Article Text

Neurology and neuropsychiatry of COVID-19: a systematic review and meta-analysis of the early literature reveals frequent CNS manifestations and key emerging narratives
  1. Jonathan P Rogers1,2,
  2. Cameron J Watson3,
  3. James Badenoch4,
  4. Benjamin Cross5,
  5. Matthew Butler6,
  6. Jia Song7,
  7. Danish Hafeez8,
  8. Hamilton Morrin9,
  9. Emma Rachel Rengasamy10,
  10. Lucretia Thomas11,
  11. Silviya Ralovska12,
  12. Abigail Smakowski2,
  13. Ritika Dilip Sundaram13,
  14. Camille Kaitlyn Hunt14,
  15. Mao Fong Lim15,
  16. Daruj Aniwattanapong6,16,
  17. Vanshika Singh17,
  18. Zain Hussain18,
  19. Stuti Chakraborty19,
  20. Ella Burchill20,
  21. Katrin Jansen21,
  22. Heinz Holling21,
  23. Dean Walton22,
  24. Thomas A Pollak6,
  25. Mark Ellul22,23,24,
  26. Ivan Koychev25,26,
  27. Tom Solomon22,23,
  28. Benedict Daniel Michael22,23,24,
  29. Timothy R Nicholson6,
  30. Alasdair G Rooney27
  1. 1 Division of Psychiatry, University College London, London, UK
  2. 2 South London and Maudsley NHS Foundation Trust, London, UK
  3. 3 Preventive Neurology Unit, Wolfson Institute of Preventive Medicine, Queen Mary University of London, London, UK
  4. 4 Medical School, University of Birmingham, Birmingham, UK
  5. 5 East Lancashire Hospitals NHS Trust, Blackburn, UK
  6. 6 Institute of Psychiatry, Psychology and Neuroscience, King's College London, London, UK
  7. 7 East London NHS Foundation Trust, London, UK
  8. 8 School of Medical Sciences, The University of Manchester, Manchester, UK
  9. 9 Maidstone & Tunbridge Wells NHS Trust, Maidstone, UK
  10. 10 Cwm Taf Morgannwg University Health Board, Abercynon, UK
  11. 11 College of Medical and Dental Sciences, University of Birmingham, Birmingham, UK
  12. 12 Department of Neurology, Psychiatry, Physiotherapy and Rehabilitation, Preventive Medicine, and Public Health, Sofia University St Kliment Ohridski, Sofia, Bulgaria
  13. 13 School of Medicine, University of Glasgow, Glasgow, UK
  14. 14 Division of Neurology, University of British Columbia, Vancouver, B.C, Canada
  15. 15 Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK
  16. 16 Department of Psychiatry, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand
  17. 17 The Wire, New Delhi, India
  18. 18 College of Medicine and Veterinary Medicine, The University of Edinburgh, Edinburgh, UK
  19. 19 Department of Physical Medicine and Rehabilitation, Christian Medical College Vellore, Vellore, Tamil Nadu, India
  20. 20 Faculty of Medicine and Life Sciences, King's College London, London, UK
  21. 21 Department of Psychology, University of Münster, Münster, Germany
  22. 22 Department of Neurology, The Walton Centre NHS Foundation Trust, Liverpool, UK
  23. 23 National Institute for Health Research Health Protection Research Unit in Emerging and Zoonotic Infection, University of Liverpool, Liverpool, UK
  24. 24 Institute of Infection, Veterinary, and Zoological Science, University of Liverpool, Liverpool, UK
  25. 25 Department of Psychiatry, University of Oxford, Oxford, UK
  26. 26 Department of Psychological Medicine, Oxford University Hospitals NHS Foundation Trust, Oxford, UK
  27. 27 Centre for Clinical Brain Sciences, The University of Edinburgh, Edinburgh, UK
  1. Correspondence to Dr Jonathan P Rogers, Division of Psychiatry, University College London, London W1T 7NF, UK;{at}


There is accumulating evidence of the neurological and neuropsychiatric features of infection with SARS-CoV-2. In this systematic review and meta-analysis, we aimed to describe the characteristics of the early literature and estimate point prevalences for neurological and neuropsychiatric manifestations.

