https://orcid.org/0000-0003-2835-3601
Figures
Abstract
Background
Tranexamic acid (TXA) is an antifibrinolytic agent commonly used to mitigate blood loss across various medical indications. Despite its widespread use, comprehensive data on its safety profile remain limited. This study aimed to systematically evaluate adverse events (AEs) associated with TXA.
Methods
Adverse event reports were extracted from the U.S. Food and Drug Administration’s Adverse Event Reporting System (FAERS) and the VigiAccess databases. Disproportionality analyses were conducted using reporting odds ratio (ROR), proportional reporting ratio (PRR), the Medicines and Healthcare products Regulatory Agency (MHRA) method, Bayesian confidence propagation neural network (BCPNN), and multi-item gamma Poisson shrinker (MGPS).
Results
A total of 17,787 TXA-related AE reports were identified. A higher proportion of reports involved females, with older adults (≥ 65 years) accounting for the largest proportion in FAERS and younger individuals (18–44 years) in VigiAccess. Overlapping PTs, including seizures, pulmonary embolism and anaphylactic reactions, were identified. Significant differences for TXA-related AEs were found by gender, age and death outcomes. Most AEs occurred within the first month, with an early failure pattern.
Citation: Feng J, Guo C, Zhao S (2026) Real‑world safety evaluation of tranexamic acid: Signal detection from FAERS and VigiAccess databases. PLoS One 21(7): e0353459. https://doi.org/10.1371/journal.pone.0353459
Editor: Mohammed Misbah Ul Haq, Deccan School of Pharmacy, INDIA
Received: January 12, 2026; Accepted: June 23, 2026; Published: July 10, 2026
Copyright: © 2026 Feng et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Raw datasets obtained from public sources were analyzed in this study. VigiAccess data can be accessed via the VigiAccess website (https://www.vigiaccess.org). FAERS data can be accessed via the FDA Industry Systems (FIS) website (https://fis.fda.gov/extensions/fpd-qde-faers/fpd-qde-faers.html). The authors confirm that they had no special access privileges that others would not have when attempting to access these datasets. Researchers can replicate the study findings by accessing these databases directly and following the methods described in the Methods section. Furthermore, all minimal datasets used to reach the conclusions are included within the paper and its Supporting Information files.
Funding: This work was sponsored and funded by the China International Medical Foundation (No.: Z-2014-08-2309-5), the Henan Medical Science and Technology Program (No.: LHGJ20240058) and the Henan Provincial Science and Technology Research Program (No.: 262102311069). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Tranexamic acid (TXA), an antifibrinolytic agent that inhibits plasminogen activation, has been widely used in clinical practice for decades [1]. Since its introduction in the 1960s, TXA has demonstrated efficacy in reducing blood loss across a variety of medical scenarios, including trauma, surgery, obstetrics, gynecology, and other bleeding disorders, making it an essential tool in the global management of hemorrhage [2,3]. Moreover, with its inclusion in the World Health Organization (WHO)’s essential medicines list, TXA has become a cornerstone for managing hemorrhagic conditions globally [4].
However, current evidence on TXA’s safety primarily comes from clinical trials [4–6], which are often constrained by small sample sizes, short follow-up periods, or narrowly defined patients. Real-world pharmacovigilance studies provide a complementary perspective to identify adverse events (AEs) signals related with TXA in broader populations, including those not studied in pre-marketing trials [7]. While numerous studies have confirmed TXA’s hemostatic benefits [8,9], its safety profile, including AE signal distribution, subgroup-specific risks, and time-to-onset (TTO) patterns, has yet to be characterized, particularly across multiple pharmacovigilance databases.
In this study, we aim to assess TXA’s safety profile using the U.S. Food and Drug Administration’s Adverse Event Reporting System (FAERS) and VigiAccess database. By using disproportionality algorithms and stratified analyses, we sought to identify significant AE signals at both the system organ class (SOC) and preferred term (PT) levels, determine subgroup-specific risks stratified by age, gender, and fatal outcomes, and perform TTO analysis of TXA-related AEs. This multidimensional approach provides evidence for clinical decision-making and helps to optimize the management of hemorrhagic disorders.
Materials and methods
Data source
Data were collected from two pharmacovigilance databases: the FAERS database, which covers reports from the first quarter of 2004 to the third quarter of 2024, and the VigiAccess database, which includes data from the drug’s market introduction to December 29, 2024. In FAERS, spontaneous AE reports were submitted by healthcare professionals, patients, and drug manufacturers [10], while VigiAccess is a user-friendly portal, allowing to search drug safety reports received by the Uppsala Monitoring Centre from around the world [11]. Although individual case safety reports (ICSRs) are not publicly accessible via VigiAccess, aggregated drug–event frequency data were retrieved from the portal. Approximate contingency tables were reconstructed from these aggregate counts to support exploratory signal detection.
The search was conducted using generic nomenclature, with AEs systematically categorized and encoded according to the Medical Dictionary for Regulatory Activities (MedDRA, version 27.1) [12]. These coded terms were arranged in a hierarchy of categories to facilitate standardized classification of specific categories (e.g., narrow PT and broad SOC).
