A plain-language reading of 34 recent studies — organised by symptom, graded by how much to trust each one, and honest about where the science disagrees.
The de-identified study now uses one short sequence throughout: observation, hypothesis, result, unknowns, and next check. Imaginary Authors — Falling Into the Sea and Maison Margiela REPLICA — Flower Market are confirmed reported triggers; Bleu de Chanel and Versace Eros are suspected or likely bad but not confirmed. All four panels share eight named ingredients. Exact English tolerated-control panels leave citronellol + alpha-isomethyl ionone as the strongest pair-level label hypothesis—not a proven cause or proof of zero presence in controls. The repeated chicken pattern matters more than the unconfirmed Gal d 7 theory; caffeine timing remains the cleanest mechanism. Read the short study or check the one-glance map.
Long COVID is defined by clusters of symptoms rather than a single lab test, and no cure exists yet. The strongest, most consistent signals are that post-exertional crashes and brain fog are real, common, and persistent, and that pacing and symptom-tailored rehabilitation help more than pushing through. Many hopeful ideas — mast-cell/histamine explanations for food-triggered flares, nerve-directed treatments, blood-filtering, vagus-nerve stimulation — remain early or unproven, and some studies flatly disagree with each other. This report keeps those disagreements visible instead of resolving them, because that is the honest state of the science.
One line per symptom area. Dots show how many papers back it, coloured by strength. Tap any card to jump to the detail.
Post-exertional malaise is a common feature of long COVID (reported somewhere between roughly a quarter and a half of patients, with low-to-moderate certainty and no standard way to measure it), and across every strength of evidence the recurring, non-directive theme is to screen for PEM first and favor pacing over graded exercise when PEM is present.
post-exertional malaise / ME-CFS overlap / pacingThe mast-cell / histamine lens is a plausible, worth-asking-about explanation for food-triggered Long COVID flares, but the evidence is early and split — symptom surveys say Long COVID can look like MCAS, while a small resting-blood study found no marker difference — and none of it can diagnose any one person.
food-triggered flares / histamine / MCASNerve involvement in Long COVID is real and measurable in some people but far from universal, and every treatment idea here (mast-cell approaches, stellate ganglion block) is still hypothesis-level — so the practical next step is asking a clinician about objective small-fiber and standing tests, and about non-antidepressant pain options.
small-fiber neuropathy / nerve painAcross one moderate review, one large-but-indirect cohort, and three small emerging studies, post-COVID standing intolerance is repeatedly linked to measurable autonomic dysfunction and POTS — enough reason to ask a clinician about a simple objective orthostatic test, but not enough to diagnose anyone, and none of these papers explains food-triggered pain, nerve pain, or post-exertional crashes.
autonomic dysfunction / POTS / standing intolerancePersistent brain fog affects roughly a fifth to a third of people after COVID and is a recognized, distinct post-COVID feature (strong meta-analysis ~22%; moderate reviews 25–32%), but every estimate is statistically shaky; non-drug options (CBT, goal-focused cognitive rehab) have only moderate evidence of helping daily function — not measured cognition — while the *BMJ* review found no compelling evidence that the antidepressant vortioxetine helps (which is not the same as proof it fails), and none of this evidence touches nerve pain, food triggers, or orthostatic symptoms.
cognitive dysfunction / brain fogNo cure exists yet: the strongest trials back exercise and rehab for heart-and-lung fitness (though the same review warns exercise can worsen post-exertional crashes) and cognitive rehab for daily functioning around brain fog, while blood-filtering apheresis failed its trial and vagus-nerve and sleep treatments remain unproven, and none of these studies tested food-triggered flares, nerve pain, or standing-collapse symptoms.
treatment / trials / cure statusLong COVID is defined by expert consensus and symptom clusters, not a lab test; the strongest evidence shows brain fog can persist for years on a separate track from mood, while standing near-collapse, nerve pain, and food-triggered flares sit outside these definitions and warrant their own work-up.
clinical guidelines & prognosis/trajectoryWhat PEM is: a delayed crash in energy, function, or symptoms after activity (physical, cognitive, or emotional) that used to be tolerated. It is the core feature long COVID shares with ME/CFS, and its presence changes what kind of activity plan is considered safe.