We searched MEDLINE, Embase, PsycINFO and CINAHL up to 18 July 2020 for randomised controlled trials, cohort studies, case-control studies, cross-sectional studies and case series. Studies reporting prevalences of neurological or neuropsychiatric symptoms were synthesised into meta-analyses to estimate pooled prevalence.

13 292 records were screened by at least two authors to identify 215 included studies, of which there were 37 cohort studies, 15 case-control studies, 80 cross-sectional studies and 83 case series from 30 countries. 147 studies were included in the meta-analysis. The symptoms with the highest prevalence were anosmia (43.1% (95% CI 35.2% to 51.3%), n=15 975, 63 studies), weakness (40.0% (95% CI 27.9% to 53.5%), n=221, 3 studies), fatigue (37.8% (95% CI 31.6% to 44.4%), n=21 101, 67 studies), dysgeusia (37.2% (95% CI 29.8% to 45.3%), n=13 686, 52 studies), myalgia (25.1% (95% CI 19.8% to 31.3%), n=66 268, 76 studies), depression (23.0% (95% CI 11.8% to 40.2%), n=43 128, 10 studies), headache (20.7% (95% CI 16.1% to 26.1%), n=64 613, 84 studies), anxiety (15.9% (5.6% to 37.7%), n=42 566, 9 studies) and altered mental status (8.2% (95% CI 4.4% to 14.8%), n=49 326, 19 studies). Heterogeneity for most clinical manifestations was high.

Neurological and neuropsychiatric symptoms of COVID-19 in the pandemic’s early phase are varied and common. The neurological and psychiatric academic communities should develop systems to facilitate high-quality methodologies, including more rapid examination of the longitudinal course of neuropsychiatric complications of newly emerging diseases and their relationship to neuroimaging and inflammatory biomarkers.

  • psychiatry
  • clinical neurology
  • systematic reviews
  • COVID-19
  • meta-analysis

This article is made freely available for use in accordance with BMJ’s website terms and conditions for the duration of the covid-19 pandemic or until otherwise determined by BMJ. You may use, download and print the article for any lawful, non-commercial purpose (including text and data mining) provided that all copyright notices and trade marks are retained.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


COVID-19 stimulated a global academic response to examine the clinical sequelae and biology of the SARS-CoV-2 virus, including its neurological and neuropsychiatric impact.1 2 Although the earliest reports naturally highlighted respiratory symptoms,1 it was quickly recognised that SARS-CoV-2, like other coronaviruses,2 can affect the central and peripheral nervous system.3 4

Many of the very earliest studies of the neurological and neuropsychiatric complications of SARS-CoV-2 infection were small retrospective case reports or series.5 6 These initial studies were feasible to deliver quickly in the context of a new and poorly understood disease. Case reports7 8 were superseded by case series,5 6 then case-control9 and cohort studies,10 11 which suggested significant morbidity and mortality from neurological or neuropsychiatric complications.12 Currently, large multicentre prospective studies are underway13 and already reporting.14 We anticipate that the quality of evidence, and our knowledge, will improve considerably as these data continue to emerge rapidly.

In response to these signals, we aimed to develop a novel, sustainable platform to evaluate emerging knowledge of the neurology and neuropsychiatry of COVID-19. This also served to assist colleagues in keeping up to date with the literature relevant to their specialty, given the extraordinary volume and pace with which research is being published. In May 2020, we started logging literature on relevant symptoms, clinical associations and putative underlying mechanisms in our blog, ‘The neurology and neuropsychiatry of COVID-19’, published weekly on the Journal of Neurology, Neurosurgery and Psychiatry website.15 This catalogue of observational studies, reviews, editorials and mechanistic studies has had over 27 000 global views, but it is essentially a library in which studies are narratively summarised and filed. We recognised the potential value of extending this platform to enable analytical summaries by synthesising evidence in the form of a systematic review and meta-analysis, which we termed Systematically Analyse and Review Studies of COVID-19 Neurology and neuropsychiatry.