Since this study was based on anonymous data available to the public, institutional review board approval and informed patient consent were not necessary.
Study design
This study employed a case/non-case approach, similar to a case-control design, to assess the safety profile of TXA [13]. For FAERS, a total of 21,964,449 reports were initially retrieved. Deduplication was performed following FDA guidance using three fields from the DEMO table: PRIMARYID, CASEID, and FDA_DT. For reports sharing the same CASEID, the record with the most recent FDA_DT was retained; where both CASEID and FDA_DT were identical, the record with the highest PRIMARYID was kept. After deduplication, 3,686,206 duplicate records were excluded, yielding 18,278,243 unique reports for background analysis, of which 1,780 were TXA-related AE reports included in this study.
For VigiAccess, as only aggregate-level frequency counts are available rather than individual ICSRs, approximate 2 × 2 contingency tables were constructed from the reported drug–event frequencies to enable exploratory disproportionality analysis. These results are considered complementary to FAERS-based signals. After deduplication and data screening, a total of 17,787 TXA-related AE reports were identified from the two databases for further analysis.
Within this cohort, descriptive analysis and disproportionality methods were utilized to identify TXA-related AE signals. If the proportion of AEs is higher in patients exposed to TXA (cases) compared to those who were not exposed (non-cases), a correlation between the drug and the event can be assumed, suggesting a disproportionality signal. It should be noted that this design does not adjust for underlying indication or disease severity. As TXA is predominantly used in high-acuity settings, observed signals may partly reflect the background risk of the treated population rather than drug-specific effects, and should be interpreted as hypothesis-generating.
Descriptive analyses
Descriptive analyses were performed to summarize the demographic characteristics of TXA-related AE reports. Variables included gender, age, reporting year, route of administration, reported countries, indications, outcomes, weight, etc. Categorical variables were presented as frequencies and percentages, while continuous variables with non-normal distributions were reported as medians with interquartile ranges (IQR).
Disproportionality analysis
Formal disproportionality analysis was conducted on FAERS data using a combination of five algorithms, based on individual-level ICSRs and full contingency tables. For VigiAccess, exploratory signal detection was performed using approximate contingency tables reconstructed from aggregated frequency counts, with the same algorithm parameter thresholds as FAERS. The five algorithms applied were: reporting odds ratio (ROR), proportional reporting ratio (PRR), Medicines and Healthcare Products Regulatory Agency (MHRA) criteria, Bayesian confidence propagation neural network (BCPNN), and the multi-item gamma Poisson shrinker (MGPS) algorithm (S1 Table and S2 Table) [14–18]. A potential positive signal was identified if TXA demonstrated a positive association across all five algorithms: (1) For ROR, the frequency of AE occurrences (a) ≥ 3, with the lower bound of the 95% confidence interval (CI) for ROR > 1; (2) For PRR, a ≥ 3, with the lower bound of the 95% CI for PRR > 1; (3) For MHRA, a ≥ 3, PRR ≥ 2, and chi-square (χ2) ≥ 4; (4) For BCPNN, the lower bound of the 95% CI for the information component (IC025) > 0; (5) For MGPS, the lower bound of the 95% CI for the empirical Bayesian geometric mean (EBGM05) > 2. This conservative approach requires simultaneous positivity across all five algorithms, combined with important medical event (IME) classification, to maximize specificity and minimize false-positive signals. Although this strategy may reduce sensitivity, concordance across distinct algorithms strengthens the evidence for retained signals. PTs meeting signal thresholds with 3–9 reports are acknowledged as potential signals but were not included in the primary tables due to statistical imprecision. These represent areas for future investigation.
To further refine the analysis, an IME list was introduced at the PT level. The IME list, published by the European Union and updated biannually in alignment with MedDRA (version 27.1), was used to refine PT-level signals; only PTs meeting predefined thresholds and classified as IMEs were included in the final analysis. To ensure statistical stability, PTs with fewer than 10 total reports were excluded from the final tabulated results.
Difference analysis
In FAERS, difference analysis was conducted to explore potential variations in AE signals across demographic and clinical parameters. Specifically, analyses were performed by gender (male vs. female), age (< 65 years vs. ≥ 65 years), and fatal outcomes (death vs. non-death). To visualize these differences, “volcano plots” were constructed, providing a graphical representation of AE signal variations. In the volcano plots, the -log10 of the P-value was plotted on the y-axis, while the log2 of the ROR was displayed on the x-axis. Between-subgroup differences in ROR were assessed for each PT using chi-square tests, with Yates’ continuity correction or Fisher’s exact test applied as appropriate. To account for multiple comparisons, Benjamini–Hochberg false discovery rate (FDR) correction was applied across all PTs within each subgroup analysis. The volcano plots display unadjusted P-values for visual clarity; FDR-corrected P-values were used to determine statistical significance, with FDR-adjusted P < 0.05 considered significant.