Two meta-analyses pooled different studies and reached different numbers. Both are shown; neither is settled.
| Source | Strength | Pooled PEM prevalence |
|---|---|---|
| *General Hospital Psychiatry*, 2024 (systematic review + meta-analysis) | strong | 55% (95% CI 38–71%) |
| *Archives of Physical Medicine and Rehabilitation*, 2025 (systematic review + meta-analysis) | moderate | 25% (95% CI 17–36%), GRADE low certainty |
| *BMC Health Services Research*, 2026 (guideline review) | weak | cites ~4–5% meeting ME/CFS criteria or reporting PEM |
Higher estimate (*General Hospital Psychiatry*, 2024, strong):
"The pooled prevalence of PEM among PACS patients at 3 months or more after COVID-19 diagnosis was 0.55 (95 % CI, 0.38, 0.71)."
Lower estimate (*Archives of Physical Medicine and Rehabilitation*, 2025, moderate; graded low certainty):
"Our research confirms that there is a large burden of PEM in adults living with PCC."
The ~4–5% figure comes from the guideline review, which quotes it from *other* studies rather than measuring it (*BMC Health Services Research*, 2026, weak). Its own quotable point is about diagnosis, not prevalence:
"The most fundamental challenge is statistical invisibility: without an activated diagnostic code, services cannot reliably identify or follow people living with Long COVID."
Why the numbers clash: there is no standard PEM definition. The 2024 review found seven different questionnaires used across studies (the DSQ-PEM most common), both meta-analyses had wide confidence intervals, and the studies feeding the 25% estimate were rated "high to very high risk of bias." So how common PEM is remains genuinely unsettled, anywhere from a minority to a majority depending on the tool and the sample.
Sources at very different evidence strengths point the same way: screen for PEM before any exercise program, and if PEM is present, favor pacing (an "energy envelope") over graded exercise with fixed progression.
> "clinical experience has shown that the presence of post-exertional malaise (PEM) is a significant barrier to physical exercise training in people with long COVID."
The reassurance that exercise "did not trigger PEM" is narrower than it sounds.
> "in the REHAB group (n=47) showed a significantly greater improvement in PCS compared with the control group (n=57) (mean difference (Δ) 2.9 points, 95% CI 0.2 to 5.7, p=0.038) at 3 weeks post baseline."
Caveats that limit how far this stretches: the effect was small (2.9 points, with the confidence interval's lower bound almost touching zero), the analysis counted only completers (104 of 132, with more dropout in the rehab arm), it was a single German centre, the cohort was mostly non-hospitalized, none had a formal ME/CFS diagnosis, and people unable to walk were excluded. So it may not represent more severe or bedbound PEM.
These two do not contradict each other. Both point toward caution and individual titration rather than a fixed push-through plan.
None of these papers address food-triggered pain flares (for example chicken or broth), burning or neuropathic pain, the mechanism of brain fog, or the decision about antidepressants. They speak only to how common PEM is and how activity or exercise might be managed. Those other symptoms would need separate, symptom-specific evaluation and their own evidence.
Mast cells are immune cells that release histamine and other chemicals. When they over-react, the result can be flushing, gut upset, pain, and reactions to certain foods, heat, or exertion. Some researchers think this process is part of Long COVID. The evidence is early and it does not all agree. Here is what each study actually shows — read the two sides together, not one alone.
Symptoms can look like MCAS — but this is survey data only. A 2021 survey (*International Journal of Infectious Diseases*, 2021) compared 136 Long COVID patients to controls and to 80 people already diagnosed with MCAS. After COVID, the Long COVID group's mast-cell symptom scores rose to nearly match the MCAS group (21.1 vs 20.1, not significantly different). Quote: *"MCA symptoms were increased in LC and mimicked the symptoms and severity reported by patients who have MCAS."* Strength: emerging. Self-reported symptoms only, no blood tests. Patients were recruited from Facebook support groups (likely sicker) and recalled their pre-COVID health after already being ill (recall bias). It shows overlap, not cause.