In the current report, we aimed to answer two questions:

  1. What were the key methodological characteristics of the early evolving literature on the neurological and neuropsychiatric consequences of COVID-19?

  2. What was the prevalence of neurological and neuropsychiatric complications in patients with COVID-19 in observational or interventional studies during this early period of evolving knowledge?

This review is the most comprehensive attempt yet to synthesise the data on the neurological and neuropsychiatric consequences of COVID-19. Other previous works are less up to date, incorporate fewer clinical parameters or have limited scope for meta-analysis.2 16–19


We conducted a systematic review and meta-analysis, based on a registered protocol (PROSPERO ID CRD42020200768) and reported according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines20 (see online supplemental table 1 for completed PRISMA checklist).

Supplemental material

Supplemental material

The overall strategy was to combine synonyms for COVID-19 infection with synonyms for neurological and neuropsychiatric syndromes. We searched Ovid MEDLINE(R) and Epub Ahead of Print, In-Process & Other Non-Indexed Citations and Daily, EMBASE (via Ovid), APA PsycINFO (via OVID) and CINAHL (via EBSCO) from 1 January 2020 to 18 July 2020. Reference lists of other systematic reviews were examined and cross-checked against our database and eligibility criteria. The full search strategy is presented in online supplemental methods 1.

Supplemental material

We included any controlled trials, cross-sectional, case–control, cohort studies or case series reporting neuropsychiatric or neurological manifestations in patients with confirmed or clinically suspected COVID-19. We excluded non-English-language reports. We excluded studies reporting on fewer than 10 infected patients to avoid the reporting biases common in small studies. Meta-analysis was conducted where a clinical manifestation was reported by three or more eligible studies. Studies were included in the meta-analysis only where they provided representative samples of patients with COVID-19 in whom the point prevalence of neurological or neuropsychiatric features could be estimated; studies where patient inclusion was based on neurological or neuropsychiatric complications (e.g. only those referred for clinical neuroimaging) were therefore excluded from the meta-analysis.

Screening of titles, abstracts and full texts for each article was conducted by two of the authors (CW, JS, AR, BC, MB, DH, JB, ERR), each blinded to the others’ ratings. Where there was disagreement about study inclusion, a third author who was a senior member of the team (AR, MB, JS or JPR) arbitrated. Zotero was used for reference management and Rayyan QCRI was used for eligibility screening.

Data extraction was performed on structured forms by two authors: one of the authors (ERR, DH, BC, CW, HM, JB) entered the data, then a second author (AR, MB, CKH, AS, JB, JS, BC, ERR, HM, DA, SR or MFL) ensured the accuracy of each data item by cross-checking against the original source. We recorded the methodological characteristics of studies and the frequency of neurological and neuropsychiatric manifestations reported by each study (see full list of variables extracted in online supplemental table 2). Where data were available for an outcome at follow-up rather than during the acute illness, prevalences at follow-up were presented separately. Where studies reported asthenia as a manifestation, this was coded as fatigue; where a paper reported both asthenia and fatigue, only the figures for fatigue were used.

Supplemental material

Levels of evidence were assessed by use of the Oxford Centre for Evidence-Based Medicine Levels of Evidence.21 Quality of studies and risk of bias were assessed using the Newcastle–Ottawa Scale, including its adaptation for cross-sectional studies.22 23 Quality was assessed by two authors in parallel with arbitration by a third author in cases of disagreement.

For the systematic review, we descriptively reported methodological characteristics of the evolving literature with analytical statistical tests where appropriate. All eligible studies were listed in a table with their study design, demographics and main findings.

