RESEARCH Association between covid-19 vaccination, SARS-CoV-2 infection, and risk of immune mediated neurological events: population based cohort and self-controlled case series analysis Xintong Li,1 Berta Raventós,2,3 Elena Roel,2,3 Andrea Pistillo,2 Eugenia Martinez-Hernandez,4 Antonella Delmestri,1 Carlen Reyes,2 Victoria Strauss,1 Daniel Prieto-Alhambra,1,5 Edward Burn,1,2 Talita Duarte-Salles2 1 Centre for Statistics in Abstract Results Medicine, Nuffield Department Objective The study included 4 376 535 people who received of Orthopaedics, Rheumatology, and Musculoskeletal Sciences To study the association between covid-19 vaccines, ChAdOx1 nCoV-19, 3 588 318 who received (NDORMS), University of Oxford, SARS-CoV-2 infection, and risk of immune mediated BNT162b2, 244 913 who received mRNA-1273, and Oxford OX3 7LD, UK neurological events. 120 731 who received Ad26.CoV.2; 735 870 people 2 Fundació Institut Universitari Design with SARS-CoV-2 infection; and 14 330 080 people per a la recerca a l’Atenció from the general population. Overall, post-vaccine Primària de Salut Jordi Gol i Population based historical rate comparison study Gurina (IDIAPJGol), Barcelona, and self-controlled case series analysis. rates were consistent with expected (background) Spain rates for Bell’s palsy, encephalomyelitis, and Setting 3 Universitat Autònoma Guillain-Barré syndrome. Self-controlled case series de Barcelona, Bellaterra Primary care records from the United Kingdom, and was conducted only for Bell’s palsy, given limited (Cerdanyola del Vallès), primary care records from Spain linked to hospital Barcelona, Spain statistical power, but with no safety signal seen for data. 4 Neurology Department, those vaccinated. Rates were, however, higher than Hospital Clinic de Barcelona Participants expected after SARS-CoV-2 infection. For example, and University of Barcelona, 8 330 497 people who received at least one dose of in the data from the UK, the standardised incidence Barcelona, Spain covid-19 vaccines ChAdOx1 nCoV-19, BNT162b2, 5 ratio for Bell’s palsy was 1.33 (1.02 to 1.74), for Department of Medical mRNA-1273, or Ad.26.COV2.S between the rollout encephalomyelitis was 6.89 (3.82 to 12.44), and Informatics, Erasmus University Medical Center, Rotterdam, of the vaccination campaigns and end of data for Guillain-Barré syndrome was 3.53 (1.83 to Netherlands availability (UK: 9 May 2021; Spain: 30 June 2021). 6.77). Transverse myelitis was rare (<5 events in all Correspondence to: The study sample also comprised a cohort of 735 870 vaccinated cohorts) and could not be analysed. D Prieto-Alhambra unvaccinated individuals with a first positive reverse daniel.prietoalhambra@ Conclusions transcription polymerase chain reaction test result for ndorms.ox.ac.uk No safety signal was observed between covid-19 (or @prieto_alhambra on Twitter; SARS-CoV-2 from 1 September 2020, and 14 330 080 vaccines and the immune mediated neurological ORCID 0000-0002-3950-6346) participants from the general population. events of Bell’s palsy, encephalomyelitis, Guillain- Additional material is published Main outcome measures online only. To view please visit Barré syndrome, and transverse myelitis. An increased the journal online. Outcomes were incidence of Bell’s palsy, risk of Bell’s palsy, encephalomyelitis, and Guillain- Cite this as: BMJ 2022;376:e068373 encephalomyelitis, Guillain-Barré syndrome, and Barré syndrome was, however, observed for people http://dx.doi.org/10.1136/ transverse myelitis. Incidence rates were estimated in with SARS-CoV-2 infection. bmj‑2021‑068373 the 21 days after the first vaccine dose, 90 days after Accepted: 27 January 2022 a positive test result for SARS-CoV-2, and between 2017 and 2019 for background rates in the general Introduction population cohort. Indirectly standardised incidence As of 13 January 2022, the covid-19 pandemic has ratios were estimated. Adjusted incidence rate ratios resulted in more than 5.5 million deaths worldwide. were estimated from the self-controlled case series. After the rapid development of anti-SARS-CoV-2 vaccines, 9.2 billion doses have been administered through national vaccination programmes.1 To date, What is already known on this topic five vaccines against SARS-CoV-2 have received a Spontaneous reporting indicated a potential association between immune conditional marketing authorisation by the European mediated neurological disorders after covid-19 vaccination Medicines Agency. These include two mRNA vaccines: The European Medicines Agency has listed Guillain-Barré syndrome as a rare side BNT162b2 (Pfizer-BioNTech) and mRNA-1273 effect of exposure to viral vector vaccines (Moderna); two viral vector vaccines: ChAdOx1 nCoV- 19 (Oxford-AstraZeneca) and Ad.26.COV2.S (Janssen/ To date, however, the results from research into the risk of immune mediated Johnson & Johnson); and one adjuvanted, recombinant neurological disorders after covid-19 vaccination have been mixed spike protein nanoparticle vaccine: NVX-CoV2373 What this study adds (Novavax).2 All of these vaccines have shown high Using routinely collected data from the UK and Spain, this study found no safety efficacy in preventing severe covid-19 and acceptable signal after covid-19 vaccination for the immune mediated neurological events of safety profiles in clinical trials.3-7 However, potential Bell’s palsy, encephalomyelitis, Guillain-Barré syndrome, and transverse myelitis adverse events related to these new vaccines have been reported, and continuous vaccine safety surveillance An increased risk of Bell’s palsy, encephalomyelitis, and Guillain-Barré syndrome is needed as mass immunisation against covid-19 was, however, seen after SARS-CoV-2 infection continues. the bmj | BMJ 2022;376:e068373 | doi: 10.1136/bmj-2021-068373 1 RESEARCH More recently, concerns have been raised about COV2.S and mRNA-1273 cohorts were only available immune mediated neurological disorders post- in SIDIAP. Vaccinated individuals were required to covid-19 vaccination. Owing to reports of people have received their first dose between the start of the developing Guillain-Barré syndrome after vaccination vaccination campaign in each country (8 December 2020 with Ad.26.COV2.S (108 of 21 million people in the UK, 27 December 2020 in Spain) and one week vaccinated as of late June 2021),8 and ChAdOx1 nCoV- before the end of data availability of each database (9 May 19 (833 people of 592 million doses administered as 2021 for CPRD AURUM, 30 June 2021 for SIDIAP), with of late July 2021),9 10 the EMA listed Guillain-Barré the vaccination date used as index date. We excluded syndrome as a rare side effect related to these vaccines. those who received more than one brand of a covid-19 Guillain-Barré syndrome has also been associated with vaccine. For the second dose cohorts, participants were mRNA vaccines in a few people.11 12 In addition, Bell’s required to have received their second dose in prespecified palsy, encephalomyelitis, and transverse myelitis intervals after the first dose. For both databases, the events have been described in case series studies interval allowed between doses of two dose regimens