The current SCCS study, based on Israel national databases, examined whether risk for acute cardiovascular events and mortality within 30 days of administration of the 2nd Pfizer vaccine dose against COVID19, was higher than the risk during the next 30 days. The results indicated that there was no substantial differences with respect to acute cardiovascular events (excluding myocarditis), even after accounting for potential moderators. No conclusions could be drawn with respect to mortality, which seemed to decrease substantially shortly after vaccine administration, a decrease that is most probably caused by bias, which we were unable to control.
Acute cardiovascular events (excluding myocarditis)
Our results with respect to acute cardiovascular events (excluding myocarditis) are in accordance with a recent meta-analysis which focused on cardiovascular safety of these vaccines in real-world studies [25]; the authors reported no increased risk for myocardial infarction (risk ratio, RR = 0.98, 95%CI 0.87–1.09) or arrhythmia (RR = 0.98, 95%CI 0.84–1.12) following COVID19 vaccination. They did disclose increased moyocarditis occurrence, especially following 2nd and 3rd vaccine doses and in mRNA vaccines [25]. Another retrospective cohort study (with over 220,000 participants) focused on thrombotic outcomes in COVID19 vaccinees, and its results supported the safety of mRNA vaccines with respect to such events [26].
Several former SCCS studies from different countries, using different outcome definitions and follow up periods, largely reported similar results. A French SCCS, focusing on pulmonary emboli, acute MI, hemorrhagic or ischemic stroke in COVID19 vaccinees aged < 75y, found no association with the administration of mRNA COVID19 vaccines during a 21d follow-up [17]. In New Zealand, a SCCS study design was used to assess the risk for different thromboembolic events following the administration of Pfizer COVID19 vaccine during a 21d follow-up. The researchers reported no risk differences, even when stratified by ethnic origin [21]. Furthermore, a SCCS study from Hong-Kong which focused on high-risk individuals with known cardiovascular disease and explored a composite outcome of acute MI, stroke, revascularization and cardiovascular death, reported no evidence of an increased risk of these events after vaccination with Pfizer COVID19 vaccine [19]. These conclusions were further consolidated by a recent Israeli SCCS study, based on the dataset of the Clalit healthcare fund which covers around 50% of the Israeli population [22]. The study population included individuals at risk for COVID19 complications (e.g., aged 60 years or older, pregnant women, residents in nursing homes, and people with particular medical conditions such as chronic lung disease, chronic renal failure, hypertension and chronic heart failure) and focused on booster doses of the Pfizer vaccine. Primary outcome measures were the cumulative number of non-COVID19 hospitalizations and the cumulative number of non-COVID19 cardiovascular hospitalizations (including, among others, thromboembolic events and stroke, acute MI, arrhythmias, cardiac failure and myocarditis). Follow-up periods were 7d, 14d, 28d and 42d following vaccination, and the analyses were stratified by vaccine dose (first or second monovalent booster or first bivalent booster). No signal was identified in any of these analyses, and the authors concluded that these results further reassure the safety of the vaccine in high-risk individuals [22].
Thus, most current evidence, including this study, support the null hypothesis of no substantial differences in the probability for acute cardiovascular events between the risk and the control periods following vaccine administration. Myocarditis is an exception, and this fact is also supported by our sensitivity analysis, which did indicate a signal for the younger age group when myocarditis was included as an outcome in the sensitivity test.
Mortality
The question of potentially higher risk for all-cause mortality in the short-term following administration of the 2nd COVID19 Pfizer vaccine dose could not be settled in this study. Our results indicated, in fact, a much lower risk for mortality in the short-term following vaccination. As mentioned earlier, a bias cannot be ruled out. First, there is an inherent bias due to death being a one-time event. While an acute cardiovascular event may occur prior to the designated follow-up time, and also in both the risk and the control periods, death occurs only once and this occurrence prevents any other. However, since death is a relatively rare event, we do not expect this bias to be large. On the other hand, Healthy Vaccinee Bias, indicating a lower mortality rate among vaccinees vs. non-vaccinees for reasons other than the administration of the vaccine itself, may be more relevant and cannot be ruled out [27]. This effect continued up to 50 days post vaccination.
