Long COVID at one year: half improved meaningfully by three years, a quarter recovered fully

Design

Prospective multisite cohort (RECOVER-Adult)

Population

1019 US adults with likely long COVID 12 months after first infection

Primary outcome

Time to two consecutive visits below the long COVID index threshold

Effect

36.2% by 24 months, 50.1% (46.1% to 54.0%) by 36 months; 26.9% symptom-free by 36 months

RECOVER-Adult is a prospective US cohort across 83 sites. This analysis followed 1019 adults who still met a research definition of likely long COVID 12 months after their first infection, and asked how many later improved.

By 24 months after infection, 36.2% no longer met the long COVID threshold on two consecutive visits; by 36 months, 50.1% (95% CI 46.1% to 54.0%). By 36 months, 56.4% had halved their symptom score with good quality of life, and 26.9% had no remaining symptoms on the index. People still affected at 18 or 24 months improved at a similar relative pace afterwards.

This gives clinicians honest numbers for a common consultation. Improvement is the more likely course, but slow, and full recovery is the minority outcome. It is observational and says nothing about which treatments help; the authors' own conclusion is that controlled trials are still needed.

• Tell patients with long COVID at one year that about half improve meaningfully by three years.
• Be honest that complete resolution occurred in about one in four by three years.
• Do not give up on improvement at 18 or 24 months; later improvement still occurred.
• Keep screening for treatable mimics: anaemia, thyroid disease, depression, sleep disorders.

Why it matters

Patients ask whether it will ever go, and until now the answer has been a guess.

Don't overread it

Observational; it describes natural course and cannot show that any treatment helped.

The statistics, in plain English

These are cumulative incidence estimates: the share expected to have reached each milestone by a given time, allowing for people who left the study. The intervals are fairly narrow because the cohort is large. The study describes what happened, not why, so it cannot say whether any treatment changed the course.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.