after covid-19 vaccination with both viral vector and (except Ad26.COV2.S) was 14 to 180 days. The SARS- mRNA vaccines.13-17 Although these events are not CoV-2 cohort included people with a first positive reverse necessarily due to covid-19 vaccines, the temporal transcription polymerase chain reaction (RT-PCR) test association between the events and vaccination result or antigen test result between 1 September 2020 warrants robust post-vaccination surveillance. Large and one week before the end of data availability of each scale epidemiological studies are required to determine database, with the test date used as index date. Data from whether covid-19 vaccination increases the risks of both RT-PCR and antigen tests were available in SIDIAP, these events above background rates in the general whereas only RT-PCR test results were available in CPRD population. AURUM. From the SARS-CoV-2 cohort we excluded We leveraged large routinely collected datasets individuals vaccinated against covid-19 before infection. including millions of vaccinated people in the UK We also identified a background population cohort that and Spain to study the potential association between included all individuals registered in CPRD AURUM and covid-19 vaccines and the short term risk of developing SIDIAP as of 1 January 2017 (index date). Bell’s palsy, encephalomyelitis, Guillain-Barré All study participants were required to be 18 years or syndrome, and transverse myelitis. To place these risks older and to have at least 365 days of data availability in context, we also studied the association between before their index date. For each specific outcome, we SARS-CoV-2 infection and risk of these immune excluded those who had experienced the outcome in mediated neurological events. the year before the index date. For each cohort, we followed participants from the index date until the Methods earliest of end of follow-up (21 days for vaccinated Data sources people, 90 days for those with a diagnosis of covid-19, For this study we used data from primary care records in and 31 December 2019 for the general population both the UK and Spain. The Clinical Practice Research cohort), first occurrence of the adverse event, end Datalink (CPRD) AURUM contains routinely collected of data availability, or until transference out of the data from primary care practices in the UK,18 19 database or death. For cohorts that had received a first representing 20% of the current UK population.20 vaccine dose, we also censored follow-up if a second Data from Spain came from the Information System dose was observed before 21 days. for Research in Primary Care (SIDIAP; www.sidiap. org), a primary care database that covers 80% of the Events of interest population in Catalonia, Spain, and is linked at an The events of interest were four immune mediated individual level to hospital data. These hospital data neurological disorders prespecified as potential included information from all public and private adverse events of special interest for covid-19 vaccine hospitals in Catalonia (Conjunt Mínim Bàsic de Dades safety: Bell’s palsy, encephalomyelitis, Guillain-Barré d’Alta Hospitalària, CMBD-AH).21 Both databases have syndrome, and transverse myelitis.2 We identified been mapped to the Observational Medical Outcomes these events using previously published clinical codes Partnership (OMOP) common data model,22 which from electronic health records.23 Supplementary allowed the same analytical code to be run against both appendix table 1 provides details of the Systematized datasets without the need to share patient level data. Nomenclature of Medicine (SNOMED) codes used to define the outcomes. Study participants The populations of interest were individuals who had Study design received at least one dose of a covid-19 vaccine and Firstly, we used the historical rate comparison people with SARS-CoV-2 infection. Vaccination cohorts method (see fig 1). Incidence rates of each outcome were constructed of people vaccinated according to the in the vaccinated and SARS-CoV-2 cohorts were used product (ChAdOx1 nCoV-19, BNT162b2, mRNA-1273, as observed rates and compared with the expected or Ad.26.COV2.S) and dose administered (first or second background incidence rates estimated from the dose, with only a single dose cohort for Ad.26.COV2.S general population cohort. For the vaccinated cohorts, as this vaccine is comprises a one dose regimen). Ad26. we estimated the rates during 1-21 days after a first 2 doi: 10.1136/bmj-2021-068373 | BMJ 2022;376:e068373 | the bmj RESEARCH vaccine dose (day 0). For the SARS-CoV-2 cohort we used for definitions of comorbidities and drug use. used a post-test period of 90 days. We estimated the observed rates during the 21 days Secondly, we used a self-controlled case series after immunisation for the vaccinated cohorts and 90 method. In this approach, only individuals who days after testing for the SARS-CoV-2 cohort. Similarly, experience the outcome are included, with participants background rates were estimated for the general acting as their own controls and thereby eliminating population from 1 January 2017 to 31 December time fixed confounding.2 Within person comparisons 2019. We calculated crude incidence rates as the total of event rates were made between the baseline number of events divided by the person time at risk per period before vaccination and the period at risk of an 100 000 person years. Indirect standardisation is used outcome. We defined the at risk period as the 0-21 days to account for differences between the age structure of after a first vaccine dose or after a SARS-CoV-2 positive the vaccinated or SARS-CoV-2 cohorts and the general test result. The 0-21 day period was subdivided into population.26 Observed and expected rates were several prespecified time periods: 0, 1-7, 8-14, 15-21 compared using standardised incidence ratios with days, as well as for the 1-21 day period overall. We corresponding 95% confidence intervals. considered events at day 0 separately, as these events For the self-controlled case series analysis, only the might precipitate hospital admissions and subsequent first event was considered for each participant. Because covid-19 testing, which could result in a positive events could potentially decrease the probability of association between SARS-CoV-2 infection and the being vaccinated, we removed a 21 day pre-risk period studied events.25 The risk period for the SARS-CoV-2 from the baseline period and reported separately. cohort was also extended to 90 days. The study period To assess potential for event dependent observation of the self-controlled case series analysis was defined periods, we plotted a histogram of the time between from 1 January 2017 to 21 days after the first vaccine the occurrence of the event and the end of observation dose or 90 days after a positive test result. for individuals censored and uncensored, and we calculated the number of deaths that occurred after Statistical analysis each event (day 0 to day 7). Conditional Poisson We characterised the participants in each cohort regression models were fitted to estimate incidence rate