The use of tranexamic acid for on‐demand and prophylactic treatment of hereditary angioedema—A systematic review

Hereditary angioedema (HAE), caused by inherited deficiency of C1 esterase inhibitor (C1‐INH), is characterized by recurring subcutaneous and/or submucosal edema. Although its efficacy remains controversial, tranexamic acid (TXA) is used to treat HAE in some countries. We analyzed TXA as an on‐demand and prophylactic treatment in patients with HAE. Published data were systematically sourced from PubMed and Embase. All retained articles underwent grading/bias assessment using the “SIGN” grading system, and the quality of retained studies was determined following assessment of design and methodology. Of 353 studies identified, 31 were included. On‐demand treatment and prophylactic treatment were assessed in five (N = 103) and 28 studies (N = 231), respectively. The majority of studies (80%) demonstrated that on‐demand TXA was ineffective for skin, abdominal, or laryngeal swellings. In a single randomized controlled trial, the median time to relief of symptoms was 2 and 12 hours for icatibant and TXA, respectively (P < 0.001). For prophylaxis, while ~50% of case series, case reports, and observational studies reported beneficial effects of TXA, newer therapies, for example, icatibant and pdC1‐INH, were more effective. One study found that breakthrough attacks during TXA prophylaxis lasted significantly longer compared with C1‐INH (median time to resolution; 7 vs 3 hours, P = 0.016). Many studies failed to report safety data (16/31, 52%); however, pruritus, vomiting, and diarrhea were noted in some patients. There is no evidence for on‐demand use of TXA in HAE and limited evidence for prophylaxis. While TXA may be more beneficial than no treatment, newer, more effective therapies should be used when available.

symptoms are considered to be caused by genetic mutations of factor XII, plasminogen, 4 or angiopoietin-1. 5 As such, diagnosis/treatment remains a challenge. 6 HAE is characterized by recurrent episodes of localized subcutaneous (SC) or submucosal edema lasting for 2-5 days that occurs primarily in the extremities, face, genitals, trunk, and abdomen. 7 Laryngeal edema (approximately 1% of all HAE type I and II attacks) may occur at any time and is potentially lifethreatening. [7][8][9] Numerous treatment options are available for HAE; these include on-demand treatment options for management of acute attacks (eg, C1-INH replacement therapy, 10 kallikrein inhibitors [ecallantide], 11 bradykinin receptor antagonists [icatibant 12 ]) and long-/short-term prophylactic treatment to reduce the frequency and/or severity of HAE attacks, for example, C1-INH replacement therapy, attenuated androgens, 13 and antifibrinolytics 14 (Figure 1). C1-INH has demonstrated considerable efficacy and safety, and while androgens may be effective in some patients, they are associated with side effects, particularly in women and children. 15 Table 1 shows the different treatment options that are used in Japan as recommended by the Japanese Association for Complement Research. Three options are used for on-demand and prophylactic treatment; the antifibrinolytic tranexamic acid (TXA), C1-INH replacement therapy, and the attenuated androgen, danazol. 16,17 Of these, plasma-derived (pd) C1-INH is specifically licensed for the treatment of HAE, 18 while TXA is licensed for urticaria and swelling, not necessarily caused by HAE. 19 Tranexamic acid was first reported by Okamoto and Okamoto in Japan, who highlighted its potent inhibitory effects on fibrinolysis. 20 A synthetic derivative of lysine, TXA works by binding to plasminogen molecules. This in turn inhibits the formation of plasmin, which is normally inhibited by C1-INH. As such, TXA blocks the activation of the complement system and immune cells such as neutrophils, thus attenuating the symptoms of HAE. 21 TXA can be administered orally or intravenously 19 and is utilized in the management of perioperative bleeding, 22 trauma, 23 and also in the treatment of melasma. 24 Although TXA is generally well tolerated with a low adverse event rate, there are concerns regarding lack of efficacy compared with other widely available treatment options, that is, C1-INH replacement therapy and attenuated androgens. 25 Moreover, many patients continue to experience intermittent swellings and require additional use of rescue medication while on TXA prophylaxis. 26 Importantly, clinical data have failed to demonstrate a dose of antifibrinolytics that is effective in a large number of patients, and furthermore, side effects including nausea, fatigue, and diarrhea have been observed in some observational studies. 27 The therapeutic efficacy of TXA is determined on an individual basis, and each patient should be treated with the lowest effective dose.
A retrospective survey in Japan found that TXA was used for prophylaxis in 39.2% of patients. 28 This is a nontypical situation compared with many other regions and is due to limited availability of HAE-specific therapies. Furthermore, self-administration of pdC1-INH is not licensed in Japan, and not all hospitals keep adequate stocks, which can make accessibility difficult for some patients. 29 Importantly, TXA is a more suitable option for patients in whom androgens are contraindicated, for example, women and expectant mothers. 15 In addition, androgens have been associated with numerous anabolic/androgenic side effects; therefore, treatment requires careful monitoring of patients for the detection of adverse events. 15 Worldwide, several other clinically effective options for the treatment of HAE are available. Recombinant C1-INH has proven to be effective for on-demand treatment, reducing the time to relief of symptoms from 152 to 90 minutes, compared with placebo (P = 0.031). 30  significantly reduced symptom severity compared with placebo. 31,32 In addition, the long-term safety and efficacy of pdC1-INH have been evaluated using real-world patient registry data, which showed that pdC1-INH administration reduces the rate of attacks and is associated with an excellent safety profile. 33  Searches were also performed on Embase to identify additional studies not found in PubMed.

