Long COVID (PASC)

Long COVID is an infection-associated chronic condition that follows SARS-CoV-2 infection. The most current U.S. definition, published by the National Academies of Sciences, Engineering, and Medicine (NASEM) in 2024, describes it as a condition present for at least 3 months after infection that can affect one or more organ systems and does not require laboratory confirmation of the original infection.^[LC3] Earlier CDC guidance framed post-COVID conditions as problems appearing four or more weeks after infection. It is also called Post-Acute Sequelae of SARS-CoV-2 (PASC) or post-COVID conditions. Critically, Long COVID can follow any severity of acute illness — including infections that were mild or even asymptomatic. Research also shows that each reinfection carries additional Long COVID risk, making prevention of repeat infections an important ongoing health concern.

The WHO estimates that 10–20% of people infected with SARS-CoV-2 experience Long COVID.^[LC5] In the United States, the CDC/NCHS Household Pulse Survey found that 29.8% of adults who had ever had COVID-19 reported ever having had Long COVID, and 5.3% of all adults — about 1 in 20 — reported currently experiencing it. Those are the figures from the survey’s final round, August 20 to September 16, 2024; the series has not been updated since, and the section below sets out what that means.^[LC4] The NIH committed $1.15 billion to the RECOVER Initiative — the largest coordinated study of Long COVID to date — which has enrolled more than 30,000 participants at research sites across the country.^[LC6]

Long COVID encompasses more than 200 reported symptoms. The NIH RECOVER Initiative's 2023 JAMA paper is the most rigorous attempt to narrow that list to something measurable. It enrolled 9,764 participants (89% SARS-CoV-2 infected; median age 47), found adjusted odds ratios of 1.5 or greater for 37 symptoms in infected versus uninfected participants, and then built a weighted "PASC score" from the twelve that contributed most.^[LC1] These are those twelve, exactly as the paper lists them — not a general list of things people with Long COVID report:

How often does it happen? Among the 2,231 RECOVER participants first infected on or after December 1, 2021 — the Omicron era — and enrolled within 30 days of infection, 224 (10%, 95% CI 8.8–11%) were PASC-positive at six months. That is the cleanest prospective figure available, and it is considerably lower than the widely quoted 10–20% WHO range, which derives from earlier, pre-Omicron, largely unvaccinated cohorts.

The NIH RECOVER Initiative and other large studies have identified several factors associated with increased Long COVID risk:^[LC2]

• Female sex — consistently higher rates in most cohort studies
• Older age — though Long COVID affects all ages including children
• Prior health conditions — particularly autoimmune conditions, asthma, and mood disorders
• Severity of acute illness — hospitalization predicts higher risk, though mild cases also lead to Long COVID
• High viral load — higher acute RNA levels associated with persistence
• Reinfection — each infection carries additional Long COVID risk; see our COVID reinfection guide for the cumulative risk evidence
• Unvaccinated status — vaccination reduces Long COVID risk by approximately 50%

Researchers have proposed several (likely overlapping) mechanisms. No single explanation accounts for all Long COVID phenotypes:

• Viral persistence — SARS-CoV-2 RNA and protein detected in gut, lymph nodes, and other tissues months post-infection
• Immune dysregulation — persistent inflammation, activated T-cells, elevated cytokines, and autoantibodies
• Microbiome disruption — gut dysbiosis correlated with Long COVID symptoms
• Reactivation of latent viruses — Epstein-Barr virus reactivation observed in a subset of Long COVID patients
• Endothelial damage and microclots — microthrombi affecting tissue perfusion, particularly relevant for brain fog and fatigue
• Autonomic dysfunction — POTS (Postural Orthostatic Tachycardia Syndrome) seen in a significant Long COVID subset

No FDA-approved treatment exists specifically for Long COVID as of 2026. Management is symptom-based and multidisciplinary. Several clinical trials are underway:

• Paxlovid (nirmatrelvir/ritonavir) — being studied for Long COVID based on the viral persistence hypothesis; RECOVER-VITAL trial ongoing
• Low-dose naltrexone — used off-label for immune modulation and fatigue; early observational evidence
• Antihistamines (famotidine, cetirizine) — some patients report improvement, possibly related to mast cell activation
• Cognitive rehabilitation — for brain fog and cognitive symptoms
• Exercise caution (pacing) — research from the NIH RECOVER Initiative and the ME/CFS literature indicates that vigorous exercise can worsen PEM; pacing, a structured approach to energy management, is the evidence-based standard of care for this symptom
• POTS management — autonomic dysfunction in Long COVID is typically evaluated by a cardiologist or autonomic specialist; management strategies documented in the literature include increased fluid and sodium intake and compression garments, while pharmacological options such as beta-blockers are prescribed and monitored by a clinician

The US national estimate most often quoted comes from the Census Bureau's Household Pulse Survey, with the Long COVID module designed by CDC's National Center for Health Statistics. It is a genuine probability-sample survey of adults, and it produced a consistent series from June 2022. Here is what it found in its final round, covering August 20 to September 16, 2024:

• 61.6% of US adults reported ever having had COVID-19 (95% CI 60.5–62.6).
• 29.8% of those who had ever had COVID-19 reported ever having had Long COVID (28.7–30.8) — that is 17.9% of all adults (17.3–18.6).
• 8.7% of those who had ever had COVID-19 reported currently having Long COVID (8.1–9.3) — 5.3% of all adults (4.9–5.7), the familiar "about 1 in 20" (CDC/NCHS).
• Of adults currently reporting Long COVID, 80.1% reported some activity limitation (77.1–82.9) and 24.3% reported significant limitation (21.7–27.2). That last figure — roughly 1.3% of all US adults significantly limited by Long COVID — is the one that best conveys the population-level burden, and it is rarely quoted.

That was the last reading. The series ran across 30 rounds from June 2022 and ends with the round closing September 16, 2024; the dataset has not been updated since. So the "1 in 20 adults" figure that circulates in 2026 is a two-year-old measurement, not a current one. Over the period it did run, the trend was gently downward: "currently experiencing" fell from 7.5% of all adults (June 2022) to 5.3%, and "ever experienced, among those ever infected" from 35.1% to 29.8%.

This is the same pattern visible everywhere else on this site: CDC and WHO instruments built during the emergency have been retired one by one, and what replaces them is not a better measurement but no measurement. A page that quotes the 5.3% without its date is, by 2026, reporting history as if it were news.

Three estimates that disagree — and why

Long COVID prevalence figures range from about 5% to about 30% depending on the study, and the spread is not sloppiness. It reflects three genuinely different questions:

• RECOVER (10%, 95% CI 8.8–11%) asked: of people enrolled within 30 days of a first Omicron-era infection, how many meet a symptom-score threshold at six months? Prospective, defined threshold, recent infections, largely vaccinated. The most internally rigorous, and the narrowest question.
• Household Pulse, CDC/NCHS (5.3% currently / 17.9% ever, of all adults) asked: do you report symptoms lasting three months or more that you did not have before COVID-19? Self-reported, no clinical adjudication, but a representative national sample covering all infections since 2020.
• WHO's 10–20% is a range drawn from the pre-Omicron literature, much of it in unvaccinated people and some of it in hospitalised cohorts, where the risk is substantially higher.

These are not competing measurements of one quantity. Vaccination status, variant era, severity of the acute illness, and the definition used each move the answer by several-fold. Anyone quoting a single number for "how common Long COVID is" without naming which of these questions they are answering is overstating what is known.

What is not known

• Current US prevalence. No national instrument has measured it since September 2024.
• Whether risk per infection is still falling. It fell sharply between the pre-Omicron and Omicron eras, in step with vaccination and prior infection. Whether it has kept falling cannot be established without a current cohort.
• A diagnostic test. There is none. The NASEM definition and the RECOVER score are clinical and symptom-based; neither is a biomarker.
• Any approved treatment. No treatment is FDA-approved specifically for Long COVID. Everything in the management list above is either symptom management or an off-label or investigational use.
• Which proposed mechanism dominates — viral persistence, immune dysregulation, microclots, autonomic dysfunction, latent-virus reactivation. They are not mutually exclusive, and no single one accounts for all presentations.

For where Long COVID sits alongside the acute risks of COVID-19 and of other infections, see our COVID-19 Risk Perspective and the hub's Virus Risk Perspective. Because each infection carries its own risk, our reinfection page covers what is known about cumulative exposure.