by personal characteristics, such as age and sex; ratios and 95% confidence intervals for each outcome,27 comorbidities any time before vaccination; and recent comparing the risk period with the baseline period. drug use during the six months before the index date. These models were estimated separately for each cohort Supplementary appendix table 2 shows the codes of interest and were adjusted for age (in five year bands) Background rates General population cohort 1 January 2017 to 31 December 2019 General population risk period Infection (day 0) SARS-CoV-2 cohort UK: 1 September 2020 to 8 May 2021 Post-infection risk period Spain: 1 September 2020 to 23 June 2021 First dose vaccination (day 0) Vaccination cohorts UK: 1 September 2020 to 8 May 2021 Post-vaccination risk period Spain: 1 September 2020 to 23 June 2021 Self-controlled case series Infection or first dose vaccination (day 0) 1 January 2017 to end of risk period Pre-infection or Post-infection or pre-vaccination post-vaccination risk period risk period Fig 1 | Study design. Potential risk period (dark blue) for vaccination cohorts was defined as time between the start of the vaccination campaign and one week before the end of data availability for each database (CPRD AURUM: 8 December 2020 to 2 May 2021; SIDIAP: 27 December 2020 to 23 June 2021). For the SARS-CoV-2 infected cohort, the potential risk period started on 1 September 2020. The baseline period for the self-controlled case series analysis (light blue) was defined from 1 January 2017 to 21 days before the day of vaccination or SARS-CoV-2 positive test result. The pre-risk period (pink) was defined as −21 to −1 days before vaccination or SARS-CoV-2 positive test result, and the risk period (orange) was defined as 1 to 21 days after vaccination and 1-90 days after a SARS-CoV-2 positive test result the bmj | BMJ 2022;376:e068373 | doi: 10.1136/bmj-2021-068373 3 RESEARCH and seasonality (four seasons). Self-controlled case 1 644 365), including 1 324 666 with ChAdOx1 nCoV- series analyses were only conducted for comparisons 19, 2 420 351 with BNT162b2, and 160 228 with with a ≤2 minimum detectable relative risk.28 mRNA-1273. Among those vaccinated, 594 407 had We applied two sensitivity analyses to assess the covid-19 before the first dose. In addition, the study impact of study design choices. Firstly, to exclude sample comprised 735 870 unvaccinated people events potentially related to covid-19, in the vaccinated with SARS-CoV-2 infection (CPRD AURUM: 447 233; cohorts we excluded individuals infected with SARS- SIDIAP: 288 637) and 14 330 080 people from the CoV-2 before the index date. For this, as RT-PCR tests general population (CPRD AURUM: 9 651 568; SIDIAP: were not routinely performed during the first wave of 4 678 512). Participants who received the first dose of the pandemic, the covid-19 definition was broadened any of the covid-19 vaccines were older (median age to include positive SARS-CoV-2 results (RT-PCR or 56-64 years in CPRD AURUM, 51-62 years in SIDIAP) antigen test) or a compatible covid-19 clinical code than the general population (median age 48 years in (see supplementary appendix table 3). Secondly, to CPRD AURUM, 47 years in SIDIAP). In CPRD AURUM, include those with little previous use of healthcare, participants with SARS-CoV-2 infection were younger we also replicated the analyses after removing the one than those vaccinated against covid-19 (median year before observation requirement for participants. age 41 years). Vaccinated individuals also had more Any subgroups with fewer than five people were comorbidities than the general population, including blinded and reported as less than five, following the autoimmune diseases, cancer, diabetes, obesity, heart requirements of information governance. disease, and renal impairment (table 1 and table 2). Patient and public involvement Bell’s palsy No patients or members of the public were directly Within 21 days after a first dose of ChAdOx1 nCoV- involved in the design or analysis of the reported 19, Bell’s palsy was observed in 144 people: 117 data. The independent scientific advisory committee observed in CPRD AURUM compared with 164.5 responsible for the approval of our protocol involved expected given background rates in the general patients in the evaluation of our data access population (standardised incidence ratio 0.71, 95% application. confidence interval 0.59 to 0.85; table 3). In SIDIAP, the corresponding values were 27 observed compared Results with 82.6 expected (0.33, 0.22 to 0.48; table 4). Thirty Figure 1 shows the study design and table 1 and table five patients (25 in CPRD AURUM; 10 in SIDIAP) were 2 show the distribution of personal characteristics, observed after the second dose of ChAdOx1 nCoV-19, comorbidities, and recent drug use for each resulting in a standardised incidence ratio of 0.26 database and cohort. The study sample comprised (0.18 to 0.39) in CPRD AURUM and 0.21 (0.11 to 0.39) 8 330 497 people who received at least one dose of a in SIDIAP (table 3, table 4, and fig 2). Bell’s palsy was covid-19 vaccine (CPRD AURUM: 5 477 859; SIDIAP: observed in 46 and 24 patients after a first and second 2 852 638), including 4 376 535 with ChAdOx1 dose of BNT162b2 in CPRD AURUM, compared with nCoV-19, 3 588 318 with BNT162b2, 244 913 with 116.4 and 99.5 expected (0.40, 0.30 to 0.53, and 0.24, mRNA-1273, and 120 731 with Ad26.COV2.S. Among 0.16 to 0.36, respectively). In SIDIAP, rates of Bell’s palsy those who received two vaccine doses (possible for after BNT162b2 (first and second doses) and mRNA- all except Ad26.COV2.S), 3 745 017 completed the 1273 and Ad26.COV2.S first doses were similar to those vaccination course (CPRD AURUM: 2 260 880; SIDIAP: expected, with equivalent standardised incidence ratios Table 1 | Baseline characteristics of study participants in UK’s Clinical Practice Research Datalink AURUM database General ChAdOx1 nCoV-19 BNT162b2 SARS-CoV-2 ­population First dose Second dose First dose Second dose ­infection Characteristics (n=9 651 568) (n=3 782 401) (n=1 093 812) (n=1 695 458) (n=1 167 068) (n=447 233) Median (IQR) age (years) 48 (33-63) 56 (47-66) 71 (59-76) 64 (49-76) 69 (52-78) 41 (30-54) Men 4 825 624 (50.0) 1 829 719 (48.4) 482 007 (44.1) 711 217 (41.9) 462 061 (39.6) 204 798 (45.8) Median (IQR) past observation time (years) 12.8 (4.9-23.7) 16.1 (6.7-27.6) 20.4 (8.2-32.4) 18.0 (7.1-30.1) 19.1 (7.5-31.3) 11.6 (4.5-22.2) Comorbidities Autoimmune disease 198 170 (2.1) 105 128 (2.8) 47 582 (4.4) 62 384 (3.7) 45 003 (3.9) 8988 (2.0) Malignant neoplastic disease 606 442 (6.3) 319 601 (8.4) 185 702 (17.0) 243 369 (14.4) 191 533 (16.4) 17 519 (3.9) Diabetes mellitus 685 313 (7.1) 383 250 (10.1) 181 246 (16.6) 263 365 (15.5) 167 689 (14.4) 32 705 (7.3) Obesity 349 425 (3.6) 190 790 (5.0) 67 163 (6.1) 98 771 (5.8) 61 431 (5.3) 20 436 (4.6) Heart disease 827 039 (8.6) 413 054 (10.9) 232 267 (21.2) 319 053 (18.8) 236 329 (20.2) 28 603 (6.4) Hypertensive disorder 1 761 421 (18.3) 915 008 (24.2) 449 684 (41.1) 592 676 (35.0) 445 346 (38.2) 59 638 (13.3) Renal impairment 512 050 (5.3) 242 170 (6.4) 157 400 (14.4) 208 112 (12.3) 168 226 (14.4) 15 982 (3.6) Drug use (183 days before to four days before index date) Systemic corticosteroids 545 205 (5.6) 187 454 (5.0) 77 668 (7.1) 103 444 (6.1) 73 838 (6.3) 18 932 (4.2) Antithrombotics and anticoagulants 190 145 (2.0) 70 593 (1.9) 43 189 (3.9) 58 379 (3.4) 44 663 (3.8) 4388 (1.0) Lipid modifying agents 290 710 (3.0) 137 925 (3.6) 71 958 (6.6) 95 123 (5.6) 69 719 (6.0) 9110 (2.0) Antineoplastic and immunomodulating agents 170 815 (1.8) 40 343 (1.1) 14 601 (1.3) 25 199 (1.5) 18 078 (1.5) 9041 (2.0) IQR=interquartile range. 