For the meta-analysis, the primary outcome was point prevalence of neurological and neuropsychiatric manifestations with 95% CIs. Given the potential for estimation errors with a double arcsine transformation of proportion,24 we used the metafor package in R V.4.0.2 to calculate generalised linear mixed models for each outcome,25 26 before then using the double arcsine transformation as a comparative sensitivity analysis.27 28 Outcome proportions were transformed using a logit transformation. Between-study heterogeneity was calculated using the I 2 statistic. We planned a priori to analyse the following subgroups: retrospective or prospective design, method of SARS-CoV-2 diagnosis, severity of COVID-19 and time point in relation to infection. Ultimately, we only conducted subgroup analysis for retrospective or prospective design and severity of COVID-19 because of lack of consistently presented data for the other subgroups. In addition, due to high heterogeneity, we conducted an additional exploratory subgroup analysis examining country of origin. Subgroup analyses were conducted on the five clinical manifestations most commonly studied: anosmia, dysgeusia, fatigue, myalgia and headache. Significance testing was performed to assess differences in reported frequencies by subgroup.


De-duplicated searches returned a total of 13 292 titles. Abstract and full-text screening generated a final list of 215 eligible studies (figure 1). A complete list of all included studies is presented in online supplemental table 3.

Supplemental material

Figure 1

Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram.

Methodological characteristics of the literature

Methodological characteristics of the studies are summarised in table 1. The most common study type was a case series (83 studies, 38.6%). To explore whether designs evolved in the first half of 2020, we considered studies that started data collection in December 2019–February 2020 to be earlier and those between March and July 2020 to be later. Among the earlier studies, 37 out of 65 (57%) were case series, whereas this proportion fell to 40 out of 115 (34.8%) among the subsequent studies, p=0.004. Change in study design is illustrated in figure 2. Overall, therefore, there was at least a 2-month lag period from the first official report of the outbreak in Wuhan by the Chinese authorities (31 December 2019) to the first group of cohort studies.

Table 1

Methodological characteristics of included studies

Studies were written by a primary author affiliated with an institution from a total of 30 countries globally (figure 3). The most frequent contributors were China (n=50 studies), USA (n=32 studies), Italy (n=28 studies) and France (n=23 studies). All but three studies starting recruitment in January 2020 were located in China. Globally, most studies (138, 64.2%) were single centre without a significant shift towards multicentre studies as the pandemic accelerated: where collection date was clear, 44 of 65 (67.7%) of earlier studies were single centre, compared with 72 of 115 (62.6%) of later studies (p=0.49).

Figure 3

Geographical distribution of studies.

Studies were predominantly in hospitalised patients (118 studies, 54.9%) and during the acute illness (144, 67.0%). There were a total of 105 638 subjects. Number of subjects in each study varied between 10 and 40 469 (median 101, IQR 196). There were 18 studies with 1000 or more subjects.

There was evidence for ethical approval and informed consent in most studies, but this was waived in a minority, frequently because of the particular circumstances of the pandemic.

Quality assessment found only 23 (10.7%) studies to be of high quality, 98 (45.6%) were of moderate quality and 94 (43.7%) were of low quality.

Prevalence of neuropsychiatric and neurological manifestations

Twenty neurological or neuropsychiatric manifestations were estimated by at least three studies, such that we included 147 studies (reporting on 99 905 infected patients) in the meta-analysis. Overall prevalences are shown in table 2 with forest plots available in figure 4 and online supplemental figures 1–20. The most often studied symptoms were headache (examined in 84 studies, n=64 613), myalgia (76 studies, n=66 268), fatigue (67 studies, n=21 101), anosmia (63 studies, 15 975) and dysgeusia (52 studies, n=13 686). The most prevalent symptoms were anosmia (43.1% (35.2% to 51.3%), n=15 975 in 63 studies), weakness (40.0% (27.9% to 53.5%), n=221 in 3 studies), fatigue (37.8% (31.6% to 44.4%), n=21 101 in 67 studies), dysgeusia (37.2% (29.8% to 45.3%), n=13 686 in 52 studies) and myalgia (25.1% (19.8% to 31.3%), n=66 268 in 76 studies). Sleep disorder was a broad term that was used in a number of studies; of the eight studies reporting a sleep problem, two specified insomnia, one sleep impairment and the remainder an unspecified sleep disorder.

