype
Characterised by good antibody coverage across variants, yet with persistent symptoms, suggesting that antibody insufficiency is not the dominant driver in this subgroup. This group accounts for the remaining ~25%.
These proportions are not incidental. They immediately challenge the assumption that Long Covid patients can be treated, trialled, or managed as a homogeneous population.
More about endotypes: https://attomarker.com/disease-areas/long-covid/understanding-endotypes/
Why this matters for trials and outcomes
Most Long Covid trials still recruit patients based on symptoms and duration alone. Immune state is rarely used as a defining inclusion criterion.
If we take Attomarker’s observed endotype distribution seriously, the implications are straightforward.
In an unstratified trial:
- around one in ten participants may sit in a hyper-immune state
- around two thirds may have immune gaps consistent with hypo-immune biology
- the remainder may sit somewhere else entirely
Now consider testing an immune-stimulating intervention across that entire cohort.
It is biologically plausible that such an intervention could benefit a hypo-immune subgroup. It is equally plausible that the same intervention could worsen symptoms or provoke harm in a hyper-immune subgroup.
When those outcomes are averaged together:
- benefit is diluted
- adverse effects become noise
- overall efficacy appears modest or absent
The conclusion becomes “the treatment didn’t work”.
A more accurate conclusion may be that the cohort was wrong.
Safety is not secondary to efficacy
This is not only about trial success. It is also about safety.
A hyper-immune endotype, by definition, reflects immune over-activity. Adding further immune stimulation in that context is not neutral. It increases the plausibility of immune-mediated harm.
The post-COVID literature increasingly documents autoimmune phenomena following SARS-CoV-2 infection. Alopecia areata is one example that illustrates the point cautiously. It is a recognised autoimmune condition, and large population studies have reported an increased incidence following COVID-19.
This does not mean Long Covid causes alopecia. It does mean that immune over-activation and autoimmunity are part of the post-COVID landscape.
If approximately one in ten Long Covid patients sits in a hyper-immune state, then ignoring immune profiling when trialling or deploying immune-modulating therapies creates avoidable risk for that subgroup.
Endotyping is therefore about prudence as much as precision.
We have seen this progression before in cancer
This is not the first time medicine has faced this problem.
For much of the twentieth century, cancers were classified primarily by anatomical location. “Breast cancer” was treated as one disease. Patients often shared symptoms and imaging features, yet outcomes varied widely and unpredictably.
Progress was incremental and slow.
The real shift occurred when cancer began to be classified by underlying biology rather than location alone. Breast cancer became hormone-receptor positive, HER2-positive, triple-negative, and other molecular subtypes.
Patients who looked similar clinically were revealed to be fundamentally different biologically.
Once that shift occurred:
- trials became more targeted
- treatment effects became clearer
- safety profiles improved
- regulatory confidence increased
- companion diagnostics became standard
Crucially, this was not simply about finding better drugs. It was about stopping the practice of treating biologically distinct patients as one cohort.
Long Covid is now at an analogous stage.
Patients often share common symptoms, fatigue, cognitive impairment, autonomic dysfunction. But as with breast cancer, similarity at the surface does not imply similarity underneath.
Progress happens when medicine stops pretending otherwise.
Why immune profiling changes the trajectory
Immune profiling does not claim to solve Long Covid. What it does is change the quality of the questions we can ask.
It allows:
- trials to be designed around plausible mechanisms
- safety risks to be anticipated rather than discovered late
- clinicians to explain variability rather than apologise for it
- patients to be grouped meaningfully rather than blended into averages
This is how ambiguity becomes interpretable.
This is also why continuing to treat Long Covid as a single cohort is no longer a neutral choice. It is a choice that perpetuates confusion and slows learning.
More about Attomarker’s COVID Antibody Spectrum Test: https://attomarker.com/covid-antibody-spectrum-test/
A challenge the field can no longer avoid
The existence of immune endotypes in Long Covid is no longer speculative. The question is whether the field is prepared to act on that reality.
If we continue to recruit, trial, and treat Long Covid as one disease, we should expect continued ambiguity. If we accept heterogeneity and design around it, progress becomes possible.
That transition requires a change in mindset: from convenience to accuracy, from averages to mechanisms, and from hope to design.
Medicine has already made this shift before.
It is time to make it again.
Important Note Regarding Acceleration Round
Attomarker is currently in an Acceleration Round.
Retail investor participation is being conducted through Crowdcube, which is authorised and regulated by the Financial Conduct Authority. All eligibility checks, approvals, and investment activity take place on the Crowdcube platform.
https://www.crowdcube.com/early-access/attomarker
Information for non-retail investors is available via a separate information page.
https://attomarker.com/investors/
This announcement is provided for information purposes only and does not constitute an offer, solicitation, or invitation to invest.