(186) Appropriate Use of Observed Versus Expected Analyses for Newly Emerging Safety Concerns: The Case of Vaccine-Induced Thrombotic Thrombocytopenia Following COVID-19 Vaccination

Publication Number: 1179

Background: A signal of rare events of clotting with concurrent low platelets following adenovirus vector COVID-19 vaccines emerged in Europe in March 2021. The syndrome was later termed ‘Vaccine-induced Immune Thrombotic Thrombocytopenia’ (VITT). The site of clotting is commonly cerebral venous sinus thrombosis (CVST).

Objectives: To determine the applicability of Observed vs Expected (O/E) analysis for events such as VITT.

Methods: An evaluation of characteristics of VITT was conducted, including the biological basis and an assessment of frequency and severity of cases of thrombosis with concurrent thrombocytopenia reported spontaneously to the United Kingdom’s (UK) Yellow Card Scheme following COVID-19 Vaccine AstraZeneca. Publicly available documentation from regulatory authorities were examined, to establish whether O/E analyses were cited to determine or explain the association between COVID-19 Vaccine AstraZeneca and VITT.

Results: In the UK, 443 cases of VITT following COVID-19 Vaccine AstraZeneca had been spontaneously reported to the Yellow Card scheme up to 25 May 2022, of which 18.3% had a reported fatal outcome. Events more frequently occurred following the first dose of COVID-19 Vaccine AstraZeneca compared with the second dose. No trends were observed by age or sex. In the UK, the incidence rate (IR) of CVST with concurrent thrombocytopenia from 2019 (i.e. pre-pandemic) has been estimated as 0.00 (95% Confidence Interval 0.00-0.11) per 100,000 person years. However, many regulators were using the concept of O/E assessments to determine or explain the association between VITT and COVID-19 vaccines. Characteristics of VITT, including for example positive PF4 “HIT” and highly elevated D-dimer, suggest that this type of assessment was not the most appropriate method to evaluate or communicate this signal.

Conclusions: VITT is a newly defined event, which appears specific to the adenovirus vector COVID-19 vaccines. There was therefore no background incidence of VITT, and by definition the event requires exposure to a vaccine. While O/E analysis is an important method for signal detection and management, it may not have been the most appropriate method to assess the occurrence of VITT following the COVID-19 vaccines. It is important to consider the characteristics of the event of interest when selecting the best method for signal assessment.