The Brighton Collaboration's GAIA initiative developed harmonised case definitions (CDs) intended to standardise safety monitoring in studies of maternal immunisation worldwide. These definitions use levels of diagnostic certainty (LOC) to enable applicability across settings with differing diagnostic capacity. Prior work, including a 2023 review, identified that around half of GAIA CDs had not been evaluated and that several key outcomes performed poorly, with many potential cases only classifiable at LOC-4. The present review aimed to update the evidence on real-world performance and implementation challenges of the GAIA maternal immunisation CDs.
The authors searched Ovid MEDLINE, EMBASE and Web of Science for English-language research studies published from 2014 through May 5, 2025, that assessed the performance of the GAIA case definitions in routine care or research settings. Two reviewers independently screened studies and extracted data, resolving discrepancies by consensus. The review protocol was registered with PROSPERO (CRD42025620433).
Database searches initially identified 4,260 records. After deduplication and a two-stage screening process, 10 studies were eligible for inclusion. Across these studies a total of 42,587 cases were assessed using GAIA definitions. The included literature covered a mixture of outcomes and settings, though several definitions have only been evaluated in high-income contexts according to the report.
The review found heterogeneity in how well individual GAIA CDs performed when applied to routine care or research data. Definitions for maternal death, neonatal death, and preterm birth generally performed well, with many potential cases classifiable at higher LOCs. By contrast, several other outcomes showed high proportions of unclassifiable cases, with many cases only achieving LOC-4.
Across studies there was marked variability: some investigations were able to classify 100% of potential cases at LOC 1–3, while others could classify as few as 50% at LOC 1–3. This variation indicates substantial heterogeneity in the ability to apply GAIA definitions across different study designs, data sources and health system contexts.
The review identified three GAIA definitions that had not been evaluated in the included literature: ectopic pregnancy, postpartum endometritis, and neurodevelopmental delay. Several other definitions have been evaluated only in high-income settings, leaving gaps in evidence from low- and middle-income countries and from routine data sources with limited diagnostic detail.
The authors highlight several likely contributors to variability in classification performance. Differences in setting-level resources, availability of diagnostic tools, and the quality and completeness of clinical documentation and data sources were cited as important determinants. The review also notes that some limitations may be inherent to the definitions themselves, which could reduce applicability across diverse health-care contexts.
Because the GAIA CDs rely on graded levels of diagnostic certainty, their performance is sensitive to how much clinical information and objective testing are available. Where documentation is sparse or diagnostics are limited, a larger proportion of potential events are only classifiable at LOC-4.
The systematic review concludes that while progress has been made in evaluating GAIA maternal immunisation case definitions, important gaps and challenges remain. Variability in performance across settings and outcomes underscores the need for targeted refinement of specific definitions and for broader validation efforts across diverse geographic and resource settings and across different data sources. Such work would support more robust, standardised and comparable safety surveillance for maternal immunisation globally.
The review emphasizes the need to prioritise evaluation of the unevaluated definitions (ectopic pregnancy, postpartum endometritis, neurodevelopmental delay) and to expand assessments into low- and middle-income settings and routine-care data sources to improve generalisability.
The authors declared that they have no known competing financial interests or personal relationships that could have influenced the reported work.