Investigators accessed linked primary care electronic health records through OpenSAFELY-TPP, covering 18.3 million adults registered with English general practices. The study defined three exposure cohorts with up to 5.5 years of follow-up: a pre-vaccination group (COVID-19 diagnosed January 2020–June 2021) and two cohorts from June–December 2021 split by vaccination status (vaccinated and unvaccinated). The authors reported adjusted hazard ratios (aHRs) for incident neurological diagnoses after a COVID-19 diagnosis, comparing incidence to periods before infection or to people without COVID-19. The manuscript is a preprint and has not been peer-reviewed.
Data indicated a pronounced rise in new neurological diagnoses in the first four weeks after COVID-19. For cognitive impairment the aHRs during weeks 0–4 were reported as 4.02 in the pre-vaccination cohort, 2.24 in the vaccinated cohort, and 4.38 in the unvaccinated cohort. For dementia the corresponding aHRs were 5.05, 2.05, and 8.69 respectively. These numbers indicate a substantially higher incidence immediately after infection across all cohorts, with the strongest relative signal seen in the unvaccinated pre-specified subgroup. The study team observed these early post-infection elevations consistently across the cohorts they examined.
The authors reported that incidence of cognitive impairment and dementia remained elevated for an extended period — up to four years after COVID-19 — particularly among people who were hospitalised with COVID-19. Among hospitalised patients the aHRs for cognitive impairment and dementia were reported in the Abstract as approximately 1.27–1.55 and 1.32–1.48 across the vaccinated, unvaccinated and pre-vaccination cohorts. The cohort showed that the relative risk attenuated from the acute window but did not entirely return to baseline within the reported follow-up, indicating a persistent excess incidence in some groups, especially those requiring hospital care.
The study compared pre-vaccination, vaccinated and unvaccinated groups diagnosed with COVID-19. Short-term aHRs for both cognitive impairment and dementia were lower in the vaccinated cohort compared with pre-vaccination and unvaccinated cohorts in the same calendar window. The authors reported larger acute relative risks in the unvaccinated cohort for both outcomes. For longer-term follow-up the Abstract reports persistent elevation across cohorts but highlights that incidence of cognitive impairment remained elevated particularly among unvaccinated individuals and those hospitalised with COVID-19. The report does not provide additional methodological details about vaccination type, number of doses, or timing beyond cohort definitions in the Abstract.
Beyond cognitive outcomes, the study team investigated a range of neurological and neurodegenerative diagnoses. These included Parkinson’s disease, restless leg syndrome, rapid eye movement sleep disorder, motor neuron disease, multiple sclerosis, and migraine. The Abstract states that incidence of most outcomes was elevated in the early post-infection period (weeks 0–4). The authors emphasise in the Abstract that cognitive impairment and dementia showed the clearest sustained increases; the preprint provides a list of other conditions analysed but does not report detailed aHRs for each condition in the Abstract text.
The authors documented ethical approvals from the Health Research Authority (REC reference 22/PR/0095) and the University of Bristol's Faculty of Health Sciences Ethics Committee (reference 117269). They also declared competing interests and funders. The paper is a preprint and, as the authors acknowledge, has not been certified by peer review. The Abstract notes that evidence on the long-term impact of COVID-19 on neurodegenerative and neurological disorders remains limited; the authors report follow-up of up to 5.5 years but do not claim definitive causation. Access to the underlying pseudonymised patient data is tightly governed; the authors state that all code for the OpenSAFELY pipeline and the data management pipeline is available on GitHub. The manuscript suggests a clinical implication: patients who were hospitalised with COVID-19 or remain unvaccinated may warrant additional routine assessment to evaluate persistent or new cognitive concerns.
Limitations documented in the preprint include its current unreviewed status and the broader caveat in the Abstract about limited long-term evidence. The preprint reports adjusted hazard ratios and cohort definitions in the Abstract but does not include further methodological detail, such as specific covariates adjusted for, vaccine dosing schedules, or granular subgroup counts in the Abstract itself. Those details may be available in the full manuscript and supplementary material referenced by the authors.
Ethical approvals, data governance statements and funding sources are provided in the preprint. Next steps explicitly referenced by the authors include continued follow-up and detailed scrutiny of persistent cognitive outcomes in routine clinical care for affected patients, particularly those hospitalised or unvaccinated.