Supplemental material

Supplemental material

Supplemental material

Supplemental material

Supplemental material

Supplemental material

Supplemental material

Supplemental material

Supplemental material

Supplemental material

Supplemental material

Supplemental material

Supplemental material

Supplemental material

Supplemental material

Supplemental material

Supplemental material

Supplemental material

Supplemental material

Supplemental material

Figure 4

Forest plots for prevalence of the five most commonly studied symptoms, subgrouped by disease severity: (A) headache; (B) myalgia; (C) fatigue; (D) anosmia; (E) dysgeusia.

Table 2

Overall meta-analytical estimates of point prevalence of neurological or neuropsychiatric symptoms

Between-study heterogeneity was mostly high with I2 ≥90% for 13 manifestations, ≥50% and <90% for 2 manifestations, and <50% for 5 manifestations. Most symptoms were recorded merely as ‘present’ or ‘absent’ by study authors. The robustness of the main analyses was assessed by repeating the analyses on headache, myalgia, anosmia, fatigue and dysgeusia using the standard random-effects model for meta-analysis with the Freeman-Tukey double arcsine transformation. The results were in line with the main analysis (see online supplemental table 4).

Supplemental material

Subgroup analyses

Subgroup analysis was conducted by study design (prospective and retrospective; table 3), case severity (outpatient, mixed non-severe, non-severe inpatients, severe but not admitted to intensive therapy unit (ITU) and admitted to ITU; table 4) and country of origin (online supplemental table 5). For headache, myalgia, anosmia and dysgeusia, there were significantly higher reported rates in prospective studies than in retrospective studies. In the severity subgroup analysis, compared with the ITU group, headache was more common in mixed non-severe and outpatient populations (p<0.001); myalgia was more common in mixed non-severe and outpatient populations (p=0.04 and <0.001, respectively); anosmia was more common in mixed non-severe and outpatient populations (p=0.05 and 0.04, respectively), and dysgeusia was more common in mixed non-severe populations (p=0.02); there were no significant differences between groups for fatigue.

Supplemental material

Table 3

Subgroup analysis by study design for five most commonly studied clinical manifestations

Table 4

Subgroup analysis by case severity for five most commonly studied clinical manifestations


To our knowledge, this is the largest and most comprehensive systematic review of the neurological and neuropsychiatric manifestations of COVID-19. We identified 215 studies, published between January and July 2020, with a total population of 105 638, containing a large variation in the size of studies. We uncovered some general findings about the methodological characteristics of the early evolving literature in response to a novel pathogen. Studies varied substantially in design, geographical location, treatment setting, illness stage, sample size, diagnostic method and clinical manifestations studied. More studies were retrospective than prospective and case series comprised a significant minority of the early literature. In terms of country of origin, after the first few weeks of the pandemic, in which the literature was dominated by studies from China, a wide range of research was produced from 30 countries, among which less economically developed countries were mostly absent. Most studies confirmed formal ethical review and most required informed consent, but these requirements were waived in a subset of cases.

In our review, we summarise point prevalence of 20 neurological and neuropsychiatric complications of COVID-19. The most frequently studied symptoms were heavily weighted towards non-specific features of systemic illness, such as headache, myalgia, fatigue, anosmia and dysgeusia, which are unlikely to be ‘primary’ neurological symptoms. It was predominantly these more non-specific symptoms that were found to have the highest prevalences, ranging from 20.7% (16.1% to 26.1%) to 43.1% (35.2% to 51.3%) (headache and anosmia, respectively). Of note, more specific neurological and neuropsychiatric symptoms such as altered mental status, depression, anxiety, sleep disorder, stroke and seizures were less frequently studied. However, the core psychiatric disorders of depression (23.0% (11.8% to 40.2%)) and anxiety (15.9% (5.6% to 37.7%)) appeared to be highly prevalent. The reported prevalence of major neurological disorders such as ischaemic stroke (1.9% (1.3% to 2.8%)), haemorrhagic stroke (0.4% (0.3% to 0.7%)) and seizure (0.06% (0.06% to 0.07%)) were substantially lower. Subgroup analyses suggested that study design (prospective vs retrospective), severity of illness and country of origin of a study affected the prevalence figures obtained. Importantly, for myalgia, fatigue, anosmia and dysgeusia, prevalences were substantially higher in prospective studies compared with retrospective studies.