4 doi: 10.1136/bmj-2021-068373 | BMJ 2022;376:e068373 | the bmj RESEARCH of 0.86 (0.70 to 1.04), 0.88 (0.71 to 1.08), 0.92 (0.54 14.9 (14.8-15.1) 135 813 (47.1) 30 542 (10.6) 66 933 (23.2) 37 800 (13.1) 65 030 (22.5) to 1.55) and 1.15 (0.52 to 2.56). Incidence rate ratios (n=288 637) 22 242 (7.7) 16 807 (5.8) 15 654 (5.4) SARS-CoV-2 46 (34-59) 5857 (2.0) 7618 (2.6) 4268 (1.5) 5046 (1.7) could only be stratified by age and sex for Bell’s palsy, ­infection as it was the only event in which some of the strata had more than five occurrences. None of the strata showed significantly higher rates after vaccination in either 15.4 (15.3-15.4) 15.4 (15.3-15.4) 15.4 (15.3-15.5) 64 435 (53.4) CPRD AURUM or SIDIAP (see appendix figure 2). Rates 15 049 (12.5) 30 154 (25.0) 15 827 (13.1) 33 102 (27.4) Ad26.COV2.S (n=120 731) of Bell’s palsy were higher than expected among those 51 (43-67) 2242 (1.9) 9397 (7.8) 4923 (4.1) 7296 (6.0) 3548 (2.9) 2934 (2.4) 1310 (1.1) infected with SARS-CoV-2, with 53 and 93 events in the 90 days after an infection registered in CPRD AURUM and in SIDIAP, compared with 39.8 and 54.7 expected events, respectively. Equivalent standardised incidence 117 409 (47.9) 73 139 (45.6) 32 964 (20.6) 19 699 (12.3) 38 727 (24.2) 24 061 (15.0) 46 745 (29.2) Second dose (n=160 228) 10 776 (6.7) 12 998 (8.1) 15 854 (9.9) ratios were 1.33 (1.02 to 1.74) in CPRD AURUM and 54 (49-60) 7838 (4.9) 4804 (3.0) 6455 (4.0) 1.70 (1.39 to 2.08) in SIDIAP. The self-controlled case series analysis was only sufficiently powered to study those with a first dose of BNT162b2, ChAdOx1 nCoV-19, and mRNA-1273 (in 38 401 (15.7) 25 669 (10.5) 55 920 (22.8) 30 494 (12.5) 60 731 (24.8) (n=244 913) 13 012 (5.3) 14 622 (6.0) 20 231 (8.3) 10 122 (4.1) mRNA-1273 SIDIAP only), and those with SARS-CoV-2 infection 53 (47-57) 6292 (2.6) 7861 (3.2) First dose (see supplementary appendix tables 4 and 5). There was some indication of event dependent observation Table 2 | Baseline characteristics of study participants in Spain’s Information System for Research in Primary Care (SIDIAP) database periods (see supplementary appendix figure 1). In 15.3 (15.2-15.4) the self-controlled case series analysis, an adjusted 527 395 (42.1) 220 040 (17.6) 243 412 (19.4) 396 021 (31.6) 367 372 (29.3) 628 916 (50.2) 187 662 (15.0) (n=1 253 283) Second dose incidence rate ratio for Bell’s palsy of 0.81 (0.66 to 40 899 (3.3) 98 008 (7.8) 59 083 (4.7) 37 985 (3.0) 18 331 (1.5) 72 (55-79) 0.98) was observed for ChAdOx1 nCoV-19 and 0.83 (0.61 to 1.10) for BNT162b2 between one and 21 days in CPRD AURUM (table 5). The values in SIDIAP were 0.92 (0.60 to 1.33) and 0.83 (0.66 to 1.02) between 15.5 (15.4-15.5) 15.3 (15.2-15.4) 103 686 (44.9) 859 915 (45.4) 241 140 (12.7) 280 876 (14.8) 502 689 (26.6) 403 280 (21.3) 701 276 (37.0) 194 837 (10.3) (n=1 892 860) 123 581 (6.5) eight and 14 days (table 6). The adjusted incidence 49 029 (2.6) 68 891 (3.6) 45 250 (2.4) 24 460 (1.3) 56 (47-75) BNT162b2 First dose rate ratio for mRNA-1273 was 0.99 (0.54 to 1.64), and for SARS-CoV-2 infection was 1.82 (1.21 to 2.61). An increased risk of Bell’s palsy was observed on the day of a SARS-CoV-2 positive test result, and 1-21 days before 23 582 (10.2) 50 787 (22.0) 26 640 (11.5) 56 782 (24.6) and 1-7 days after the test result (table 5 and table 6). Second dose (n=230 854) 18 504 (8.0) 11 213 (4.9) 60 (42-63) 5075 (2.2) 5650 (2.4) 4851 (2.1) 5310 (2.3) 3277 (1.4) Encephalomyelitis Overall, 16 encephalomyelitis events were observed after a first dose of ChAdOx1 nCoV-19: standardised ChAdOx1 nCoV-19 15.3 (15.2-15.3) incidence ratio 1.45 (0.80 to 2.62) in CPRD AURUM and 270 830 (45.6) 161 047 (27.1) 200 957 (33.8) 60 074 (10.1) 80 492 (13.5) 87 455 (14.7) (n=594 134) 14 501 (2.4) 21 359 (3.6) 32 229 (5.4) 16 968 (2.9) 18 339 (3.1) 0.82 (0.34 to 1.97) in SIDIAP (table 3 and table 4). Nine 62 (59-65) 7097 (1.2) First dose events were observed after a first dose of BNT162b2 in SIDIAP (0.77, 0.40 to 1.48). No events were reported after a second dose of mRNA-1273, and fewer than five occurred among the remaining vaccination cohorts. General population Conversely, 28 encephalomyelitis events occurred after 2 287 403 (48.9) 1 139 337 (24.4) 11.0 (11.0-11.0) 863 167 (18.4) 568 547 (12.2) (n=4 678 512) 339 004 (7.2) 464 603 (9.9) 229 136 (4.9) 258 775 (5.5) 110 809 (2.4) SARS-CoV-2 infection: 11 observed to 1.6 expected Drug use (183 days before to four days before index date) 79 044 (1.7) 80 561 (1.7) Antineoplastic and immunomodulating agents 59 360 (1.3) 47 (36-63) (6.89, 3.82 to 12.44) in CPRD AURUM and 17 observed to 4.5 expected (3.75, 2.33 to 6.02) in SIDIAP. Guillain-Barré syndrome In CPRD AURUM, 11 Guillain-Barré syndrome events Median (IQR) past observation time (years) were observed after a first dose of ChAdOx1 nCoV- 19 and fewer than five after a second dose. The Antithrombotics and anticoagulants corresponding values in SIDIAP were fewer than five Malignant neoplastic disease and zero events. Five or fewer events were observed after the first and second dose of BNT162b2 and the Median (IQR) age (years) Systemic corticosteroids Lipid modifying agents Hypertensive disorder IQR=interquartile range. second dose of mRNA-1723, with no events after the Autoimmune disease Renal impairment Diabetes mellitus first dose of mRNA or Ad26.COV2.S. However, 26 Characteristics Comorbidities Heart disease events occurred after SARS-CoV-2 infection: nine observed to 2.6 expected (3.53, 1.83 to 6.77) events in Obesity CPRD AURUM and 17 observed to 2.9 expected (5.92, Men 3.68 to 9.53) events in SIDIAP. the bmj | BMJ 2022;376:e068373 | doi: 10.1136/bmj-2021-068373 5 RESEARCH Table 3 Association between covid-19 vaccination or SARS-CoV-2 infection and occurrence of immune mediated neurological disorders of special interest in UK’s Clinical Practice Research Datalink AURUM database No of Standardised incidence Event by vaccine and dose ­participants Person years Observed events Expected events ratio (95% CI) Bell’s palsy ChAdOx1 nCoV-19: First dose 3 776 803 384 250 117 164.5 0.71 (0.59 to 0.85) Second dose 1 093 258 218 629 25 94.7 0.26 (0.18 to 0.39) BNT162b2: First dose 1 693 453 275 333 46 116.4 0.40 (0.30 to 0.53) Second dose 1 166 571 235 351 24 99.5 0.24 (0.16 to 0.36) SARS-CoV-2 positive test result 446 851 106 342 53 39.8 1.33 (1.02 to 1.74) Encephalomyelitis ChAdOx1 nCoV-19: First dose 3 778 358 384 420 11 7.6 1.45 (0.80 to 2.62) BNT162b2: First dose 1 694 161 275 451 <5 5.4 Second dose 1 167 028 235 447 <5 4.6 SARS-CoV-2 positive test result 447 037 106 391 11 1.6 6.89 (3.82 to 12.44) Guillain-Barré syndrome ChAdOx1 nCoV-19: First