A mechanism has been proposed — but it is a hypothesis, not proof. A 2026 review (*Journal of Neuropathology & Experimental Neurology*, 2026) argues that mast cells sit next to nerves and, when triggered by the virus, release chemicals that could drive burning/nerve pain, fatigue, and dizziness. Quote: *"Mast cell activation mirrors patterns seen in small-fiber neuropathy and myalgic encephalomyelitis/chronic fatigue syndrome, suggesting a shared immune-mediated etiology."* Strength: emerging. A narrative review (author-picked studies, no systematic search); much of the biology comes from lab and animal work, and it notes treatment "responses remain variable."
A second review names food as a possible trigger — but no specific food. A 2023 review (*Asia Pacific Allergy*, 2023) repeats the symptom-overlap claim and lists triggers including *"histamine-releasing foods,"* plus exercise, stress, temperature changes, and sleep deprivation. Quote: *"A study by Weinstock et al. indicates that patients with long COVID-19 suffer the same clinical syndrome as patients with mast cell activation syndrome (MCAS)."* Strength: weak. Narrative review, single-source claim, and MCAS symptoms overlap many conditions. It does not name any specific food — no mention of chicken or broth.
Mast cells and IgE were raised in severe COVID — but not in people like an outpatient. A 2025 case series (*Life*, 2025) found high IgE in about half of hospitalized patients and mast cells in autopsy lung tissue. Quote: *"Elevated IgE (>100 IU/mL) occurred in 10/21 patients, with two patients exhibiting levels exceeding 1000 IU/mL."* Strength: weak. These were severely ill or deceased hospital patients, with no healthy comparison group and no histamine or tryptase measured. It suggests the virus *can* involve mast cells in severe disease; it says little about mild, food-triggered flares.
A blood-marker study found no mast-cell difference. A 2024 case-control study (*Scandinavian Journal of Immunology*, 2024) measured two mast-cell blood proteins (tryptase and CPA3) in 24 Long COVID patients vs 24 recovered controls. No difference (p=0.93 and p=0.82). Quote: *"Mast cell activation does not appear to be part of long-term pathogenesis of long-COVID, at least in the majority of patients."* Strength: weak. Small (24 per group), so it can miss a subgroup. Blood was drawn at rest — not during a food-triggered flare — and tryptase is often normal between episodes. It did not measure histamine and did not test food triggers.
The symptom survey and the blood-marker study do not cancel out — they answer different questions. Symptoms can look like MCAS while a single resting blood test looks normal, especially if the blood was not drawn during a flare. So the evidence is genuinely split, and neither side proves nor rules out a mast-cell / histamine process for any one person.
On food: these papers point at *"histamine-releasing foods"* as a trigger category but name no specific foods. The idea that slow-cooked broths and aged or leftover meats are higher in histamine is general nutrition science, not a finding in any of these studies — so treat any chicken/broth link as a question to test, not an established fact.
Small fibers are the tiny nerve endings that carry pain and temperature, and help control blood pressure when you stand. When they misfire, people get burning pain, pins and needles, and sometimes lightheadedness on standing.
Four 2026 papers touch this axis. None is strong evidence — the best-rated are "emerging," one is "weak." Here is what each actually says, and where they pull in different directions.
---
*"Microneurography Reveals Unmyelinated Small Nerve Fiber Dysfunction in Long COVID," Annals of Neurology, 2026. Strength: emerging.*
Researchers used a fine needle to record single nerve fibers in 36 Long COVID patients who already had nerve pain and standing symptoms. Nearly 9 in 10 showed objectively abnormal nerve activity.
"A total of 32 patients (88.9%) had objective electrophysiological abnormalities in peripheral C fibers, including spontaneous nociceptor activity (61.1%), peripheral sensitization (27.8%), and multiple spikes (11.1%)."
What it means: in people like this, the burning maps to measurable nerve dysfunction. The catch: these 36 were hand-picked from a specialist clinic *because* they had these symptoms, so 89% is not the rate in Long COVID generally. No matched control group, no p-values, and this test exists at very few centers. It shows a pattern; it does not prove COVID caused it.
---
*"Neuropathic Symptoms and Persistent Pain After Hospitalization for COVID-19," Healthcare, 2026. Strength: weak.*
A different picture. 80 people were followed for 8 months after being hospitalized for COVID. Screen-positive nerve pain was uncommon (about 11% at the start, dropping to roughly 3%), and most pain improved by month 4.