There are several limitations to our study, relating both to the quality of the underlying evidence and to the data synthesis. Major limitations in the study design were the frequent absence of comparison groups, limiting conclusions about the specificity of symptoms to COVID-19; retrospective study designs, which meant that only those symptoms that happened to be enquired about were included; and small sample sizes, which risk reporting bias. In terms of populations, the frequent use of hospital inpatients is unrepresentative of the majority of patients with COVID-19, who are not admitted to hospital. Regarding clinical manifestations, the main limitations were reliance on self-report measures, which risks recall biases; lack of baseline assessment, which prevents estimation of incidence; and a focus on non-specific neuropsychiatric symptoms rather than on major neurological and neuropsychiatric disorders. In addition, some of the most commonly studied symptoms (such as weakness and fatigue) have some conceptual overlap,29 so it is possible that the prevalences found in this review may be underestimated. Terminology connoting altered mental status varied, with terms such as delirium and encephalopathy chosen in different studies, despite existing recommendations on standardisation of the nomenclature.30 The finding that only 10.7% of the studies were of high quality limits the strength of any conclusions that can be drawn. In terms of the data synthesis, we were limited by excluding studies not published in English, which may particularly have reduced the number of important studies included from China, and the generalisability of our results may be limited by the geographical scope of the studies. The rapidly evolving literature means that any review on this subject risks becoming out of date. Furthermore, the high heterogeneity between studies, even after subgroup analyses, suggests that variation in populations, outcomes and measurement techniques might account for much of the differences between studies. Finally, the cross-sectional nature and the focus on acute presentations of most studies reported to date limit our ability to draw conclusions about the long-term impact of neuropsychiatric post-COVID-19 symptom burden. Future well-designed prospective cohorts, such as the UK-based Post-hospitalisation COVID-19 Study (, may be able to address this gap in the knowledge.

There are several implications of this review for future research. First, while retrospective studies are important in identifying associations in large patient populations, they are likely to underestimate the prevalence of important symptoms. This may particularly be the case with some neuropsychiatric disorders such as depression and delirium, which are known to be generally under-recognised.31 32 Therefore, even in the context of a pandemic, there is a need to improve the speed with which the academic community can produce prospectively designed studies, which are based on registered protocols and use validated and objective measures. Standardised case definitions and record forms for common neurological manifestations of viral infections were produced by the Brain Infections Global Network from early in the pandemic33 and made freely available. These have been modified by other international groups,34 and are being incorporated into the WHO case report forms.35 More studies are required of those not admitted to hospital and the timing of neurological and neuropsychiatric symptoms relative to diagnosis must be specified. In terms of the clinical manifestations, many of the common and debilitating neurological symptoms (such as headache, myalgia and anosmia) were assessed systematically by a large number of studies, allowing for meaningful prevalence estimates and subgroup analyses. However, some severe neurological and neuropsychiatric disorders, such as depression, stroke and seizures, received comparatively scant attention and would benefit from similar study. Finally, the occasional waivers of ethical review and the more frequent waivers of informed consent in these studies illustrate that some aspects of study review may be overly burdensome—and therefore potentially neglected—during a pandemic. While we acknowledge the need for proper ethical and institutional oversight, COVID-19 may be an opportunity for this process to be streamlined across the field, especially for non-interventional studies, where the risks to participants are minimal, so that studies during a pandemic (and beyond) can start quickly and inform urgent policy needs.