TTO analysis
To examine variations in AE occurrences over time, we performed TTO analysis. TTO was defined as the interval between TXA initiation and AE occurrence, calculated from drug start date and event date fields. Records with missing, or erroneous dates were excluded. Cases with a TTO of 0 days were retained, as same-day events are clinically plausible for intravenous administration (e.g., infusion-related reactions). However, we acknowledge that 0-day records may also reflect date imprecision or reporting conventions rather than true same-day onset. Of the 1,780 FAERS reports, 391 (21.97%) contained evaluable TTO data and were included in this analysis; the remaining 1,389 (78.03%) were excluded due to missing date information. Cumulative distribution curves were plotted to visualize TTO patterns, and gender-stratified analyses were performed to explore potential differences.
In addition to median onset times and IQR, Weibull’s shape parameter (WSP) test was applied to classify TTO patterns [19,20]. The WSP includes parameters of scale (α) and shape (β). The shape parameter β was performed to classify AE risk patterns: (1) Early failure, characterized by a decreased AE hazard over time (β < 1, and 95% confidence interval [CI] < 1); (2) Random failure, characterized by a constant AE hazard over time (β is equal to or close to 1, and 95% CI contains 1); (3) Wear-out failure, characterized by an increased AE hazard over time (β > 1, and 95% CI > 1) [21].
Furthermore, in TTO analysis, we introduced Standardized MedDRA Queries (SMQs). The SMQs consist of groupings of MedDRA terms ordinarily at the PT level that relate to a defined medical condition or area of interest [22].
Statistics analysis and reporting guidelines
All statistical analyses and visualizations were carried out using SAS software (version 9.4; SAS Institute Inc., Cary, NC), R software (version 4.4.2; R Foundation for Statistical Computing, Vienna, Austria), Python (version 3.10; Python Software Foundation, Wilmington, United States), and Prism (version 9.5; GraphPad Software, San Diego, CA). A P-value < 0.05 was considered statistically significant.
This study followed the REporting of A Disproportionality analysis for drUg Safety signal detection using individual case safety reports in PharmacoVigilance (READUS-PV) guidelines (S3 Table) [23].
Results
Descriptive analysis
A total of 17,787 TXA-related AE reports were collected from the two pharmacovigilance databases: 1,780 reports from FAERS and 16,007 reports from VigiAccess (Table 1). The VigiAccess database provided demographic characteristics based on four dimensions: gender, age, reporting year, and continent. Excluding reports with unknown gender, the female-to-male ratio for TXA-related AEs was approximately 1.80–2.09 times higher in both databases (53.26% vs. 29.66% in FAERS; 65.96% vs. 31.63% in VigiAccess). Regarding age, FAERS reported a median age of 52.5 years (IQR: 37.0–70.0), with patients aged 65 and older accounting for the highest reporting proportion of TXA-related AEs (25.51%, n = 454). Conversely, in VigiAccess, the 18–44 age group reported the highest proportion of TXA-related AEs, representing 36.62% (n = 5,862). In terms of geographical distribution, the majority of reports came from the Americas (45.11%) in FAERS and Asia (72.14%) in VigiAccess. TXA-related AE reports exhibited an increasing trend in recent years (S1 Fig and S2 Fig), with more than half of the reports from both databases occurring between 2019 and 2024 (FAERS: 60.11%; VigiAccess: 57.46%).
Further, in FAERS, over four-fifths of all reports (85.06%) were submitted by healthcare professionals, with the majority of reports originating from the United States (39.38%). The most common route of administration was not specified (49.72%), followed by intravenous (20.90%). Notably, 94.27% of the reports were classified as serious, with hospitalization being the most frequently reported outcome (32.02%). In terms of indications, excluding not specified or unknown indications, the most common TXA-related indication was heavy menstrual bleeding (n = 204), followed by hemorrhage (n = 136) and hemorrhage prophylaxis (n = 94).
Distribution at SOC level for TXA
AEs were categorized based on the SOC classification. In VigiAccess, the most frequently reported AEs were related to gastrointestinal disorders (n = 7,386, 23.50%), skin and subcutaneous tissue disorders (n = 4,861, 15.47%), and general disorders and administration site conditions (n = 4,489, 14.28%). In contrast, in FAERS, the top three SOC categories for TXA-related AEs were nervous system disorders (n = 972, 18.81%), injury, poisoning, and procedural complications (n = 685, 13.25%), and general disorders and administration site conditions (n = 547, 10.58%). Despite differences in ranking, a considerable overlap was observed in the SOC levels between the two databases (Fig 1).
TXA, Tranexamic acid; FAERS, The U.S. Food and Drug Administration (FDA)’s Adverse Event Reporting System; SOC, System organ class.
Distribution at PT level for TXA
We analyzed positive signals at the PT level, focusing on PTs with at least 10 reports. The cumulative number of reports for TXA-related PTs was 1,109 in FAERS and 2,807 in VigiAccess. A total of 34 PT signals were identified in FAERS, while 25 positive PT signals were observed in VigiAccess. The analysis highlighted the most frequently reported PTs and those with the highest signal intensity.