dose 3 778 430 384 427 11 14.9 0.74 (0.41 to 1.33) Second dose 1 093 778 218 737 <5 9.5 BNT162b2: First dose 1 694 167 275 452 <5 10.4 Second dose 1 167 031 235 448 <5 9.0 SARS-CoV-2 positive test result 447 029 106 390 9 2.6 3.53 (1.83 to 6.77) Transverse myelitis ChAdOx1 nCoV-19: First dose 3 778 434 384 430 <5 7.6 BNT162b2: First dose 1 694 190 275 456 <5 4.7 SARS-CoV-2 positive test result 447 041 106 393 <5 1.9 Events with fewer than five occurrences have been omitted for privacy reasons. No events were observed for omitted cohorts. Transverse myelitis occurred in less than 0.5% participants across both Transverse myelitis events were only observed after databases (see supplementary appendix table 9). the first dose of ChAdOx1 nCoV-19 (in CPRD AURUM), first dose of BNT162b2 (in CPRD AURUM and SIDIAP) Discussion and second dose of mRNA-1723 (in SIDIAP), with In this study, which included more than eight million fewer than five events. Fewer than five events were also vaccine recipients from two countries, we observed no observed after SARS-CoV-2 infection in CPRD AURUM. safety signal between BNT162b2, ChAdOx1 nCoV-19, mRNA-1273, or Ad26.COV2.S and risk of developing Sensitivity analyses the immune mediated neurological events of Bell’s We found similar results in sensitivity analysis after palsy, encephalomyelitis, Guillain-Barré syndrome, excluding individuals with previous covid-19 among and transverse myelitis. An increased risk of Bell’s those vaccinated, removing the requirement for at least palsy, encephalomyelitis, and Guillain-Barré syndrome one year of history in all cohorts (see supplementary was, however, observed in two cohorts of 447 233 and appendix table 6). We obtained similar results after 288 637 people with SARS-CoV-2 infection. reducing the follow-up period from 90 to 21 days in the SARS-CoV-2-infected cohorts, although standardised Findings in context incidence ratios were higher and confidence intervals Immune mediated neurological disorders have been wider than those obtained in the main analysis. The identified as adverse events of special interest by first sensitivity analysis was also performed in the self- regulators, such as the Food and Drug Administration in controlled case series, with similar results obtained the US and the EMA in Europe, and organisations such to those of the main analysis (see supplementary as the Brighton Collaboration. These adverse events appendix table 7). The results from primary of special interest have been closely monitored during and secondary analyses are available at https:// immunisation campaigns,23 29 30 and several severe livedataoxford.shinyapps.io/VaxAEsNeuroimmune. neurological disorders were reported as rare adverse events during the first clinical trials of covid-19 vaccines.4 Hospital admissions and deaths after outcome events 5 31-35 Although a causal relationship was not established, In SIDIAP, the percentage of outcome events associated the recommendation for further monitoring has led to with hospital admission ranged from 35.9% for Bell’s the publication of many case reports worldwide.36 palsy to 86.3% for encephalomyelitis (see supplementary However, research findings from observational studies appendix table 8). Deaths in the first week after an event on the association between covid-19 vaccines and 6 doi: 10.1136/bmj-2021-068373 | BMJ 2022;376:e068373 | the bmj RESEARCH Table 4 | Association between covid-19 vaccination or SARS-CoV-2 infection and occurrence of immune mediated neurological disorders of special interest in Spain’s Information System for Research in Primary Care (SIDIAP) database Event by vaccine and dose No of participants Person years Observed events Expected events Standardised incidence ratio (95% CI) Bell’s palsy ChAdOx1 nCoV-19: First dose 592 365 89 055 27 82.6 0.33 (0.22 to 0.48) Second dose 230 692 55 659 10 47.3 0.21 (0.11 to 0.39) BNT162b2: First dose 1 890 434 102 623 100 116.7 0.86 (0.70 to 1.04) Second dose 1 251 680 74 638 85 97.1 0.88 (0.71 to 1.08) mRNA-1723: First dose 244 467 17 648 14 15.2 0.92 (0.54 to 1.55) Second dose 159 995 12 563 5 11.3 0.44 (0.18 to 1.06) Ad26.COV2.S: First dose 120 470 5,228 6 5.2 1.15 (0.52 to 2.56) SARS-CoV-2 positive test result 288 030 66 603 93 54.7 1.70 (1.39 to 2.08) Encephalomyelitis ChAdOx1 nCoV-19: First dose 592 843 89 121 5 6.1 0.82 (0.34 to 1.97) Second dose 230 844 55 696 <5 3.4 BNT162b2: First dose 1 892 409 102 738 9 11.7 0.77 (0.40 to 1.48) Second dose 1 253 202 74 733 <5 10.3 mRNA-1273: First dose 244 744 17 669 <5 1.2 Ad26.COV2.S: First dose 120 588 5,233 <5 0.4 SARS-CoV-2 positive test result 288 374 66 689 17 4.5 3.75 (2.33 to 6.02) Guillain-Barré syndrome ChAdOx1 nCoV-19: First dose 592 860 89 123 <5 5 BNT162b2: First dose 1 892 423 102 739 5 6.3 0.79 (0.33 to 1.91) Second dose 1 253 201 74 733 <5 5.3 mRNA-1273: Second dose 160 213 12,58 <5 0.7 SARS-CoV-2 positive test 288 392 66 693 17 2.9 5.92 (3.68 to 9.53) Transverse myelitis BNT162b2: First dose 1 892 510 102 744 <5 0.9 mRNA-1723: Second dose 160 222 12 581 <5 0.1 Events with fewer than five occurrences have been omitted for privacy reasons. No events were observed for omitted cohorts. immune mediated neurological events have been mixed. 15-21 days after vaccination. Moreover, this association For mRNA vaccines, for example, although two studies between ChAdOx1 nCoV-19 and Guillain-Barré underpinned by safety surveillance databases found no syndrome was subsequently confirmed using analogous association between mRNA vaccines and Bell’s palsy,37 38 methods with Scottish data.42 In our study we found no two others observed an increased risk,39 40 with the case such safety signals for ChAdOx1 nCoV-19. Meanwhile, series and nested case-control study from Hong Kong also in another study comparing events reported in a vaccine reporting an increased risk for recipients of an inactivated surveillance system with background rates, Ad26. vaccine (CoronaVac; Sinovac Biotech). Another study COV2.S was associated with increased risks of Guillain- from Hong Kong, however, found no such increased risk Barré syndrome (observed to expected rate ratio of 4.2 when comparing 1.1 million recipients of BNT162b2 or (3.5 to 5.0) in the 42 days after vaccination).43 In our CoronaVac with 2.7 million unvaccinated individuals.41 study, we observed no safety signal for Ad26.COV2.S, Our findings are in line with this study, with no safety but this vaccine was only available for the data from signal seen for either BNT162b2 or mRNA-1273. Spain, and was the smallest vaccine cohort. Although a study of more than 20 million people In line with our findings, several other studies have vaccinated with ChAdOx1 nCoV-19 and 12 million also reported increased risks of immune mediated vaccinated with BNT162b2 in the UK similarly did neurological events after SARS-CoV-2 infection.42 44 45 not observe any increased risk of hospital admissions One study, for example, found increased risks of Bell’s for neurological events with BNT162b2, it did find an palsy; encephalitis, meningitis, and myelitis; and association for ChAdOx1 nCoV-19.42 For example, the Guillain-Barré syndrome after SARS-CoV-2 