Two former publications implied that COVID19 vaccines may be associated with higher risk for cardiac arrest [15] or cardiovascular mortality [16], fueling public concerns through scientific, popular and social media. The 2022 ecological study by Sun et al. [15] relied on a 2019–2021 dataset from the Israel National Emergency Medical Services and disclosed an increase of over 25% in calls due to cardiac arrest and acute coronary syndrome in 16–39y old subjects during January–May 2021, compared with the years 2019–2020. The authors reported that weekly emergency call counts were significantly associated with the rates of the 1st and 2nd vaccine dose administered to this age group but not with COVID19 infection rates; however, these associations were purely ecological. Furthermore, the validity of the recorded main reason for these calls is unknown, and the study dataset excluded other emergency services in Israel [15]. It should also be mentioned, that no increase in sudden cardiac death in the young age group was noted during this time period in Israel, based on national data. Sun et al.'s [15] results were not replicated by another ecological study which relied on national databases and looked at out-of-hospital cardiac arrest (OHCA) in young people (1-50y) in Victoria, Australia [20]. The main results did not demonstrate increased rates of overall OHCA, myocarditis causing OHCA, or unascertained OHCA during the pandemic or after the introduction of COVID19 vaccination. Furthermore, causes of death in young people experiencing fatal OHCA within 30d of their COVID19 vaccination (any dose) were consistent with pre-pandemic causative profiles for this age group [20].
The SCCS report published by the Florida Health Department [16] indicated no differences in all-cause mortality in the first 28d following vaccination as compared to the control period, but rather a statistically significant mortality decrease in participants aged 60 + y (relative incidence = 0.97, 95%CI 0.94–0.99). On the other hand, cardiac-related mortality risk was significantly higher during the risk period for males (relative incidence = 1.09, 95%CI = 1.03–1.15) but not for females, and this effect was mostly seen for mRNA vaccines (male relative incidence = 1.11, 95% CI = 1.05–1.18) [16]. However, this report was not published in a peer-reviewed journal. A recent meta-analysis used three different SCCS publications on all-cause and cardiac-related mortality following COVID19 vaccination [28]. While the pooled hazard ratio (HR) revealed no significant association of COVID19 vaccination with all-cause mortality (HR = 0.89, 95%CI 0.71–1.10), the pooled HR for cardiac-related mortality was significantly increased (HR = 1.06, 95%CI 1.02–1.11), mainly in males [28]. Interestingly, these results were mostly driven by the Florida's Health Department report [16], since the other two SCCS studies included in this meta-analysis [29, 30] showed no increase in mortality risk in the short-term period following COVID19 vaccine administration; the Italian SCCS study [29] focused on mRNA vaccines only and showed no increase for all-cause mortality on days 3,7,14 and 30 following vaccine administration [29] while the British SCCS study [30] focused on several types of COVID19 vaccines, and disclosed no significant increase in all-cause or cardiac mortality in the 12 weeks following COVID19 vaccination [30]. Thus, despite our study inability to draw any conclusions on mortality following Pfizer mRNA COVID19 2nd vaccine dose, the current scientific evidence support no such association.
Our study has several limitations. First, the follow up period spanned only 60 days following vaccine administration thus potential effects that may have occurred later than that were not captured. However, adverse effects of vaccines are usually expected to present earlier, not later, and the occurrence of acute myocarditis within 7 days of the 2nd dose administration [9–11] is a good example. Second, we focused on the 2nd vaccine dose only. The main reason for that was, that vaccination regime in Israel was based on a 3 weeks gap between the 1st and the 2nd dose thus follow-up after the 1st vaccine dose would have been confounded by the administration of the next dose. In addition, our previous studies on acute myocarditis as an adverse effect of COVID19 mRNA vaccines indicated that it was mainly observed in 2nd dose vaccinees, and usually within the first seven days following vaccination [9–11]. Another limitation of the current study is that we only had data on acute cardiovascular events which caused hospitalizations. Having said that, it should be kept in mind that most serious effects do require hospitalization. Furthermore, we introduced only few potential moderators into our models due to lack of information on others. However, the inherent SCCS study design reduces confounding by time-insensitive covariates. Last, we were unable to study the effect of the 2nd dose administration on all-cause mortality, probably due to bias. We also had no information of specific death causes, since the relevant data for the year 2021 are not yet available in Israel.
On the other hand, the study has some advantages too; it was based on the total population in Israel and on national validated databases. Its design accounted for potential confounding by time-insensitive individual covariates, such as comorbidities, and we were able to include in our models moderators such as former cardiovascular hospitalizations and seasonal risk. Last, we were able to exclude COVID19 positive individuals, thus reducing potential confounding by acute infection and by Long-COVID19 status.