In FAERS (S3 Fig), the three most frequently reported PTs for TXA were seizure (n = 151), pulmonary embolism (PE) (n = 119), and generalized tonic-clonic seizure (GTCS) (n = 89). The top three PT signals with the highest ROR were myoclonic epilepsy (ROR = 165.01), coronary artery thrombosis (ROR = 123.15), vascular stent thrombosis (ROR = 118.45) (Table 2). In VigiAccess (S4 Fig), the three most frequently reported PT signals for TXA were cardiac flutter (n = 813), anaphylactic reaction (n = 299), anaphylactic shock (n = 280). The PT signals with the highest ROR in VigiAccess were renal cortical necrosis (ROR = 542.55), myoclonic epilepsy (ROR = 37.36), cardiac flutter (ROR = 35.99) (Table 3). Of the 25 PT signals identified in VigiAccess, 15 (60%) were also detected in FAERS. When ranked by case numbers, the top three overlapping PTs were seizure (n = 390), PE (n = 364), and anaphylactic reaction (n = 351). In addition to the overlapping PTs, 19 unique PTs were reported in FAERS, while 10 unique PTs were identified in VigiAccess (Fig 2).
PT, Preferred term; TXA, Tranexamic acid; FAERS, The U.S. Food and Drug Administration (FDA)’s Adverse Event Reporting System; CVT, Cardiac ventricular thrombosis; MODS, Multiple organ dysfunction syndrome; DIC, Disseminated intravascular coagulation; GTCS, Generalised tonic-clonic seizure.
Difference analysis
In exploratory subgroup analyses, females showed higher reporting disproportionality for headache, PE, and back pain, whereas males demonstrated higher signals for hypertension, hypoxia, and GTCS (Fig 3A). Age-based analysis suggested that individuals aged ≥ 65 years may have higher disproportionality signals for AEs such as GTCS, atrial fibrillation, and seizure, while younger individuals (< 65 years) showed higher reporting proportions for nausea, rash, and headache (Fig 3B). These subgroup findings are hypothesis-generating and should be interpreted with caution given the exploratory nature of the analyses. In the analysis of fatal outcomes, certain AEs were more frequently reported in association with death. Cardiac arrest, multiple organ dysfunction syndrome (MODS), and PE showed higher reporting disproportionality in the death group, while acute myocardial infarction (AMI) was more frequently reported in the non-death group (Fig 3C). These findings should be interpreted with caution, as patients who died may differ substantially from survivors in terms of disease severity, clinical indication, comorbidities, and concomitant medications.
Note: Statistical significance was assessed using chi-square tests (with Yates’ continuity correction or Fisher’s exact test for sparse data), with Benjamini–Hochberg false discovery rate (FDR) correction applied for multiple comparisons (FDR-adjusted P < 0.05). FAERS, The U.S. Food and Drug Administration (FDA)’s Adverse Event Reporting System; TXA, Tranexamic acid; ROR, Reporting odds ratio; GTCS, generalised tonic-clonic seizure; PE, Pulmonary embolism; DVT, Deep vein thrombosis; MODS, Multiple organ dysfunction syndrome; AKI, Acute kidney injury; AMI, Acute myocardial infarction.
TTO analysis
In FAERS, after removing duplicates and erroneous reports, 391 cases provided TTO data. The majority of AEs occurred within the first month of treatment (n = 350, 89.51%) (Fig 4A), with a median time to event of 0 days (IQR 0.0–5.0 days) (Fig 4B). However, TTO data were available for only 391 of 1,780 cases (21.97%) and a substantial proportion had a recorded onset of 0 days, which may partly reflect same-day reporting conventions or date imprecision rather than true immediate onset. Among cases with evaluable TTO, Weibull shape parameter values (β < 1) were consistent with an early failure pattern across all SMQ categories (Table 4).
AE, Adverse event; TXA, Tranexamic acid; IQR, Interquartile range.
Among 367 cases with gender data, AE reports were more frequent in females (n = 260) than in males (n = 107) (Fig 4C). Notably, the cumulative incidence of AEs showed a significant difference between males and females. The median onset time for females was 1 day (IQR: 0.0–6.0 days), compared to 0 days (IQR: 0.0–2.0 days) for males (Wilcoxon test, P = 0.0014) (Fig 4D).
Discussion
This study represents the first pharmacovigilance investigation to jointly utilize the FAERS and VigiAccess databases for analyzing TXA-related AEs. To our knowledge, this is among the largest pharmacovigilance analyses of TXA to date, encompassing 17,787 AE reports from two independent databases. By leveraging multidimensional pharmacovigilance tools, we identified demographic trends, AE distributions, and safety signals. Our findings offer a detailed perspective on TXA-related AEs, emphasizing safety considerations for clinical decision-making.
A notable finding was the predominance of AE reports among females, suggesting potential gender differences in AE reporting patterns. TXA is widely prescribed for bleeding-related disorders that disproportionately affect women, such as heavy menstrual bleeding, postpartum hemorrhage, and other gynecological conditions [24]. This widespread use in female-specific conditions may partly explain the observed sex differences. Regarding age distribution, older adults (≥ 65 years) were more represented in FAERS, while younger adults (18–44 years) predominated in VigiAccess. Additionally, differences in reporting culture, indication mix, and the inclusion of non-prescription use across regions may further contribute to this discrepancy. Given these structural differences, findings from the two databases should be interpreted independently, with VigiAccess results serving as exploratory and complementary evidence. The 60% concordance rate between the two databases for PT-level signals suggests reasonable reproducibility of key safety signals across independent pharmacovigilance systems. However, both databases share the inherent limitations of spontaneous reporting, including indication bias and absence of exposure denominators.