infection.42 incidence rate ratios in the self-controlled case series The increase in risks after SARS-CoV-2 infection much analysis for Bell’s palsy and Guillain-Barré syndrome exceed any of the previously mentioned associations were 1.3 (1.1 to 1.6) and 2.9 (2.2 to 3.9), respectively, reported after vaccination. the bmj | BMJ 2022;376:e068373 | doi: 10.1136/bmj-2021-068373 7 RESEARCH Standardised incidence Standardised incidence ratio (95% CI) ratio (95% CI) Bell’s palsy ChAdOx1 nCoV-19 first dose ChAdOx1 nCoV-19 second dose BNT162b2 first dose BNT162b2 second dose mRNA-1273 first dose mRNA-1273 second dose Ad26.COV2.S first dose Covid-19 positive test result Encephalomyelitis ChAdOx1 nCoV-19 first dose BNT162b2 first dose Covid-19 positive test result Guillain-Barré syndrome ChAdOx1 nCoV-19 first dose BNT162b2 first dose Covid-19 positive test result 0 1 5 10 0 1 5 10 CPRD AURUM SIDIAP Fig 2 | Standardised incidence ratios of immune mediated neurological disorders of special interest. CPRD AURUM=Clinical Practice Research Datalink AURUM (UK); SIDIAP=Information System for Research in Primary Care (Spain) Limitations of this study historical cohorts might be limited because of changes Our study has limitations. Firstly, CPRD AURUM only over time, including seasonal variations47 in outcomes included primary care data from the UK. Therefore, and changes in exposure to other viral infections during diagnoses from inpatient settings might not be captured the pandemic, which might have reduced the incidence and the absolute risk could be underestimated. A of immune mediated neurological events.48 In previous study has, however, shown that CPRD AURUM addition, given public concerns over vaccine safety and captures immune mediated neurological disorders such complications from SARS-CoV-2 infection, individuals as Guillain-Barré syndrome relatively accurately, even vaccinated against or infected with SARS-CoV-2 might in the absence of linked hospital data.46 Meanwhile, have been more prone to seek care when showing the SIDIAP database did have patient level hospital symptoms. Although seasonality was not included in linkage, and results were consistent between databases. the historical comparison method, in our self-controlled Secondly, confounding by indication might affect the case series analysis we adjusted for age and seasonality historical comparator method. Although we account to control for time varying confounding. Fourthly, for differences in age, individuals vaccinated also had because we excluded individuals with a recent history more comorbidities than the comparator cohort and of the same immune mediated neurological events, our may well differ in other, unobserved characteristics. results are not generalisable to people with chronic Thirdly, comparisons between the vaccinated and relapsing conditions such as multiple sclerosis or Table 5 | Incidence rate ratios (95% confidence intervals) for covid-19 vaccination or SARS-CoV-2 infection and occurrence of Bell’s palsy using self-controlled case series analysis in UK’s Clinical Practice Research Datalink AURUM database ChAdOx1 nCoV-19 first dose BNT162b2 first dose SARS-CoV-2 infection No of Incidence rate ratio No of Incidence rate ratio No of Incidence rate ratio events (95% CI) events (95% CI) events (95% CI) Baseline 7403 Ref 4564 Ref 2335 Ref Days pre-vaccination/pre-infection: −21 to −1 83 0.68 (0.54 to 0.84) 41 0.74 (0.54 to 1.00) 14 1.03 (0.58 to 1.68) Day 0 5 0.83 (0.30 to 1.79) <5 <5 Days post-vaccination/post-infection: 1-7 25 0.59 (0.38 to 0.85) 14 0.75 (0.42 to 1.22) 6 1.30 (0.52 to 2.65) 8-14 34 0.81 (0.56 to 1.11) 17 0.91 (0.54 to 1.41) <5 15-21 43 1.03 (0.75 to 1.37) 16 0.84 (0.49 to 1.33) 7 1.51 (0.64 to 2.93) 1-21 102 0.81 (0.66 to 0.98) 47 0.83 (0.61 to 1.10) * 0.94 (0.51 to 1.55) 1-90 48 0.76 (0.56 to 1.01) Events with fewer than five occurrences have been omitted for privacy reasons. *Since event number <5 was blinded, the total number of events during the period was also blinded. 8 doi: 10.1136/bmj-2021-068373 | BMJ 2022;376:e068373 | the bmj RESEARCH Table 6 | Incidence rate ratios (95% confidence intervals) for covid-19 vaccination or SARS-CoV-2 infection and the occurrence of Bell’s palsy estimated using self-controlled case series analysis in in SIDIAP ChAdOx1 nCoV-19 BNT162b2 mRNA-1273 SARS-CoV-2 infection No of Incidence rate ratio No of Incidence rate ratio No of Incidence rate ratio No of Incidence rate ratio events (95% CI) events (95% CI) events (95% CI) events (95% CI) Baseline 4588 Ref 9150 Ref 3791 Ref 3892 Ref Days pre-vaccination/pre-infection: −21 to −1 12 0.44 (0.23 to 0.73) 74 0.69 (0.54 to 0.86) 11 0.80 (0.41 to 1.38) 44 2.93 (2.13 to 3.92) Day 0 <5 <5 <5 8 11.13 (5.07 to 20.80) Days post-vaccination/post-infection: 1-7 7 0.77 (0.33 to 1.49) 22 0.62 (0.39 to 0.91) <5 17 3.38 (2.01 to 5.28) 8-14 7 0.77 (0.33 to 1.50) 35 1.02 (0.72 to 1.40) 6 1.37 (0.54 to 2.79) 6 1.21 (0.48 to 2.47) 15-21 11 1.22 (0.63 to 2.10) 27 0.85 (0.57 to 1.22) <5 <5 1-21 25 0.92 (0.60 to 1.33) 84 0.83 (0.66 to 1.02) * 0.99 (0.54 to 1.64) * 1.82 (1.21 to 2.61) 1-90 77 1.31 (1.03 to 1.64) Events with fewer than five occurrences have been omitted for privacy reasons. *Since event number <5 was blinded, the total number of events during the period was also blinded. Devic’s disease. This exclusion criteria would also have the decision to submit for publication. TD-S, EB, and DP-A are guarantors. The corresponding author attests that all listed authors meet authorship led to a depletion-of-susceptible bias in the cohorts that criteria and that no others meeting the criteria have been omitted. received a second vaccine dose (because those who Funding: This work was partially funded by the UK National Institute had an event after the first dose were excluded from for Health Research (NIHR) and European Health Data and Evidence the second dose cohort). Fifthly, we assessed outcomes Network (EHDEN). EHDEN has received funding from the Innovative Medicines Initiative 2 Joint Undertaking under grant agreement No using routinely recorded diagnoses without a formal 806968. The Innovative Medicines Initiative 2 Joint Undertaking adjudication, and outcome misclassification might receives support from the European Union’s Horizon 2020 research and therefore have been possible. However, we used code innovation programme and The European Federation of Pharmaceutical Industries and Associations (EFPIA). The study funders had no role in the lists and algorithms for the identification of immune conceptualisation, design, data collection, analysis, decision to publish, or mediated neurological disorders previously published preparation of the manuscript. ER and EMH were supported by Instituto as part of a study on the background rates of covid-19 de Salud Carlos III, Spain (grant No CM20/00174, and JR17/00012). related adverse events of special interest.23 Competing interests: All authors have completed the ICMJE disclosure form at http://www.icmje.org/disclosure-of-interest/ and Our study focused on short term adverse events after declare the following interests: DP-A receives funding from the UK vaccination, using