"Older age, baseline dyspnea, anxiety, and higher baseline LANSS scores were independently associated with lower EQ-VAS scores at 8 months, whereas only higher baseline LANSS scores remained independently associated with persistent pain."
Plain version: the more neuropathic the pain looked early, the more likely it lingered. The catch: this used a questionnaire (LANSS), not nerve testing — so "neuropathic" here is a screen, not a confirmed diagnosis. It was a hospitalized, single-center group, and the key finding rests on very few cases, so it is fragile.
Where papers 1 and 2 disagree — plainly: The nerve-recording study found dysfunction almost everywhere it looked; the hospitalized cohort found neuropathic symptoms uncommon and mostly improving. They aren't truly contradicting each other — they measured *different people* (a hand-picked symptomatic clinic vs. a general hospitalized group) with *different tools* (direct nerve recording vs. a symptom questionnaire). Together they say: nerve involvement is real in some, but how common it looks depends entirely on who you measure and how.
---
*"Long COVID neuropathy: The role of mast cells," Journal of Neuropathology and Experimental Neurology, 2026. Strength: emerging (narrative review).*
This is a review arguing one idea: that immune cells called mast cells — which release histamine — could be a shared driver behind nerve pain, energy crashes, and dysautonomia all at once.
"Mast cell activation mirrors patterns seen in small-fiber neuropathy and myalgic encephalomyelitis/chronic fatigue syndrome, suggesting a shared immune-mediated etiology."
Why it may interest you: it ties several separate-seeming symptoms to one mechanism, and it points toward *non-antidepressant* directions (antihistamines, mast-cell stabilizers). Histamine is also relevant to food-triggered reactions. The catch: this is a hypothesis paper, not a study — no patients, no measurements — and the senior author is a long-time proponent of this model. It states treatment "responses remain variable." It is a reason to ask a question, not an answer.
---
*"Stellate Ganglion Block in the Treatment of Long COVID: A Systematic Review," 2026. Strength: emerging.*
A stellate ganglion block (SGB) is an injection near a nerve cluster in the neck that calms the "fight-or-flight" nervous system. This review pooled 7 small studies.
"Response rates ranged from 55.8% to 100%. The most robust improvements (> 80% patients reporting relief) were seen in cough, dyspnea, headache, joint pain, pain interference/intensity, pins/needles, subjective relief."
The reported numbers look striking (pins/needles relief 9/9; autonomic symptoms improved in about 77%). The big catch: none of the 7 studies had a control group, and outcomes were self-reported — so there is no way to separate the injection from placebo or natural recovery. SGB is itself an invasive injection with risks, and the authors call for controlled trials before the benefit can be believed.
---
None of these four papers can diagnose you, and none directly addresses your food/broth triggers, energy crashes, or brain fog. They raise questions worth asking — not answers.
*What the research says about post-COVID standing problems — and what it cannot say about you. This is about questions to raise with a clinician, not a diagnosis.*
One documented near-collapse on standing is a reasonable reason to ask about a simple objective test. Several Long COVID studies tie standing intolerance to measurable autonomic dysfunction. None studied you, and none can tell you what you have.
A 2024 narrative review in *Nature Reviews Cardiology* (moderate strength — an expert synthesis, not a systematic review or meta-analysis) estimates that cardiovascular autonomic dysfunction:
"CVAD is an important component of post-COVID-19 syndrome, also termed long COVID, and might affect one-third of highly symptomatic patients with COVID-19."
Read that carefully: one-third of the highly symptomatic subgroup — not one-third of everyone with COVID or Long COVID. The word "might," and the absence of any confidence interval, mean this is a rough estimate, not a measured rate.
A 2023 prospective case-control study in the *American Journal of Medicine* (emerging strength — 33 Long COVID vs 33 POTS vs 33 healthy controls, single center) found a much higher figure in its own sample:
"The majority (79%) of those with PASC met the internationally established criteria for POTS."
That 79% is 26 of 33 people, in a young (median age 32), 86%-female group partly recruited from an online Long COVID support group — likely enriched for the most symptomatic. The two numbers (about one-third vs 79%) do not contradict each other; they count different populations. Neither is a general-population rate, and neither applies to any one person.