There are several clinical implications of our study. First, practitioners should be aware that neurological and neuropsychiatric symptoms are very common with four (anosmia, weakness, dysgeusia and fatigue) estimated to occur in more than 30% of patients. Second, these non-specific neurological and neuropsychiatric symptoms appear to be the most common. Neuropsychiatric disorders such as anxiety and depression occupy an intermediate space with prevalence of between 15.9% (5.6% to 37.7%) and 23.0% (11.8% to 40.2%), while major neurological disorders such as stroke and seizures are much rarer. However, because of the very high number of individuals infected with SARS-CoV-2 worldwide, even less frequent symptoms may still result in a substantial increase in the burden of disease. This means that services for those with common mental illnesses and neurological rehabilitation should be resourced and equipped for an increase in case numbers. Many of these disorders can become chronic, so the neurological and psychiatric impact of the pandemic may substantially outlast the current phase. Third, given the multitude of symptoms reported, neurological and neuropsychiatric comorbidity is likely to be the norm rather than the exception in COVID-19, so there must be accessible advice and input from these specialties for patients who are acutely unwell. Finally, although there is a relative lack of data on non-hospitalised patients, the data available suggest that several symptoms, such as anosmia, dysgeusia, fatigue, headache and myalgia, are common even among those with milder illness. Although long-term evidence from this earliest literature was sparse, it gives some initial indication that the symptoms described in ‘long COVID’ may be a continuation of some of those experienced in the acute phase of the illness.36 Long COVID is, however, likely to be a heterogeneous entity with a multifactorial aetiology, including viral persistence, inflammatory changes, physical deconditioning and psychological factors. Our finding that the most frequently reported neurological symptoms actually occurred more frequently in those with less severe COVID-19 suggests that neurological symptoms are not necessarily correlated with systemic or respiratory symptoms, implying that different mechanisms or timing of mechanisms may be involved.

In conclusion, COVID-19 is accompanied by a wide range of neurological and neuropsychiatric symptoms from the common, such as fatigue and anosmia, to the more infrequent but severe, such as stroke and seizure. There is substantial psychiatric morbidity, but a lack of control groups limits to what extent causality can be attributed.

Ethics statements

Ethics approval

This study is secondary research that synthesised the results of original papers; as such, it is exempt from ethical approval.


We wish to thank the many healthcare workers who have contributed inestimably to the care of these patients but are seldom recognised in the research literature.


Supplementary materials


  • JPR and CJW are joint first authors.

  • TRN and AGR are joint senior authors.

  • Twitter @drjprogers, @Camer0nWatson, @BadenochJamie, @BenCross23, @mattbutlerpsych, @jiasongpsych, @danish_hafeez1, @Hammy_UK, @emmarachelsamy, @lucretiaathomas, @ralovska, @abismakowski, @maotweets, @EarlNeuroPsy, @vanshika_writes, @ella_burchill, @tompollak, @Melibeus, @IvanKoychev, @RunningMadProf, @BenedictNeuro, @Tim_R_Nicholson, @allyrooney

  • Contributors TRN and AR conceived the study. AR and JPR led and coordinated the study. MB compared the work with other systematic reviews. CW, BC, MB, JS, DH, ERR, LT, AR, MFL and JB screened studies for eligibility. AR, JPR, JS and MB consulted on study inclusion. CW, BC, DH, HM, ERR, LT, SR, RDS and JB extracted the data. JB, MB, JS, DH, HM, ERR, LT, SR, AS, RDS, CKH, MFL, DA, AR and BC checked data extraction. AR conducted OCEBM ratings. JPR calculated descriptive statistics. CW and KJ conducted the meta-analysis, supported by HH. CW and DH created figures. JPR, BC, MB, JS, DH, HM, LT, SR, AS, RDS, CKH, MFL, VS, ZH, SC, EB, DW, TAP, ME, IK, TRN, AR and JB conducted quality assessment. JPR and AS supervised and arbitrated quality assessment. JB made the PRISMA flow chart. JPR, MB and JB sorted references. JB checked adherence to PRISMA guidelines. AR, JPR, JB, ERR, VS and MB drafted the manuscript. JPR, MB, JS and JB checked the completed manuscript. ERR and ZH created tables. MFL sorted funding statements. JPR and EB formatted the manuscript. DW, ME, IK, TS, BDM, TRN, AR and TAP provided senior review of the manuscript. JPR and AR are responsible for the overall content of the study.

