A search of the medical English literature was conducted using PubMed/MEDLINE, EMBASE, and the Cochrane CENTRAL up to 1 August 2023. A thorough review was conducted on randomized controlled trials (RCTs) exploring the potential link between influenza vaccination and the subsequent risk of developing CVDs.
The search terms were: myocardial disease and influenza vaccines (Table S1). This study adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement (PRISMA) for its design and reporting (Table S2) (Prospero ID: CRD42023450694)11.
The records obtained from the database searches were amalgamated, and redundancy was addressed through employment of EndNote X7 (Thomson Reuters, Toronto, ON, Canada). Subsequently, two evaluators (MZ and MJN) carried out individual assessments of the records, considering both their title/abstract and full text, in order to exclude any entries that did not align with the study’s objectives.
The studies included in the analysis met the following criteria based on the PICOs:
Participants: The patients with a diagnosis of CVDs.
Intervention: Patients with CVDs who received influenza vaccine.
Comparison: Patients with CVDs who received a placebo.
Outcome: Lower risk of cardiovascular events.
Conference abstracts, case reports, and studies comparing high and low doses of influenza vaccination were excluded.
Data extraction was conducted by two investigators (MZ and MJN) and recorded in a Microsoft Excel spreadsheet (XP professional edition; Microsoft Corp, Redmond, WA). The extracted information encompassed various key elements, including the first author’s name, publication date, study type, definitions of cases and controls, the number of cases and controls, and the outcomes of the studies.
The quality assessment of the included studies was conducted by two reviewers (MJN and MZ) using the Cochrane tool12. In case of any discrepancies, a third reviewer (FO) was involved. This instrument encompasses a range of domains, which encompass random sequence generation, allocation concealment, blinding of both participants and personnel, blinding of assessors for outcomes, completeness of outcome data, and additional considerations such as selective reporting and potential sources of bias. Each study was categorized as having a low risk of bias when no concerns regarding bias were identified, a high risk of bias when there were concerns about bias or an unclear risk of bias when there was insufficient information available.
The combined outcomes were presented as risk ratios along with corresponding 95% confidence intervals (CIs). To evaluate the level of variability among the studies, both the I2 value and p-value were employed. In cases where minimal statistical heterogeneity was observed (I2 ≤ 50% or p ≥ 0.1), the fixed-effect model was applied. Conversely, if substantial inter-study heterogeneity was identified (I2 > 50% or p < 0.1), the random-effects model was utilized. Assessment of between-study heterogeneity involved the application of Cochran’s Q test and the I2 statistic. To assess potential publication bias, Begg’s test was employed, where a significance level of p < 0.05 indicated noteworthy publication bias. All analytical procedures were conducted using Comprehensive Meta-Analysis Software, Version 3.0 (Biostat, Englewood, NJ).