Our analysis also revealed a clear temporal trend, with over 50% of AE reports occurring between 2019 and 2024. This aligns with the increasing acceptance and expanded indications for TXA, which has demonstrated well-documented benefits in a variety of clinical settings, not only for trauma, but can also be utilized across multiple specialties to manage hemorrhage [25]. However, the increase of TXA-related AE reports may partly reflect expanded use across broader clinical contexts, which warrants pharmacovigilance attention [26].
The SOC distribution for TXA revealed both similarities and differences between the two databases. In VigiAccess, gastrointestinal disorders, skin and subcutaneous tissue disorders, and general disorders and administration site conditions were the most frequently reported, whereas in FAERS, stronger signals were observed for nervous system disorders and injury, poisoning, and procedural complications. Importantly, some findings were consistent with TXA’s mechanism of action, particularly its role in modulating fibrinolysis [27]. Theoretically, TXA might be expected to increase the risk of thrombosis and subsequent vascular and cardiovascular complications due to its antifibrinolytic activity [28]. However, data from other studies [29–31] suggested no significant association between TXA administration and the rate of thrombosis-related complications, and the protective effects of TXA were maintained. The SOC-level analysis highlights the importance of vigilance in monitoring AEs across multiple organ systems, particularly in high-risk patients with conditions such as active thromboembolic disease or imbalances in thrombosis and hemostasis [26].
The PT-level analysis further identified specific safety concerns associated with TXA. Shared PT signals, such as seizure, PE, and anaphylactic reaction, underscore the necessity of monitoring for neurological complications, thromboembolism, and skin symptoms during TXA treatment. TXA-related seizures may be attributed to its role as a competitive antagonist of the gamma-aminobutyric acid receptor type A (GABAA). TXA crosses the blood–brain barrier and may directly excite neurons, thereby lowering the seizure threshold [32]. In in vitro and animal experiment, higher TXA level in the cerebral spinal fluid, correlated with serum concentration, was associated with the incidence of seizures [32,33]. And indeed, some clinical studies indicated that high-dose TXA administration was linked to a higher incidence of seizures [28,34]. Beyond common PTs, certain serious AEs, like myoclonic epilepsy, renal cortical necrosis, and coronary artery thrombosis, exhibited exceptionally high ROR values. However, these signals were based on relatively small case numbers, and high ROR values with low absolute counts are susceptible to statistical instability; they should therefore be interpreted with caution and regarded as signals warranting further investigation rather than confirmed safety concerns.
It should be noted that TXA is predominantly used in high-acuity clinical settings, including trauma, major surgery, and critical illness, where thromboembolic events (e.g., PE, AMI), cardiac arrest, and MODS are complications of the underlying condition rather than necessarily drug-related effects. The observed disproportionality signals for these PTs may therefore be subject to confounding by indication, whereby the severity of the hemorrhagic condition or the surgical context, rather than TXA exposure itself, contributes to AE reporting. In the absence of indication-restricted analyses or an active comparison (e.g., other hemostatic agents), these signals should be interpreted with caution.
Furthermore, TXA is approved across heterogeneous indications, including heavy menstrual bleeding, surgical hemorrhage prophylaxis, and trauma-associated hemorrhage, with signals pooled across these different clinical contexts. As FAERS does not consistently capture structured indication data for all reports, signals across these indications may obscure indication-specific safety profiles. Future pharmacovigilance studies with indication-stratified analyses would provide more refined signal characterization.
Our study evaluated disproportionality signal differences across subgroups stratified by age, gender, and fatal outcomes. Females showed higher reporting disproportionality for headache and PE, while males exhibited stronger signals for hypertension and GTCS. Older individuals (≥ 65 years) demonstrated higher disproportionality signals for GTCS and atrial fibrillation, whereas younger individuals (< 65 years) showed higher reporting proportions for nausea and rash, and more death-related AEs caused by cardiac arrest and MODS. These differences in reporting patterns may result from individual variations, underlying comorbidities, or differential prescribing patterns. Factors like hormone levels, gastrointestinal physiology, and hepatorenal function may influence drug absorption, distribution, metabolism, and excretion, potentially contributing to sex- and age-related differences in AE profiles [35]. Of note, the death versus non-death comparison is particularly susceptible to confounding. Patients with fatal outcomes likely had more severe underlying conditions, more complex surgical or trauma contexts, and greater exposure to concomitant medications, any of which could independently contribute to the observed AEs. The higher disproportionality signals for cardiac arrest, MODS, and PE in the death group therefore cannot be attributed to TXA exposure per se, and should not be interpreted as evidence that TXA directly causes these fatal events.