  • Funding JPR is supported by a Wellcome Trust Clinical Training Fellowship (102186/B/13/Z). MB is a National Institute of Health Research (NIHR) Academic Clinical Fellow (ACF-2019-17-008). HH is supported by the Deutsche Forschungsgemeinschaft (German Research Foundation) (HO1286/16-1). TAP is supported by an NIHR Clinical Lectureship (no award/grant number). ME is supported by the Association of British Neurologists through a Clinical Research Training Fellowship (no award/grant number). ME and TS are supported by the NIHR Health Protection Research Unit in Emerging and Zoonotic Infections at University of Liverpool in partnership with Public Health England, in collaboration with Liverpool School of Tropical Medicine and the University of Oxford (NIHR200907). IK is supported by the Medical Research Council (Dementias Platform UK Grant MR/L023784/2) and the Oxford Health Biomedical Research Centre (no award/grant number). DA is supported by the Faculty of Medicine, Chulalongkorn University, Thailand (no award/grant number). AR is supported by the Royal College of Physicians of Edinburgh, John, Margaret, Alfred and Stewart Sim Fellowship 2018–2020 (no award/grant number). BDM is supported by the UKRI/MRC COVID-CNS grant (MR/V03605X/1), the MRC-CSF (MR/V007181/1), and the MRC/AMED grant (MR/T028750/1).

  • Disclaimer The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.

  • Map disclaimer The depiction of boundaries on the map(s) in this article does not imply the expression of any opinion whatsoever on the part of BMJ (or any member of its group) concerning the legal status of any country, territory, jurisdiction or area or of its authorities. The map(s) are provided without any warranty of any kind, either express or implied.

  • Competing interests JPR has received payment from the Alberta Psychiatric Association for a lecture and has held one unpaid advisory meeting with representatives from Promentis Pharmaceuticals regarding drug development. LT is in receipt of a bursary as part of the Royal College of Psychiatrists PsychStar scheme. By winning a prize from the Royal College of Psychiatrists, she has received prize money and free attendance at a meeting. She is President of the University of Birmingham Psychiatry Society. IK has been supported by the Medical Research Council through Dementias Platform UK and by the National Institute for Health Research (NIHR) through the Oxford Health Biomedical Research Centre. He has been a medical advisor to Mantrah and Sharp Therapeutics, digital technology start-ups. He holds stock options in Sharp Therapeutics. TS is supported by the NIHR Health Protection Research Unit in Emerging and Zoonotic Infections (grant nos. IS-HPU-1112-10117 and NIHR200907), NIHR Programme Grant for Applied Research (no. RP-PG-0108-10,048), NIHR Global Health Research Group on Brain Infections (no. 17/63/110), and the European Union’s Horizon 2020 research and innovation programme ZikaPLAN (Preparedness Latin America Network), grant agreement no. 734584. He receives royalties from Oxford University Press, Elsevier, Liverpool University Press and Cambridge University Press. He is a consultant for the MHRA Vaccine Benefit Risk Expert Working Group. He filed for a patent on a test for bacterial meningitis based on a blood test (no. GB1606537.7 14 April 2016). He was on the Data Safety Monitoring Committee of the GSK Study to Evaluate the Safety and Immunogenicity of a Candidate Ebola Vaccine in Children GSK3390107A (ChAd3 EBO-Z) vaccine. He chaired the Siemens Healthineers Clinical Advisory Board (1) Data Safety Monitoring Board: Study of Ebola vaccine ChAd3-EBO-Z-Commercial entity. He holds shares in Medefer Solutions. BDM has received payment for a lecture for Valneva.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.