similar time periods to other studies.41 National Institute for Health Research (NIHR) in the form of a senior 42 The time at risk after vaccination was defined according research fellowship and from the Oxford NIHR Biomedical Research to the shortest recommended interval between first and Centre. XL receives the Clarendon Fund and Brasenose College scholarship (University of Oxford) to support her DPhil study. DP-A’s second vaccine doses. Although the second dose is not research group has received research grants from the European a consideration for Ad26.COV2.S, we used the same Medicines Agency; the Innovative Medicines Initiative; and Amgen, time at risk to ensure comparability between the results. Chiesi, and UCB Biopharma; and consultancy or speaker fees from Astellas, Amgen, AstraZeneca, and UCB Biopharma. Therefore, further evidence is required to understand the Ethical approval: The protocol for this research was approved by long term adverse events of vaccination and SARS-CoV-2 the independent scientific advisory committee for Medicine and infection. Larger cohorts are also needed to study the Healthcare Products Regulatory Agency database research (protocol effect of vaccination on different age groups, particularly No 20_000211) and by the clinical research ethics committee of Fundació Institut Universitari per a la recerca a l’Atenció Primària de among younger populations. Moreover, although it is Salut Jordi Gol i Gurina (project code: 21/052-PCV). Informed consent reassuring that so few outcomes of transverse myelitis of individual patients was not required as anonymised information were seen in this study, additional data might allow a was obtained from medical records. more detailed study of this event. Data sharing: Patient level data cannot be shared without approval from data custodians owing to local information governance and data protection regulations. Analytical code and detailed definitions Conclusion of algorithms for identifying the events are available in a GitHub We found no safety signal for any of the studied immune repository (https://github.com/SIDIAP/VaxAEsNeuroimmune). mediated neurological events after vaccination against The lead authors (XL and BR) affirm that this manuscript is an honest, accurate, and transparent account of the study being reported; that covid-19. Infection with SARS-CoV-2 was, however, no important aspects of the study have been omitted; and that any associated with an increased risk of Bell’s palsy, discrepancies from the study as planned (and, if relevant, registered) encephalomyelitis, and Guillain-Barré syndrome. have been explained. We thank healthcare professionals of the UK and Catalonia who daily Dissemination to participants and related patient and public register information in the populations’ electronic health records; the communities: We will disseminate a lay summary of our findings Institut Català de la Salut and the Programa d’analítica de dades per through our Twitter and other social media accounts. a la recerca i la innovació en salut for providing access to the different Provenance and peer review: Not commissioned; externally peer data sources accessible through SIDIAP; and Jennifer A de Beyer reviewed. (Centre for Statistics in Medicine, University of Oxford) for help with This is an Open Access article distributed in accordance with the English language editing. terms of the Creative Commons Attribution (CC BY 4.0) license, which Contributors: XL and BR are joint first authors. DP-A, EB, and TD-S are permits others to distribute, remix, adapt and build upon this work, joint senior authors. XL, BR, DP-A, TD-S, EB, and VS conceived the study for commercial use, provided the original work is properly cited. See: and contributed to the study design. XL, BR, ER, and AP conducted the http://creativecommons.org/licenses/by/4.0/. statistical analyses. XL, BR, ER, VS, DP-A, EB, and TD-S interpreted the results and wrote the manuscript. All authors contributed to writing the 1 WHO. COVID-19 Dashboard. Geneva: World Health Organization, manuscript, approved the final version, and had final responsibility for 2020. https://covid19.who.int (accessed 14 January 2022). the bmj | BMJ 2022;376:e068373 | doi: 10.1136/bmj-2021-068373 9 RESEARCH 2 European Medicines Agency. COVID-19 vaccines: authorised. www. 27 Suchard MA, Simpson SE, Zorych I, Ryan P, Madigan D. Massive ema.europa.eu/en/human-regulatory/overview/public-health- parallelization of serial inference algorithms for a complex threats/coronavirus-disease-covid-19/treatments-vaccines/ generalized linear model. ACM Trans Model Comput Simul 2013;23. vaccines-covid-19/covid-19-vaccines-authorised#authorised-covid- doi:10.1145/2414416.2414791 19-vaccines-section (accessed 14 January 2022). 28 Musonda P, Farrington CP, Whitaker HJ. Sample sizes for self- 3 Polack FP, Thomas SJ, Kitchin N, et al, C4591001 Clinical Trial Group. controlled case series studies. Stat Med 2006;25:2618-31. Safety and Efficacy of the BNT162b2 mRNA Covid-19 Vaccine. N Engl doi:10.1002/sim.2477 J Med 2020;383:2603-15. doi:10.1056/NEJMoa2034577 29 bc-coordinator. Priority list of adverse events of special interest: 4 Baden LR, El Sahly HM, Essink B, et al, COVE Study Group. Efficacy COVID-19. 2020. https://brightoncollaboration.us/priority-list-aesi- and Safety of the mRNA-1273 SARS-CoV-2 Vaccine. N Engl J covid/ (accessed 11 Mar 2021). Med 2021;384:403-16. doi:10.1056/NEJMoa2035389 30 Center for Biologics Evaluation and Research Office of Biostatistics 5 Voysey M, Clemens SAC, Madhi SA, et al, Oxford COVID Vaccine and Epidemiology. CBER Surveillance Program Background Trial Group. Safety and efficacy of the ChAdOx1 nCoV-19 vaccine Rates of Adverse Events of Special Interest for COVID-19 Vaccine (AZD1222) against SARS-CoV-2: an interim analysis of four Safety Monitoring Protocol. www.bestinitiative.org/wp-content/ randomised controlled trials in Brazil, South Africa, and the UK. uploads/2021/02/C19-Vaccine-Safety-AESI-Background-Rate- Lancet 2021;397:99-111. doi:10.1016/S0140-6736(20)32661-1 Protocol-FINAL-2020.pdf (accessed 11 Mar 2021). 6 Sadoff J, Gray G, Vandebosch A, et al, ENSEMBLE Study Group. Safety 31 Food and Drug Administration. Briefing Document. Pfizer-BioNTech and Efficacy of Single-Dose Ad26.COV2.S Vaccine against Covid-19. COVID-19 Vaccine. In: Vaccines and Related Biological Products N Engl J Med 2021;384:2187-201. doi:10.1056/NEJMoa2101544 Advisory Committee Meeting. www.fda.gov/media/144245/ 7 Heath PT, Galiza EP, Baxter DN, et al, 2019nCoV-302 Study Group. download (accessed 15 September 2021) Safety and Efficacy of NVX-CoV2373 Covid-19 Vaccine. N Engl J 32 Knoll MD, Wonodi C. Oxford-AstraZeneca COVID-19 vaccine efficacy. Med 2021;385:1172-83. doi:10.1056/NEJMoa2107659 Lancet 2021;397:72-4. doi:10.1016/S0140-6736(20)32623-4 8 European Medicines Agency. COVID-19 Vaccine Janssen: Guillain- 33 Sadoff J, Le Gars M, Shukarev G, et al. Interim Results of a Barré syndrome listed as a very rare side effect. www.ema.europa.eu/ Phase 1-2a Trial of Ad26.COV2.S Covid-19 Vaccine. N Engl J en/news/covid-19-vaccine-janssen-guillain-barre-syndrome-listed- Med 2021;384:1824-35. doi:10.1056/NEJMoa2034201 very-rare-side-effect (accessed 19 October 2021). 