TXA-related AEs were characterized by early failure patterns, with over four-fifths of AEs reports occurring within the first month of treatment, and gradually decreased with prolonged exposure time. This finding aligns with TXA’s pharmacokinetics, which exerts its effects rapidly after administration [36]. However, in the context of TXA’s use in perioperative and acute hemorrhagic settings, early-onset AEs may also reflect the severity of underlying conditions, surgical interventions, or concomitant therapies rather than direct drug toxicity. The predominance of 0-day TTO records, while clinically plausible for intravenous administration, may additionally reflect date imprecision or same-day reporting conventions. Therefore, the early failure pattern observed here should be interpreted as a reporting pattern rather than a pharmacological characteristic of TXA. Besides, gender differences were also observed in TTO analysis. Compared with males, females exhibited a longer median TTO, which may reflect sex-based differences in TXA pharmacodynamics or underlying pathophysiology.
The findings of this study have important clinical implications. The disproportionality signals for thromboembolic and neurological events suggest that enhanced monitoring may be warranted, particularly in patients with a history of thrombotic events or seizure disorders. Hypersensitivity reaction signals, including anaphylaxis, suggest the importance of emergency preparedness during TXA administration, particularly in acute settings. Moreover, close monitoring during the initial treatment period may be necessary, as TTO analysis identified a higher AE reports in this period; however, this reflects reporting patterns rather than confirmed incidence rates. Special attention should be focused to high-risk populations, including older adults, females, and individuals with pre-existing comorbidities, to minimize the risk of adverse outcomes.
Limitations
This study has several limitations. First, the reliance on spontaneous reporting databases introduces potential reporting biases, including underreporting, over-reporting, and incomplete data. Second, substantial missingness was present in key covariates, including route of administration, TTO, and body weight. Analyses involving these variables may introduce selection bias due to limited representativeness of available subsets. In particular, age data were missing for 438 of 1,780 FAERS reports (24.61%), which may have introduced selection bias into age-based subgroup analyses. Third, the median TTO of 0 days may reflect date imprecision, or reporting conventions, particularly given that route of administration was unspecified in nearly half of reports. The early failure pattern should therefore be considered hypothesis-generating. On the other hand, same-day AEs in perioperative settings may reflect the underlying condition or surgical context rather than direct drug effects.
Fourth, VigiAccess provides only aggregated drug–event frequency data without access to individual ICSRs; therefore, VigiAccess-derived signals were based on approximate contingency tables and should be regarded as exploratory findings requiring further validation. Fifth, confounding by indication is a major limitation of this analysis. Many identified signals, including PE, AMI, cardiac arrest, and MODS, are well-recognized complications of trauma, surgery, and critical illness. It is not possible to determine whether these signals reflect true drug effects or are attributable to the severity of the underlying condition. Sixth, signals were pooled across heterogeneous TXA indications (e.g., heavy menstrual bleeding, surgical prophylaxis, trauma), which may mask indication-specific safety patterns.
Seventh, the restriction of tabulated PT-level results to those with ≥10 reports, while consistent with standard practice for statistical stability, may have excluded rarer signals with 3–9 reports that could be clinically meaningful. Future studies with larger datasets may be better powered to characterize low-frequency signals. Additionally, subgroup comparisons involved simultaneous testing across multiple PTs. Although FDR correction was applied, the exploratory nature of these analyses and small subgroup samples limit the reliability of individual PT-level findings. Finally, disproportionality analysis, while valuable for signal detection, cannot provide definitive evidence of a causal (or inverse) relationship between drug exposure and AEs. Further validation from independent data sources, along with insight into potential mechanisms and prevention of TXA-related AEs, are essential to confirm the causal nature of these signals.
Conclusion
This study provides a comprehensive assessment of TXA-related safety profile based on real-world pharmacovigilance data. By analyzing key demographic trends, SOC and PT distributions, specific high-risk populations, and TTO patterns, we characterized the multidimensionality of TXA-related AE reporting patterns. Thromboembolic, neurological, and hypersensitivity events demonstrated stronger disproportionality signals, emphasizing the importance of close monitoring during the initial period. Subgroup analyses identified differential reporting patterns by age and gender. However, these represent hypothesis-generating signals rather than confirmed differential incidence risks, and should be interpreted in the context of the inherent limitations of spontaneous reporting data. Future studies are warranted to validate these findings and further optimize its clinical application.
Supporting information
S1 Table. Four table of measure of disproportionality.
https://doi.org/10.1371/journal.pone.0353459.s001
(DOCX)
S2 Table. ROR, PRR, MHRA, BCPNN, and MGPS methods, formulas, and thresholds.
https://doi.org/10.1371/journal.pone.0353459.s002
(DOCX)
S1 Fig. Annual distribution of AE reports for TXA in VigiAccess.
https://doi.org/10.1371/journal.pone.0353459.s004
(TIF)
S2 Fig. Annual distribution of AE reports for TXA in FAERS.
https://doi.org/10.1371/journal.pone.0353459.s005
(TIF)
S3 Fig. The top 50 Signal strength of AEs at the PT level ranked by ROR for TXA in VigiAccess.
https://doi.org/10.1371/journal.pone.0353459.s006
(TIF)
S4 Fig. The top 50 Signal strength of AEs at the PT level ranked by ROR for TXA in FAERS.
https://doi.org/10.1371/journal.pone.0353459.s007
(TIF)
Acknowledgments
We would like to thank FAERS and WHO for the free access of data, and allowing the use of it.
