A comparison of four self-controlled study designs in an analysis of COVID-19 vaccines and myocarditis using five European databases (preprint)

IntroductionThe aim of this study was to assess the possible extent of bias due to violation of a core assumption (event-dependent exposures) when using self-controlled designs to analyse the association between COVID-19 vaccines and myocarditis. MethodsWe used data from five European databases (Spain: BIFAP, FISABIO VID, and SIDIAP; Italy: ARS-Tuscany; England: CPRD Aurum) converted to the ConcePTION Common Data Model. Individuals who experienced both myocarditis and were vaccinated against COVID-19 between 1 September 2020 and the end of data availability in each country were included. We compared a self-controlled risk interval study (SCRI) using a pre-vaccination control window, an SCRI using a post-vaccination control window, a standard SCCS and an extension of the SCCS designed to handle violations of the assumption of event-dependent exposures. ResultsWe included 1,757 cases of myocarditis. In unadjusted analyses, agreement between study designs varied by vaccine brand. There was good agreement between all designs for AstraZeneca and Pfizer, but for Moderna we found harmful incidence rate ratios (IRR) using the standard and extended SCCS (standard SCCS: IRR = 3.12, 95%CI = 1.53 - 6.40; extended SCCS: IRR = 2.43, 95%CI = 1.11 - 5.33) compared with no association with the SCRIs (SCRI-pre: IRR = 0.60, 95%CI = 0.27 - 1.33; SCRI-post: IRR = 0.86, 95%CI = 0.34 - 2.19), although confidence intervals were wide. There was very good agreement between all designs for the unadjusted second dose analyses, confirming the known harmful association between the second dose of Moderna and Pfizer vaccines and myocarditis. ConclusionsIn the context of the known association between COVID-19 vaccines and myocarditis, we have demonstrated that two forms of SCRI and two forms of SCCS led to largely comparable results, possibly because of limited violation of the assumption of event-dependent exposures.

The Validation Of Covid-19 information In The Pharmacoepidemiological Research Database for Public Health System by vaccination status (preprint)

Purpose: To validate Covid-19 information records in The Pharmacoepidemiological Research Database for Public Health System (BIFAP), commonly used for pharmacoepidemiological research in Spain. Methods: The recorded Covid-19 cases in primary care (PC) or positive test registries (gold-standard) were identified among vaccinated patients against SARS-CoV-2 infection of any age. They were matched with unvaccinated controls by birth year, vaccination date, region, and sex, between December 2020-October 2021. The sensitivity (SE), specificity (SP), positive (PPV), negative (NPV) predictive values, and date accurateness were estimated for PC by vaccination status and age brands. Results: Among 21,702 patients with positive tests and 20,866 with recorded Covid-19 diagnoses, the SE, SP, PPV, and NPV were, respectively, 79.98%, 99.95%, 80.24% and 99.94% among vaccinated, and 78.67%, 99.96%, 84.51% and 99.94% among controls. For those aged ≥70 years old, SE (71.15-72.85%) was lower while PPV (84.68-88.04%) was higher compared to <70 years old participants. 94.12% of the total true positive cases (N=17,191) were recorded within ±5 days from the date of the test result. Conclusions: PC Covid-19 diagnosis recorded in BIFAP showed high validation parameters. SE was similar and PPV was slightly lower among vaccinated than unvaccinated controls. Correction of vaccines effectiveness estimates by such misclassification is recommended. Data shows the influence of age. Among the elderly, Covid-19 diagnosis was less recorded but when recorded is more accurate than among younger patients. These findings permit the design of informed algorithms for performing Covid-19-related research.

Cohort monitoring of 29 Adverse Events of Special Interest prior to and after COVID-19 vaccination in four large European electronic healthcare data sources (preprint)

Setting Primary and/or secondary health care data from four European countries: Italy, the Netherlands, the United Kingdom, Spain Participants Individuals with complete data for the year preceding enrollment or those born at the start of observation time. The cohort comprised 25,720,158 subjects. Interventions First and second dose of Pfizer, AstraZeneca, Moderna, or Janssen COVID-19 vaccine. Main outcome measures 29 adverse events of special interest (AESI): acute aseptic arthritis, acute coronary artery disease, acute disseminated encephalomyelitis (ADEM), acute kidney injury, acute liver injury, acute respiratory distress syndrome, anaphylaxis, anosmia or ageusia, arrhythmia, Bells’ palsy, chilblain-like lesions death, erythema multiforme, Guillain Barré Syndrome (GBS), generalized convulsion, haemorrhagic stroke, heart failure, ischemic stroke, meningoencephalitis, microangiopathy, multisystem inflammatory syndrome, myo/pericarditis, myocarditis, narcolepsy, single organ cutaneous vasculitis (SOCV), stress cardiomyopathy, thrombocytopenia, thrombotic thrombocytopenia syndrome (TTS) venous thromboembolism (VTE) Results 12,117,458 individuals received at least a first dose of COVID-19 vaccine: 54% with Comirnaty (Pfizer), 6% Spikevax (Moderna), 38% Vaxzevria (AstraZeneca) and 2% Janssen Covid-19 vaccine. AESI were very rare <10/100,000 PY in 2020, only thrombotic and cardiac events were uncommon. After adjustment for factors associated with severe COVID, 10 statistically significant associations of pooled incidence rate ratios remained based on dose 1 and 2 combined. These comprised anaphylaxis after AstraZeneca vaccine, TTS after both AstraZeneca and Janssen vaccine, erythema multiforme after Moderna, GBS after Janssen vaccine, SOCV after Janssen vaccine, thrombocytopenia after Janssen and Moderna vaccine and VTE after Moderna and Pfizer vaccines. The pooled rate ratio was more than two-fold increased only for TTS, SOCV and thrombocytopenia. Conclusion We showed associations with several AESI, which remained after adjustment for factors that determined vaccine roll out. Hypotheses testing studies are required to establish causality.