As of March 2023, when the Office for National Statistics stopped collecting data on this condition, 1.879 million individuals had self-assessed as having long covid—symptoms lasting more than 12 weeks following acute covid-19 infection. Of these, the proportion of individuals with symptoms lasting two years or more is around 42%, suggesting a decline in new cases of long covid but a persistence of those with ongoing symptoms.1 Some systematic reviews and meta-analyses have reported that up to a third of such individuals have persistent symptoms of cognitive impairment,23 but estimates vary widely and are complicated by methodological heterogeneity—eg, study size, assessment approach, follow-up duration, and different sampling frames (from self-reported surveys4 to large retrospective matched cohort studies of health records5), as discussed in a recent meta-analysis.6
The pathological underpinnings and potential therapeutic possibilities for cognitive impairment in long covid are also uncertain. The bulk of evidence to date is mechanistic (using basic science, animal models, or human tissue), observational (using longitudinal cohort studies), or hypothetical (reasoning from basic principles); this literature has been well summarised by the RECOVER Consortium.7 Because of the methodological heterogeneity, even when individual studies have been rigorously conducted, it is difficult to know to what extent their findings can be extrapolated and generalised across those with long covid. A few randomised controlled trials of potential treatments (pharmacological and non-pharmacological) …