The objective definition comes from the 2021 *American Autonomic Society statement* (emerging strength — an expert position paper with no patients enrolled), which frames Long-COVID POTS as an open question:
"Some of the symptoms and signs include those of postural tachycardia syndrome (POTS)."
POTS is defined there as a sustained heart-rate rise of >30 bpm within 10 minutes of standing (>40 bpm ages 12–19), with symptoms lasting about 3 months, and no large blood-pressure drop. It is measured with a simple lying-vs-standing test (active stand, 10-minute "NASA lean," or tilt table). This 2021 statement made no prevalence claim at all — it said Long-COVID POTS was largely unstudied and flagged that autonomic-clinic waitlists can run 6–12 months.
The 2023 *American Journal of Medicine* study measured exactly this: heart-rate rise on a 10-minute stand was about 36 bpm in Long COVID vs about 15 bpm in healthy controls (p<0.001).
A 2026 nationwide Japanese claims cohort in *Annals of Medicine* (moderate strength — 3.07 million matched pairs, but based on prescriptions, not diagnoses) found:
"Among 3,074,329 matched pairs, over a median follow-up of 8 months, 13011 composite outcome were observed, and COVID-19 infection was associated with a 36% relative increase in prescriptions (HR 1.36, 95%CI 1.32–1.41)."
Meaning: after COVID, people started autonomic-support medicines more often than matched uninfected people. Every drug in that study (midodrine, fludrocortisone, droxidopa, amezinium) is a non-antidepressant blood-pressure/volume medicine. The authors' own caveat: people diagnosed with COVID see doctors more, which alone can raise prescription rates — so this is an association, not proof of cause, and it cannot diagnose anyone.
A 2026 tilt-table study in *Scientific Reports* (emerging strength — 39 Long COVID vs 22 controls, small, p-values only, several borderline near 0.03) found autonomic problems that a basic reflex test missed:
"SG participants exhibited marked autonomic dysfunction during tilt."
Here the Long COVID group showed a *blunted* sympathetic response on standing even though baroreflex sensitivity looked normal — the authors describe dysautonomia that stays "hidden" until standing is actually tested. Given the small sample, the modest takeaway is that measured autonomic patterns in Long COVID are not uniform, and a normal simple test does not always rule dysautonomia out.
| Paper (journal, year) | Type | Strength | What it shows |
|---|---|---|---|
| *Nature Reviews Cardiology*, 2024 | Narrative review | Moderate | Autonomic dysfunction "might affect one-third of highly symptomatic" Long COVID patients (rough estimate) |
| American Autonomic Society statement, 2021 | Position statement | Emerging | Defines POTS (>30 bpm rise in 10 min); says prevalence unknown, needs research |
| *American Journal of Medicine*, 2023 | Case-control, n=99 | Emerging | 79% of its 33 Long COVID patients met POTS criteria (small, selected sample) |
| *Annals of Medicine*, 2026 | Claims cohort, 3.07M pairs | Moderate | 36% more autonomic-drug starts after COVID (HR 1.36); prescriptions, not diagnoses |
| *Scientific Reports*, 2026 | Tilt-table study, n=61 | Emerging | "Hidden" dysautonomia on tilt; blunted sympathetic response, normal baroreflex |
Brain fog after COVID is real, recognized, and common. What follows is what the studies found, how sure we can be, and what to ask a clinician. None of it is a diagnosis of you or advice.
Three reviews put lasting cognitive trouble at roughly a fifth to a third of people months after infection. The exact figure is unstable because studies measured it very differently.
| Source | Strength | Found (≥12 wk / ≥3 mo after COVID) |
|---|---|---|
| Systematic review + meta-analysis, *Brain, Behavior, and Immunity*, 2022 | Strong | Cognitive impairment ~22% (95% CI 17–28) |
| Systematic review, *Medicine*, 2025 | Moderate | Brain fog ~30%; cognitive impairment ~25% |
| Meta-analysis, *Journal of the Neurological Sciences*, 2022 | Moderate | Brain fog ~32%; memory ~28%; attention ~22% |
The strongest estimate (meta-analysis of 43 studies, n=13,232):
"The proportion of individuals exhibiting cognitive impairment was 0.22 (95% CI, 0.17, 0.28; p < 0.001; n = 13,232; I2 = 98.0)." — *Brain, Behavior, and Immunity*, 2022
Read that as "roughly a fifth," not exactly 22%. The `I2 = 98%` means the studies disagreed enormously, so the precise number is soft.