34 Márquez Loza AM, Holroyd KB, Johnson SA, Pilgrim DM, Amato 9 Matarneh AS, Al-Battah AH, Farooqui K, Ghamoodi M, Alhatou M. AA. Guillain-Barré Syndrome in the Placebo and Active Arms of a COVID-19 vaccine causing Guillain-Barre syndrome, a rare potential COVID-19 Vaccine Clinical Trial: Temporal Associations Do Not Imply side effect. Clin Case Rep 2021;9:e04756. doi:10.1002/ccr3.4756 Causality. Neurology 2021;10.1212/WNL.0000000000011881. 10 Razok A, Shams A, Almeer A, Zahid M. Post-COVID-19 vaccine doi:10.1212/WNL.0000000000011881 Guillain-Barré syndrome; first reported case from Qatar. Ann Med 35 Phase III Double-blind, Placebo-controlled Study of AZD1222 for the Surg (Lond) 2021;67:102540. doi:10.1016/j.amsu.2021.102540 Prevention of COVID-19 in Adults. NCT04516746. National Library 11 Waheed S, Bayas A, Hindi F, Rizvi Z, Espinosa PS. Neurological of Medicine (US). https://clinicaltrials.gov/ct2/show/NCT04516746 Complications of COVID-19: Guillain-Barre Syndrome Following Pfizer [cited 2021 Apr 1] COVID-19 Vaccine. Cureus 2021;13:e13426. doi:10.7759/cureus.13426 36 Food and Drug Administration. Briefing Document. Moderna COVID-19 12 Hughes DL, Brunn JA, Jacobs J, Todd PK, Askari FK, Fontana RJ. Vaccine. In: Vaccines and Related Biological Products Advisory Guillain-Barré syndrome after COVID-19 mRNA vaccination in a liver Committee Meeting. www.fda.gov/media/144434/download transplantation recipient with favorable treatment response. Liver 37 Klein NP, Lewis N, Goddard K, et al. Surveillance for Adverse Events Transpl 2022;28:134-7. doi:10.1002/lt.26279 After COVID-19 mRNA Vaccination. JAMA 2021;326:1390-9. 13 Goss AL, Samudralwar RD, Das RR, Nath A, ANA Investigates. ANA doi:10.1001/jama.2021.15072 Investigates: Neurological Complications of COVID-19 Vaccines. Ann 38 Renoud L, Khouri C, Revol B, et al. Association of Facial Paralysis With Neurol 2021;89:856-7. doi:10.1002/ana.26065 mRNA COVID-19 Vaccines: A Disproportionality Analysis Using the 14 Román GC, Gracia F, Torres A, Palacios A, Gracia K, Harris D. Acute World Health Organization Pharmacovigilance Database. JAMA Intern Transverse Myelitis (ATM):Clinical Review of 43 Patients With COVID- Med 2021;181:1243-5. doi:10.1001/jamainternmed.2021.2219 19-Associated ATM and 3 Post-Vaccination ATM Serious Adverse 39 Sato K, Mano T, Niimi Y, Toda T, Iwata A, Iwatsubo T. Facial nerve palsy Events With the ChAdOx1 nCoV-19 Vaccine (AZD1222). Front following the administration of COVID-19 mRNA vaccines: analysis Immunol 2021;12:653786. doi:10.3389/fimmu.2021.653786 of a self-reporting database. Int J Infect Dis 2021;111:310-2. 15 Nishizawa Y, Hoshina Y, Baker V. Bell’s palsy following the Ad26. doi:10.1016/j.ijid.2021.08.071 COV2.S COVID-19 vaccination. QJM 2021;114:657-8. doi:10.1093/ 40 Wan EYF, Chui CSL, Lai FTT, et al. Bell’s palsy following vaccination qjmed/hcab143 with mRNA (BNT162b2) and inactivated (CoronaVac) SARS-CoV-2 16 Zuhorn F, Graf T, Klingebiel R, Schäbitz WR, Rogalewski A. vaccines: a case series and nested case-control study. Lancet Infect Postvaccinal Encephalitis after ChAdOx1 nCov-19. Ann Dis 2022;22:64-72. doi:10.1016/S1473-3099(21)00451-5 Neurol 2021;90:506-11. doi:10.1002/ana.26182 41 Li X, Gao L, Tong X, et al. Autoimmune conditions following 17 Vogrig A, Janes F, Gigli GL, et al. Acute disseminated encephalomyelitis mRNA (BNT162b2) and inactivated (CoronaVac) COVID-19 after SARS-CoV-2 vaccination. Clin Neurol Neurosurg vaccination: a descriptive cohort study among 1.1 million 2021;208:106839. doi:10.1016/j.clineuro.2021.106839 vaccinated people in Hong Kong.MedRxiv 2021:21265314. 18 Wolf A, Dedman D, Campbell J, et al. Data resource profile: doi:10.1101/2021.10.21.21265314 Clinical Practice Research Datalink (CPRD) Aurum. Int J 42 Patone M, Handunnetthi L, Saatci D, et al. Neurological complications Epidemiol 2019;48:1740-1740g, g. doi:10.1093/ije/dyz034 after first dose of COVID-19 vaccines and SARS-CoV-2 infection. Nat 19 Herrett E, Gallagher AM, Bhaskaran K, et al. Data resource Med 2021;27:2144-53. doi:10.1038/s41591-021-01556-7 profile: Clinical practice research datalink (CPRD). Int J 43 Woo EJ, Mba-Jonas A, Dimova RB, Alimchandani M, Zinderman CE, Epidemiol 2015;44:827-36. doi:10.1093/ije/dyv098 Nair N. Association of Receipt of the Ad26.COV2.S COVID-19 Vaccine 20 Clinical Practice Research Datalink. CPRD Aurum June 2021 (Version With Presumptive Guillain-Barré Syndrome, February-July 2021. 2021.06.001). Clinical Practice Research Datalink. June 7,2021. JAMA 2021;326:1606-13. doi:10.1001/jama.2021.16496 doi:10.48329/PYC2-WE97 44 Paterson RW, Brown RL, Benjamin L, et al. The emerging spectrum of 21 Burn E, Fernández-Bertolín S, Voss EA, et al. Establishing and COVID-19 neurology: clinical, radiological and laboratory findings. characterising large COVID-19 cohorts after mapping the Information Brain 2020;143:3104-20. doi:10.1093/brain/awaa240 System for Research in Primary Care in Catalonia to the OMOP 45 Varatharaj A, Thomas N, Ellul MA, et al, CoroNerve Study Common Data Model. MedRxiv 2021:21266734. [cited 2021 Jan 5] Group. Neurological and neuropsychiatric complications of doi:10.1101/2021.11.23.21266734 COVID-19 in 153 patients: a UK-wide surveillance study. Lancet 22 Hripcsak G, Schuemie MJ, Madigan D, Ryan PB, Suchard MA. Drawing Psychiatry 2020;7:875-82. doi:10.1016/S2215-0366(20)30287-X reproducible conclusions from observational clinical data with OHDSI. 46 Stowe J, Andrews N, Wise L, Miller E. Investigation of the temporal Yearb Med Inform 2021;30:283-9. doi:10.1055/s-0041-1726481 association of Guillain-Barre syndrome with influenza vaccine and 23 Li X, Ostropolets A, Makadia R, et al. Characterising the background influenzalike illness using the United Kingdom General Practice Research incidence rates of adverse events of special interest for covid-19 Database. Am J Epidemiol 2009;169:382-8. doi:10.1093/aje/kwn310 vaccines in eight countries: multinational network cohort study. 47 Webb AJ, Brain SA, Wood R, Rinaldi S, Turner MR. Seasonal BMJ 2021;373;n1435. doi:10.1136/bmj.n1435 variation in Guillain-Barré syndrome: a systematic review, 24 Petersen I, Douglas I, Whitaker H. Self controlled case series meta-analysis and Oxfordshire cohort study. J Neurol Neurosurg methods: an alternative to standard epidemiological study designs. Psychiatry 2015;86:1196-201. doi:10.1136/jnnp-2014-309056 BMJ 2016;354:i4515. doi:10.1136/bmj.i4515 48 Wouk J, Rechenchoski DZ, Rodrigues BCD, Ribelato EV, Faccin-Galhardi 25 Fonseca-Rodríguez O, Fors Connolly AM, Katsoularis I, Lindmark LC. Viral infections and their relationship to neurological disorders. K, Farrington P. Avoiding bias in self-controlled case series studies Arch Virol 2021;166:733-53. doi:10.1007/s00705-021-04959-6 of coronavirus disease 2019. Stat Med 2021;40:6197-208. doi:10.1002/sim.9179 26 Sharples LD, Kirkwood BR. Essentials of medical statistics. J R Stat Supplementary information: additional tables 1-9 Soc Ser A Stat Soc 1989;152:263. doi:10.2307/2982934 and figures 1 and 2 10 doi: 10.1136/bmj-2021-068373 | BMJ 2022;376:e068373 | the bmj