Vaccine safety surveillance during a pandemic is challenging. In 2020, the global spread of the SARS-CoV-2 virus to pandemic proportions brought about the rapid development of vaccines to fight it. At the end of 2020 and beginning of 2021, the first COVID-19 vaccines became available for use in large, often population-wide, vaccination campaigns. Although large clinical trials had been conducted for these vaccines, there is still a risk that rare or previously unknown adverse reactions to the vaccine could still occur when administered at the scale of a global vaccination campaign.
As for any drug or vaccine that has been newly released to the market, risk groups for developing Adverse Events Following Immunisation (AEFIs) are not yet well established, and potential long-term effects are not yet entirely known. During large-scale vaccination campaigns, it can be difficult to differentiate between events that occur coincidentally within a plausible time frame after vaccination and those that are actually caused by the vaccine. Another consideration is that vaccines are biological products that need special storage conditions, so there is a risk of safety issues relating to incorrect storage or transportation of vaccine batches that may occur and need to be traced quickly before the vaccines are administered to large groups of people.
The spontaneous reporting system in the Netherlands is maintained by the pharmacovigilance (PV) centre Lareb. In the Netherlands, the COVID-19 vaccination campaign started on 6 January 2021, and a high volume of AEFI reports followed. To date, more than 234,000 reports have been made, with a total of more than 1,130,000 adverse reactions reported. Vaccinated people themselves reported the majority of these adverse reactions.
Since Lareb had expected a high volume of reports that needed to be processed and assessed within a short period, it developed a dedicated system to allow for near real-time processing of reports and vaccine safety monitoring. This started with a specific COVID-19 vaccine web-based reporting form with prespecified questions for case assessments. AEFIs that were listed in the product information for the vaccine at the time of marketing authorisation were prespecified on the reporting form to enable faster processing. Vaccine brands could be chosen from a prespecified list as well. Closed questions specific to COVID-19 were also asked, for example whether the individual had been previously infected with COVID-19. By working with this structured reporting form, more than 33% of all incoming reports could be processed automatically in the first year of immunisation. Those who had been vaccinated were given a leaflet with information on their vaccine after each vaccination and this allowed them to fill in the batch number of their vaccine on the reporting form. When not known, and with permission from the reporter, batch numbers were retrieved from the national vaccination registry maintained by the National Institute for Public Health and the Environment (RIVM).
The remaining 67% of incoming reports that could not be automatically processed were triaged daily by a team of PV assessors with expertise in vaccine AEFIs and signal detection. They selected high-priority cases based on the seriousness of the adverse event, listing them as an adverse event of special interest or signal value. These cases were then assessed by a team of dedicated vaccine experts. Brighton Collaboration case definitions were used for these assessments. In the first year of the COVID-19 campaign, 4% of cases were deemed high priority.
Since Lareb had expected a high volume of reports that needed to be processed and assessed within a short period, it developed a dedicated system to allow for near real-time processing of reports and vaccine safety monitoring.
Case-by-case analysis of high-priority cases remained a driving force of the signal-detection work. Daily signal detection meetings were held to discuss high-priority cases, for instance, those of thrombosis with thrombocytopenia syndrome (TTS) after vaccination with the Oxford/AstraZeneca vaccine.
Assessors could consult with an external clinical advisory board consisting of medical specialists working in clinical practice in fields such as immunology, haematology, and vascular medicine. As not all reports could be manually assessed, a custom-built line listing was made to review weekly vaccine data and reporting rates calculated from the number of reports received against vaccinations given in the Netherlands. These line listings could, for instance, focus on specific populations, such as teenagers and children. Although it was also incorporated as a feature in the line-listing, the influence of such large amounts of COVID-19 vaccine AEFI data on disproportionality analysis was not known. Therefore, for the pandemic vaccination campaign, disproportionality was not used as a primary trigger for further analysis. In addition, a batch analysis was performed where reporting patterns of batches of the same brand of vaccine were compared. The batch analysis triggered no signal requiring action so far.
An automated standardized weekly report was generated by combining data from the spontaneous reporting system with that of a large cohort event monitoring study, which was started in early 2021 and involved monitoring individuals for 6 months following vaccination.
When much of your population is being vaccinated, it is important to consider the background incidence of clinical events of interest to identify potential signals. The background incidence rate of many potential events had already been mapped out in the European project ACCESS, so these were used for Observed/Expected analysis of adverse events. Stratification of background incidences by age groups and sex was also performed. Data on the vaccinated population were provided by the RIVM based on the Dutch vaccination registry. Additional background incidences specific to the Dutch population were provided on request by the PHARMO Institute based on electronic patient records of general practitioners as well as hospital data.
Following the initial COVID-19 immunisation campaign, booster vaccination campaigns were held using vaccines adapted to other strains of the virus. However, the SARS-CoV-2 virus has not been eradicated yet and may require future vaccination campaigns. It is important that national pharmacovigilance centers learn from the experience of the past two years and identify areas where innovation could help in future data management and signal detection.
