Researchers in the UK have reported results of what is likely the longest follow-up study to date of cognitive and psychiatric symptoms in adults who were hospitalized with acute COVID-19 infections.

Overall, they found that the burden of such symptoms increased 2 to 3 years following hospitalization, when compared with symptoms 6 and 12 months after. This was due to both worsening of existing symptoms and the emergence of new symptoms. Results appeared in the journal Lancet Psychiatry. They call new attention to the problem of “long COVID.”

The team, which was led by Paul J. Harrison, FRCPsych, of the University of Oxford, a 2004 BBRF Independent Investigator, surveyed a subset of the large Posthospitalization COVID-19 Study (PHOSP-COVID) that involved nearly 8,000 individuals who were released after being treated for COVID at 83 UK National Health Service-participating hospitals between 2020 and 2021.

The new study, called C-Fog, involved 475 PHOSP-COVID participants who consented to be recontacted for follow-up research. This cohort, 40% female, average age 58, completed a multipart computerized cognitive assessment and various questionnaires about their mental health and employment status. These were submitted between 2 and 3 years following participants’ original admission to hospital for COVID.

The many shorter-term studies on COVID-19 that have been conducted since the emergence of the virus have established that infection is associated with increased risks of neuropsychiatric disorders, including depression, anxiety, and cognitive deficits, either in isolation or as part of a “post-COVID syndrome,” often called long COVID. These risks are higher in people who were admitted to the hospital for COVID, studies based on electronic health records have indicated.

Importantly, longer-term data has been lacking, making it impossible for doctors and patients to know what to expect more than a year beyond hospitalization. Other questions that the C-Fog study hoped to answer were whether symptoms experienced early in COVID illness predict later outcomes, as well as whether specific symptoms affect subsequent functioning at work.

Cognitive testing was performed in the C-Fog cohort remotely, via computer-delivered tasks. These made possible assessment of eight cognitive domains: object memory (immediate), reaction speed, 2-dimensional mental manipulation, cognitive control, spatial working memory, spatial planning, verbal analogies, and object memory (delayed). Following cognitive testing, participants were asked to complete several questionnaires online: for depression, anxiety, fatigue, subjective (self-)assessment of cognitive changes, and regarding occupational changes, if any.

Most participants (75%) reported mild or greater depression at the 2- to 3-year follow up, following hospitalization. The comparable figure for anxiety was 53%; for fatigue, 62%; and subjective cognitive decline, 52%. Severe symptoms of depression were reported by 22%, severe fatigue 24%, and severe subjective cognitive decline 25%.

Overall, the researchers said, participants “had worse overall cognitive scores than would be expected for people with the same sociodemographic characteristics” who did not have COVID. “Significant deficits” in cognition “were observed across all cognitive domains."

For those who reported depression at 6 months post-hospitalization, depression scores increased an average of 2 points on a 27-point scale by the 2- to 3-year follow-up. Importantly, “there was evidence of both worsening of persistent depressive symptoms” and the emergence of new symptoms among participants who had been without symptoms at 6 months. The same patterns were observed for anxiety, with the average anxiety score rising from the 6-month assessment to the 2-3-year follow-up by about 1 point on a 21-point scale, as well as the emergence of new symptoms in those who had not reported symptoms at 6 months.

Fatigue showed improvement between the 6- and 12-month post-hospitalization assessments, but then “significantly deteriorated” from 12 months to the 2- to 3-year follow-up.

Broadly speaking, the team said, “individuals admitted to hospital with COVID-19 who were included in the C-Fog cohort continued to experience substantial cognitive and psychiatric burden up to 3 years after hospital admission. Almost 1 in 2 experienced moderate to severe depression, 1 in 4 reported severe cognitive decline, and 1 in 9 had objective signs of severe cognitive deficits that would equate to a difference of 30 points on a typical IQ scale.” Fatigue, where present, added to the burden, they said.

The evidence suggested to the team that much of the symptom burden could be chalked up to the persistence of symptoms already present at 6 and 12 months following hospitalization. But they did not think this persistence of older symptoms could account for the worsening seen in many, or for the appearance of new symptoms. There was “robust evidence for both” in depression, while the evidence in anxiety and fatigue was less compelling due to the size of the cohort. But the team downplayed mere delayed diagnosis of symptoms at the 6- and 12-month assessments as explanation for the worsening that was seen at 2-3 years.

The team said that results of the study suggest that a neuropsychiatric “syndrome” may emerge over time from “a few core symptoms” including those seen at 6 and 12 months post-hospitalization. If further research supports this, “then early interventions targeting the core symptoms might be a viable strategy to limit long-term symptom burden,” they said. For instance, since anxiety at 6 months predicted in this cohort multiple other symptoms at 2-3 years, interventions targeting anxiety in the early period following hospitalization might prevent or reduce other symptoms years later. “Adults with severe ongoing health impairments at 6 months are at particularly high risk of severe symptoms at 2-3 years,” they noted.

More than 1 in 4 participants in the study reported changes in occupation since having COVID-19. In the bulk of cases, this was attributed by participants to cognitive problems. Those who changed occupation “had difficulties executing complex tasks with changing demands,” such as task-switching. This suggests the possible benefit of interventions such as brain training for task switching to reduce the impact of long COVID on cognitively affected patients, the researchers said.

In addition to looking for biological mechanisms underpinning different symptom trajectories, future research might also include individuals affected by the delta and subsequent variants of COVID, which had not arisen when the C-Fog cohort was infected. The impact of vaccination on symptom trajectories could also be explored, the team said.