References
- 1. Gruen RL, Mitra B, Bernard SA, McArthur CJ, Burns B, et al, PATCH-Trauma Investigators and the ANZICS Clinical Trials Group. Prehospital Tranexamic Acid for Severe Trauma. N Engl J Med. 2023;389(2):127–36. pmid:37314244
- 2. Murao S, Nakata H, Yamakawa K. Safety of tranexamic acid in thrombotic adverse events and seizure in patients with haemorrhage: a protocol for a systematic review and meta-analysis. BMJ Open. 2020;10(6):e036020. pmid:32571860
- 3. Ockerman A, Vanassche T, Garip M, Vandenbriele C, Engelen MM, Martens J, et al. Tranexamic acid for the prevention and treatment of bleeding in surgery, trauma and bleeding disorders: a narrative review. Thromb J. 2021;19(1):54. pmid:34380507
- 4. Klein A, Agarwal S, Cholley B, Fassl J, Griffin M, Kaakinen T, et al. A review of European guidelines for patient blood management with a particular emphasis on antifibrinolytic drug administration for cardiac surgery. J Clin Anesth. 2022;78:110654. pmid:35065393
- 5. Ker K, Sentilhes L, Shakur-Still H, Madar H, Deneux-Tharaux C, Saade G, et al. Tranexamic acid for postpartum bleeding: a systematic review and individual patient data meta-analysis of randomised controlled trials. Lancet. 2024;404(10463):1657–67. pmid:39461793
- 6. Karl V, Thorn S, Mathes T, Hess S, Maegele M. Association of Tranexamic Acid Administration With Mortality and Thromboembolic Events in Patients With Traumatic Injury: A Systematic Review and Meta-analysis. JAMA Netw Open. 2022;5(3):e220625. pmid:35230436
- 7. Lawati KA, Sharif S, Maqbali SA, Rimawi HA, Petrosoniak A, Belley-Cote EP, et al. Efficacy and safety of tranexamic acid in acute traumatic brain injury: a systematic review and meta-analysis of randomized-controlled trials. Intensive Care Med. 2021;47(1):14–27. pmid:33079217
- 8. Shi J, Zhou C, Pan W, Sun H, Liu S, Feng W, et al. Effect of High- vs Low-Dose Tranexamic Acid Infusion on Need for Red Blood Cell Transfusion and Adverse Events in Patients Undergoing Cardiac Surgery: The OPTIMAL Randomized Clinical Trial. JAMA. 2022;328(4):336–47. pmid:35881121
- 9. Reuben A, Appelboam A, Stevens KN, Vickery J, Ewings P, Ingram W, et al. The Use of Tranexamic Acid to Reduce the Need for Nasal Packing in Epistaxis (NoPAC): Randomized Controlled Trial. Ann Emerg Med. 2021;77(6):631–40. pmid:33612282
- 10. Morris R, Ali R, Cheng F. Drug Repurposing Using FDA Adverse Event Reporting System (FAERS) Database. Curr Drug Targets. 2024;25(7):454–64. pmid:38566381
- 11. Bhardwaj K, Alam R, Pandeya A, Sharma PK. Artificial Intelligence in Pharmacovigilance and COVID-19. Curr Drug Saf. 2023;18:5–14.
- 12.
Medical Dictionary for Regulatory Activities. Introductory Guide MedDRA Version 27.1. 2022 [cited 10 Sept 2024]. Available: https://admin.meddra.org/sites/default/files/guidance/file/intguide_27_1_English.pdf
- 13. Xing X, Zhang X, Wang K, Wang Z, Feng Y, Li X, et al. Post-marketing safety concerns with lecanemab: a pharmacovigilance study based on the FDA Adverse Event Reporting System database. Alzheimers Res Ther. 2025;17:15.