The 2025 review lands close:
"The pooled prevalence of brain fog was 30% (95% CI: 28–32), while cognitive impairment was 25% (95% CI: 23–27)." — *Medicine*, 2025
But treat its tidy figures cautiously: the paper states it did no meta-analysis, yet still reports pooled percentages with narrow confidence intervals — an internal contradiction its own appraisal flagged. Moderate strength.
The neurology meta-analysis names the cluster plainly:
"Fatigue, cognitive dysfunction (brain fog, memory issues, attention disorder) and sleep disturbances appear to be key features of post-COVID-19 syndrome." — *Journal of the Neurological Sciences*, 2022
Its brain-fog estimate was 32%, but with a very wide band (reported 10–54%) — again reflecting how differently studies measured it. Moderate strength.
Where they agree: cognitive symptoms are a common, recognized post-COVID feature, and the neurology meta-analysis treats them as *distinct* from mood/psychiatric symptoms. Where they're soft: every estimate carries extreme heterogeneity (I2 ~98%), and most relied on self-report rather than objective testing — so no single percentage is precise.
The 2025 review reported brain fog rising over time (~22% at 3–6 months to ~34% at ≥12 months) and being more common in women and in unvaccinated people. Because that paper's synthesis is internally inconsistent (above), read the direction as a signal, not a settled fact — the trend is worth noting, the exact numbers are not reliable. Moderate strength.
Two moderate-strength findings point to non-drug options aimed at daily function:
"an online programme of cognitive behavioural therapy (CBT) probably reduces fatigue (mean difference −8.4 ...) and probably improves concentration (mean difference −5.2, −7.97 to −2.43 ...)." — *BMJ*, 2024
"goal attainment was significantly greater in the CR compared with the TAU group (adjusted mean difference, 2.88 [95% CI, 2.03-3.73]; P < .001; Cohen d = 1.57), with CR providing a large and clinically meaningful treatment effect." — *JAMA Network Open*, 2026
Caveats on both:
On medication for brain fog, the *BMJ* living review found no evidence that it helps — which is different from proof that it fails. Its exact wording (*BMJ*, 2024, moderate certainty overall):
"No compelling evidence was found to support the effectiveness of other interventions, including, among others, vortioxetine, leronlimab, combined probiotics-prebiotics..."
So the honest reading is: there is *no supporting evidence* for the antidepressant vortioxetine in long-COVID cognition — not a demonstrated harm or a proven absence of benefit. Declining an antidepressant *for brain fog* is consistent with that lack of supporting evidence. (The CBT above is offered as a coping tool for concentration and fatigue — not as evidence that the symptoms are psychological.)
_Correction (2026-07-05): an earlier draft called this "high-certainty evidence that vortioxetine does not improve cognition." The source says only that no compelling evidence supports its effectiveness. Corrected against the verbatim abstract — an example of the model overstating a source it half-remembered._
All five papers are about how common cognitive symptoms are, or whether a therapy helps daily function. None address cause, mechanism, or the other symptoms in this picture — none examined nerve/burning pain, food-triggered flares, post-exertional energy crashes, or near-collapse on standing (possible POTS). Their silence on those is *not* reassurance; it means those need their own evidence base and work-up.
There is no cure for Long COVID yet. The studies below test ways to *manage* symptoms. Most tested one symptom at a time, and the largest trials did not screen for post-exertional crashes. So read each as "what helped an average patient," not "what will help you." The last section lists what none of them tested.
---
Evidence: strong — a meta-analysis of 51 randomized trials, 4,026 people. *Effects of therapeutic interventions on long COVID: a meta-analysis of randomized controlled trials* (eClinicalMedicine, 2025).
Exercise training improved walking distance, leg strength, breathing, and mental well-being. It has the biggest evidence base and the only "high-certainty" grade in the paper.
"Exercise training should be prioritized for improving cardiopulmonary function and exercise capacity in Long COVID, supported by high-certainty evidence."
