key: cord-0911660-1ggytw5q
authors: Miller, Elizabeth
title: Rapid evaluation of the safety of Covid-19 vaccines: how well have we done?()
date: 2022-01-06
journal: Clin Microbiol Infect
DOI: 10.1016/j.cmi.2021.12.018
sha: 5c83104449dacfbddcaa5a2499ee59e33c5c1305
doc_id: 911660
cord_uid: 1ggytw5q

nan

Vaccine safety surveillance has two complementary components -an initial signal generation step in which adverse events of interest are identified followed by analytic epidemiological studies to assess the association with vaccination. Safety signals can be generated from pre-licensure trials or during vaccine implementation from reports of unexpected or rare clinical events that occur in temporal association with vaccination. The accompanying paper in this edition by Shasha et al (1) provides a nice "real-world" example of this process. The authors used electronic records from the Meuhedet Health Maintenance Organisation (HMO) in Israel to investigate various safety signals raised for the BNT162b2 mRNA vaccine from pre-licensure trials and post-marketing case reports. Their study showed no difference between a vaccinated and matched unvaccinated cohort in the incidence of Bell's-palsy, consistent with the results of a case-control study from Hong-Kong (2) and a matched cohort study from another Israeli HMO, Clalit (3). However, a different analysis using the Clalit HMO dataset and a historical cohort as the comparator, rather than a contemporary matched unvaccinated cohort, showed a small excess of Bell's palsy cases within a month of receipt of BNT162b2 vaccine. Using standardised incidence ratios stratified by age, gender and dose, the study showed the highest attributable risk (4.46 per 100,000) was in females aged 65 years and over after a first dose (4). study. An obvious example is the risk of herpes zoster which is highly age and co-morbidity dependent, factors which are also included in deciding priority groups for vaccination. Use of a matched cohort or case control design, and/or adjustment for the effect of confounding variables by multivariate regression analysis, can help control for such biases but is only able to control for those factors that are captured and measured in the dataset. The self-controlled case series (SCCS) method is another way of controlling for individual level confounders as it assesses in a cohort of vaccinated individuals with the outcome event the incidence in a pre-defined post-vaccination period compared with the incidence of the same event outside the risk period (7) . While the SCCS method has been employed for assessing the risk of certain adverse events after Covid-19 vaccines (6) , its use as a method for rapid investigation of safety signals has been limited by inadequate observation time to assess the baseline incidence in post-vaccination periods after the putative risk period. When the period before Covid-19 vaccination is used to assess baseline incidence this can result in bias if the event is considered a contraindication to, or a reason to recommend, vaccination (8) . While the potential for unmeasured confounding inevitably remains in most observational safety studies of Covid-19 vaccines, when an elevated risk is shown for one vaccine but not another deployed in the same study population, as in those showing a risk of GBS with adenovirus vectored but not mRNA vaccines (5, 6) , the results are more likely to be valid . Similarly when elevated risks are shown for the same outcome with the same vaccine in different populations and with different study designs then the evidence for a causal association is strengthened.

Despite their limitations, the post-marketing safety surveillance systems in place in different countries have proven remarkably effective in identifying rare side effects associated with Covid-19 vaccines that could not detected in clinical trials. Acute myocarditis after mRNA vaccines and thrombosis with thrombocytopenia after adenovirus vector vaccines were rapidly identified as safety signals from passive reports from clinicians during the roll-out of the vaccines. Prompt investigation of these safety signals by epidemiologists using electronic health records (8, 9 ) , accompanied by The author declares no competing interests.

Real-world safety data for the Pfizer BNT162b2 SARS-CoV-2 vaccine: historical cohort study

Bell's palsy following vaccination with mRNA (BNT162b2) and inactivated (CoronaVac) SARS-CoV-2 vaccines: a case series and nested case-control study

Safety of the BNT162b2 mRNA Covid-19 Vaccine in a Nationwide Setting

Association between vaccination with the BNT162b2 mRNA COVID-19 vaccine and Bell's palsy: a population-based study. The Lancet regional health

Guillain-Barré Syndrome after COVID-19 Vaccination in the Vaccine Safety Datalink

Neurological complications after first dose of COVID-19 vaccines and SARS-CoV-2 infection

Use of the self-controlled case-series method in vaccine safety studies: review and recommendations for best practice

Risk of venous thrombotic events and thrombocytopenia in sequential time periods after ChAdOx1 and BNT162b2 COVID-19 vaccines: a national cohort study in England. Lancet Regional Health: Europe

Myocarditis after Covid-19 Vaccination in a Large Health Care Organization

Clinical features of vaccine induced immune thrombocytopenia and thrombosis

Communicating the potential benefits and harms of the Astra-Zeneca COVID-19 vaccine. The Winton Centre for Risk