- 14. Moore N, Thiessard F, Begaud B. The history of disproportionality measures (reporting odds ratio, proportional reporting rates) in spontaneous reporting of adverse drug reactions. Pharmacoepidemiol Drug Saf. 2005;14(4):285–6. pmid:15782397
- 15. Rothman KJ, Lanes S, Sacks ST. The reporting odds ratio and its advantages over the proportional reporting ratio. Pharmacoepidemiol Drug Saf. 2004;13(8):519–23. pmid:15317031
- 16. Bate A, Lindquist M, Edwards IR, Olsson S, Orre R, Lansner A, et al. A Bayesian neural network method for adverse drug reaction signal generation. Eur J Clin Pharmacol. 1998;54(4):315–21. pmid:9696956
- 17. Sakaeda T, Tamon A, Kadoyama K, Okuno Y. Data mining of the public version of the FDA Adverse Event Reporting System. Int J Med Sci. 2013;10(7):796–803. pmid:23794943
- 18. Chen S, Fang W, Zhao L, Xu H. Safety assessment of cenobamate: real-world adverse event analysis from the FAERS database. Front Pharmacol. 2024;15:1369384. pmid:38560357
- 19. Tan H, Yan X, Chen Y, Huang G, Luo L, Li W, et al. A real-world pharmacovigilance study of drug-induced QT interval prolongation: analysis of spontaneous reports submitted to FAERS. Front Cardiovasc Med. 2024;11:1363382. pmid:38803662
- 20. Van Holle L, Zeinoun Z, Bauchau V, Verstraeten T. Using time-to-onset for detecting safety signals in spontaneous reports of adverse events following immunization: a proof of concept study. Pharmacoepidemiol Drug Saf. 2012;21(6):603–10. pmid:22383219
- 21. Cai H, Jia B, Fu Z, Chen B, Liu Y, Zhao S. Real-world safety of icosapent ethyl: analysis based on spontaneous reports in FAERS database. Expert Opin Drug Saf. 2024;23(3):373–83. pmid:37873598
- 22. Shimada K, Hasegawa S, Nakao S, Mukai R, Matsumoto K, Tanaka M, et al. Adverse event profiles of ifosfamide-induced encephalopathy analyzed using the Food and Drug Administration Adverse Event Reporting System and the Japanese Adverse Drug Event Report databases. Cancer Chemother Pharmacol. 2019;84(5):1097–105. pmid:31502115
- 23. Fusaroli M, Salvo F, Begaud B, AlShammari TM, Bate A, Battini V, et al. The REporting of A Disproportionality Analysis for DrUg Safety Signal Detection Using Individual Case Safety Reports in PharmacoVigilance (READUS-PV): Explanation and Elaboration. Drug Saf. 2024;47(6):585–99. pmid:38713347
- 24. Saccone G, Della Corte L, D’Alessandro P, Ardino B, Carbone L, Raffone A, et al. Prophylactic use of tranexamic acid after vaginal delivery reduces the risk of primary postpartum hemorrhage. J Matern Fetal Neonatal Med. 2020;33(19):3368–76. pmid:30704334
- 25. Mergoum AM, Mergoum AS, Larson NJ, Dries DJ, Cook A, Blondeau B, et al. Tranexamic Acid Use in the Surgical Arena: A Narrative Review. J Surg Res. 2024;302:208–21. pmid:39106732
- 26. Tian N, Sun Y, Liu Y, Jin J, Chen S, Han H, et al. Safety assessment of tranexamic acid: real-world adverse event analysis from the FAERS database. Front Pharmacol. 2024;15:1388138. pmid:38863974
- 27. Beverly A, Ong G, Kimber C, Sandercock J, Dorée C, Welton NJ, et al. Drugs to reduce bleeding and transfusion in major open vascular or endovascular surgery: a systematic review and network meta-analysis. Cochrane Database Syst Rev. 2023;2(2):CD013649. pmid:36800489
- 28. Murao S, Nakata H, Roberts I, Yamakawa K. Effect of tranexamic acid on thrombotic events and seizures in bleeding patients: a systematic review and meta-analysis. Crit Care. 2021;25(1):380. pmid:34724964
- 29. Sabbag OD, Abdel MP, Amundson AW, Larson DR, Pagnano MW. Tranexamic Acid Was Safe in Arthroplasty Patients With a History of Venous Thromboembolism: A Matched Outcome Study. J Arthroplasty. 2017;32(9S):S246–50. pmid:28262452
- 30. Carbone A, Poeran J, Zubizarreta N, Chan J, Mazumdar M, Parsons BO, et al. Administration of tranexamic acid during total shoulder arthroplasty is not associated with increased risk of complications in patients with a history of thrombotic events. J Shoulder Elbow Surg. 2021;30(1):104–12. pmid:32807373
- 31. Asaadi S, Mukherjee K, Abou-Zamzam AM, Ji L, Luo-Owen X, Tabrizi MB, et al. Tranexamic acid is not associated with a higher rate of thrombotic-related reintervention after major vascular injury repair. J Trauma Acute Care Surg. 2024;96(4):596–602. pmid:38079274
- 32. Lecker I, Wang D-S, Whissell PD, Avramescu S, Mazer CD, Orser BA. Tranexamic acid-associated seizures: Causes and treatment. Ann Neurol. 2016;79(1):18–26. pmid:26580862
- 33. Schlag MG, Hopf R, Zifko U, Redl H. Epileptic seizures following cortical application of fibrin sealants containing tranexamic acid in rats. Acta Neurochir (Wien). 2002;144(1):63–9. pmid:11807648
- 34. Myles PS, Smith JA, Forbes A, Silbert B, Jayarajah M, Painter T, et al. Tranexamic Acid in Patients Undergoing Coronary-Artery Surgery. N Engl J Med. 2017;376(2):136–48. pmid:27774838
- 35. Chobanov JD, Wang Z, Man KKC, Dayib E, Lip GYH, Hingorani AD, et al. Sex-specific comparative outcomes between oral anticoagulants in patients with atrial fibrillation: a systematic review and meta-analysis. Open Heart. 2024;11(2):e002792. pmid:39019498
- 36. Muhunthan K, Balakumar S, Navaratnaraja TS, Premakrishna S, Arulkumaran S. Plasma Concentrations of Tranexamic Acid in Postpartum Women After Oral Administration. Obstet Gynecol. 2020;135(4):945–8. pmid:32168220