Two honest limits, both from the same paper:
"may pose a risk of symptom exacerbation rather than improvement"
So one paper holds both the headline "prioritize exercise" and the warning that it can backfire in people who crash. That tension is real, not a mistake. (The same review found respiratory-muscle and tele-rehab helped modestly; steroid nasal sprays and olfactory training showed no benefit.)
Ask a clinician: whether your pattern fits PEM or post-exertional symptoms first — because in that case the same evidence points toward pacing rather than graded exercise, and toward autonomic/standing testing before any upright exercise plan.
---
Evidence: moderate — one randomized trial, 78 adults, England. *Cognitive Rehabilitation and Functional Outcomes in Long COVID-Related Cognitive Impairment: A Randomized Clinical Trial* (JAMA Network Open, 2026).
Ten weekly one-hour sessions helped people make progress toward their own chosen daily-life goals — a large effect.
"At 3 months after randomization, goal attainment was significantly greater in the CR compared with the TAU group (adjusted mean difference, 2.88 [95% CI, 2.03-3.73]; P < .001; Cohen d = 1.57), with CR providing a large and clinically meaningful treatment effect."
The catch: that big effect was self-reported. Objective memory and thinking tests improved only a little, and there was no improvement in fatigue, mood, or sleep. Read it as learning to function around brain fog, not restoring it. It is a non-drug option (not an antidepressant). Note that 10 sessions plus homework is itself activity that could provoke a crash.
Ask a clinician: whether formal testing shows cognitive impairment (the trial's entry criterion), and whether goal-focused rehab could be paced to avoid triggering post-exertional crashes.
---
Evidence: moderate — a randomized, sham-controlled crossover trial, 40 adults. *Immunoadsorption Versus Sham Treatment for Post-COVID Syndrome: A Randomised Sham-Controlled Crossover Trial* (Lancet Regional Health – Europe, 2026).
This filters "autoantibodies" out of the blood and is heavily marketed by private clinics. In the trial it *did* remove the target antibodies, but it changed no symptom score versus a fake (sham) procedure, and the serious harms — jugular-vein clots — happened after the real treatment.
"There was no difference in change in symptom severity between immunoadsorption and sham; odds ratio in Post-Covid-19 Functional Scale, OR = 1.17 (95% CI, 0.41–3.36; p = 0.771)"
This is a plain disagreement with the marketing story: even in a group where nearly everyone tested autoantibody-positive, removing the antibodies did not help. "You have autoantibodies" did not mean the treatment worked. (This was a small, single-centre trial with a manufacturer among its funders, so a small missed benefit cannot be fully excluded — but every one of six symptom measures was null.)
Ask a clinician: given the null result and the clot risk, what evidence would justify paying for antibody-removal apheresis.
---
Evidence: emerging — a review of 5 small studies, 154 people total; too varied to pool. *Transcutaneous Auricular Vagus Nerve Stimulation for Post-COVID-19 Condition: A Systematic Review and Critical Appraisal of Clinical Evidence* (Journal of Clinical Medicine, 2026).
An ear-clip nerve stimulator. It was safe (no serious harms across all 154 people), but no controlled trial beat a sham. In the best-powered trial, the sham group's fatigue improved *more* than the real device — even though the device did shift heart-rate-variability numbers.
"Paradoxically, in the best-powered RCT (Percin et al.), sham stimulation produced significantly greater fatigue improvement than active taVNS, despite active taVNS producing significant HRV increases consistent with cardiac autonomic modulation."
So the biology moved but the symptoms did not. Every efficacy outcome was graded very-low certainty. Safe and drug-free, but not established.
Ask a clinician: whether it is worth trying as an experimental, low-risk option — while being clear it has not yet beaten placebo, and that standing symptoms would be worth characterizing (active-stand / tilt) first.
---
Evidence: emerging / insufficient — a review of 14 randomized trials, but only 2 could be combined. *Effect of interventions for the management of sleep disturbances in patients with long COVID: a systematic review and meta-analysis of randomized controlled trials* (Journal of Clinical Sleep Medicine, 2025).
"Out of 3,352 retrieved studies, 14 were included in the systematic review and 2 in the meta-analysis."
