Long covid

The short version: long COVID is real, it is multisystem, and the evidence shows it can follow even mild infections, but the burden of proof for any intervention that claims to prevent or treat it has not been met.

The evidence you are seeing comes from a wide range of observational studies, not from randomized controlled trials designed to test a specific treatment. That matters because observational studies can show associations but cannot prove cause and effect. The German population-based CoMoLo study, for example, found that people with a prior SARS-CoV-2 infection had about four times the odds of smell and taste disorders and about three and a half times the odds of shortness of breath roughly a year later, compared to people who were never infected [4]. A separate systematic review of 45 studies found that about 1 in 10 people reported persistent gastrointestinal symptoms after infection, compared to about 1 in 20 controls [1]. These numbers give you a sense of the scale, but they come from studies that compare infected to uninfected people, not from trials that test a specific drug or vaccine against a placebo.

The mechanisms being proposed to explain long COVID are multiple and overlapping. One leading hypothesis is that the virus or its components persist in tissues. A 2026 study using colorectal biopsies found evidence of SARS-CoV-2 persistence in about 27% of people with long COVID, compared to about 8% of recovered controls, though the difference was not statistically significant in that small sample [9]. Another thread points to immune dysregulation: a 2025 study found that people with long COVID had persistently higher antibodies targeting the envelope and nucleocapsid proteins of the virus, along with expanded populations of certain T cells and proinflammatory cytokine signatures, suggesting ongoing immune activation [10]. A third mechanism involves microclotting and damage to the lining of blood vessels. The European Academy of Neurology position paper notes that about 55% of COVID-19 patients show MRI abnormalities in the brain and about 20% have cerebral microbleeds, which fits a picture of vascular injury [5]. These mechanisms are not mutually exclusive, and they may operate differently in different people.

What is missing from this evidence is a clear answer to the question that matters most to you: what can you do about it? The studies describe the problem but do not test solutions. The German survey of post-COVID clinics found that care is personalized and symptom-oriented, meaning doctors treat the symptoms as they arise rather than targeting a root cause [8]. A case report of a young woman with severe dysautonomia after mild COVID showed improvement over a year with lifestyle changes and a graduated exercise program, but that is a single case, not a proven protocol [7]. The microbiome review discusses probiotics and fecal transplants as experimental, not established [6]. The evidence for vaccination reducing long COVID risk is mixed: one Portuguese study found that prior vaccination was associated with about a 55% lower risk of long-term cardiovascular events after hospitalization [2], but a small German study of dental workers found that 22% of those who had been infected still reported long COVID symptoms despite high vaccination rates [3]. The Brazilian cross-sectional study found that unvaccinated people had about 3.6 times the prevalence of chest pain compared to vaccinated people, but the finding was inconsistent within the same analysis [11].

The bottom line is that long COVID is a real condition with plausible biological mechanisms and measurable population burden, but the evidence base for any specific prevention or treatment is thin. The studies that exist are mostly observational, mostly funded by public sources rather than manufacturers, and mostly descriptive rather than interventional. No trial has yet proven that any drug, supplement, or vaccine reliably prevents or cures long COVID. The burden of proof remains on the intervention, and that burden has not been met.

My call: long COVID is a documented post-infectious syndrome with multiple proposed mechanisms, but no intervention has been proven to prevent or treat it in a randomized controlled trial. Confidence: moderate for the existence and mechanisms of long COVID; low for any specific intervention.