The single pooled analysis (two herbal-medicine trials) found no benefit. Some individual trials showed signals, including non-drug relaxation, but the base is thin and low-to-moderate certainty. Two cautions from the paper: some "non-drug" sleep arms were exercise-based (same PEM caution as section 1), and several common sleep and nerve-pain prescriptions (amitriptyline, trazodone, mirtazapine) are antidepressant-class drugs.
Ask a clinician: whether low-exertion sleep approaches (relaxation, sleep hygiene) can be separated from exercise-based ones, and whether any offered sleep/pain drug is an antidepressant.
---
Across all five studies, nothing tested treatments for food-triggered flares, mast-cell / histamine reactions, burning or nerve pain, or orthostatic collapse / POTS. The exercise trials were not screened or stratified for standing intolerance. (One supplement, PEA-LUT, appears in the exercise review but was tested only for smell recovery, not pain.) For those symptoms, this body of evidence is silent — it neither supports nor rules out any treatment, and each would need its own workup.
This section covers two things: how experts *define* Long COVID, and what research says about how it *changes over time*. None of it diagnoses you. It gives you the vocabulary and timelines a clinician uses, plus questions to bring to an appointment.
---
There are two main definitions, built two different ways. They do not fully agree, and neither is a lab test.
The WHO clinical definition *A clinical case definition of post-COVID-19 condition by a Delphi consensus (Lancet Infectious Diseases, 2021).* Strength: strong as a standard — but it is expert agreement, not a measurement.
"post-COVID-19 condition occurs in individuals with a history of probable or confirmed SARS-CoV-2 infection, usually 3 months from the onset, with symptoms that last for at least 2 months and cannot be explained by an alternative diagnosis. Common symptoms include, but are not limited to, fatigue, shortness of breath, and cognitive dysfunction, and generally have an impact on everyday functioning."
Three things to know:
The RECOVER research definition *Development of a Definition of Postacute Sequelae of SARS-CoV-2 Infection (JAMA, 2023).* Strength: strong cohort (~9,700 adults) — but the authors call it preliminary, not yet a diagnostic tool.
This study built a 12-symptom score from data. The 12 selected: post-exertional malaise, fatigue, brain fog, dizziness, gastrointestinal symptoms, palpitations, change in sexual desire/capacity, loss or change of smell/taste, thirst, chronic cough, chest pain, and abnormal movements. How common was a positive score?
"Among 2231 participants first infected on or after December 1, 2021, and enrolled within 30 days of infection, 224 (10% [95% CI, 8.8%-11%]) were PASC positive at 6 months."
Where they differ: WHO is a panel's clinical consensus; RECOVER is a statistical model from one cohort. Different time windows, different symptom sets. Neither is a blood test, and both explicitly expect to change.
---
*Risk Factors Associated With Post-COVID-19 Condition: A Systematic Review and Meta-analysis (JAMA Internal Medicine, 2023).* Strength: strong (860,000+ patients) — but very high study-to-study variation and modest effect sizes.
"The findings of the meta-analysis showed that female sex (OR, 1.56; 95% CI, 1.41-1.73), age (OR, 1.21; 95% CI, 1.11-1.33), high BMI (OR, 1.15; 95% CI, 1.08-1.23), and smoking (OR, 1.10; 95% CI, 1.07-1.13) were associated with an increased risk of developing PCC."
Read carefully:
---
*Neurological and psychiatric risk trajectories after SARS-CoV-2 infection (Lancet Psychiatry, 2022).* Strength: strong (1.28 million matched pairs) — but observational, and risks are measured against other respiratory infections, not healthy people.
This 2-year study found a split: some risks faded within weeks; others stayed.
"By contrast, risks of cognitive deficit (known as brain fog), dementia, psychotic disorders, and epilepsy or seizures were still increased at the end of the 2-year follow-up period."
---
*The impact of post-acute sequelae of COVID-19 on cardiac function and structure (American Journal of Preventive Cardiology, 2026).* Strength: moderate — the authors' own GRADE certainty rating is low.
"This review indicates that PASC is associated with modest, subclinical alterations in cardiac function. These alterations appear more pronounced in older adults and those with cardiometabolic comorbidities, highlighting the potential value of risk-stratified cardiovascular surveillance in individuals with PASC. The long-term clinical relevance of these changes remains unclear and warrants further study."
---
---
All 34 confirmed on Europe PMC by DOI. Sorted strongest evidence first.