| Selection of studies
Titles and abstracts of all retrieved articles were assessed to identify eligible studies. Relevance was defined according to inclusion and exclusion criteria listed in Table 2. Eligible studies were published in English, conducted in human subjects, and contained original clinical data on the efficacy or safety of TXA in type I/type II HAE. Narrative and systematic review articles were not included. Titles and abstracts were initially screened followed by full-text analysis in cases of uncertainty.

| Data extraction and assessment
The following data were extracted: type of study, patient population, efficacy outcomes, for example, prevention of attacks or alleviation of symptoms, and safety outcomes, for example, adverse reactions or side effects. All retained articles underwent grading and bias assessment using the "SIGN" grading system; the quality of studies was determined following assessment of study design and methodology 35 (Table S1). Bias was defined as follows: performance bias, that is, differences in treatment given to comparison groups; selection bias, that is, significant differences in baseline characteristics between groups; detection bias, that is, small sample size; recall bias, that is, data/outcomes based on subjective memory; reporting bias, that is, selective outcome reporting that favors one group over another. Studies were judged to have a low, acceptable, or high risk of bias and a quality grading between 1 and 3 (Table S1).
Using this system, RCTs were given a score of 1 (the highest rating), whereas case reports were given a score of 3 (the lowest rating). All other articles, including observational studies, were given a score of 2. Studies were further up or downgraded based on the robustness of the study design, whereby ++, +, orwas added, to denote low, acceptable, or high risk of bias, respectively. The evidence base for this review was formed mainly of case reports/series (N = 16) 36-51 and observational studies (N = 10). [52][53][54][55][56][57][58][59][60][61] There was one RCT, 62 two crossover studies, 63,64 one questionnaire, 65 and one survey. 66 Three studies evaluated TXA in on-demand treatment, 60

| Bias Assessment
Two studies were assigned a grading of 1, indicating a weak-to-moderate evidence base overall. Several of the included studies were case reports with no valid comparator; these reports were assigned a quality grading of 3 and did not undergo bias assessment (N = 16).
The majority of the observational studies were assigned a grading of 2+, meaning that they were deemed to have an acceptable risk of bias, apart from three studies. 53,60,61 One randomized crossover study was assigned a grading of 1−, 63 while the other nonrandomized crossover study was assigned a grading of 2-. 64  including studies that relied on patient-reported data. Detection bias due to small sample size was identified in 87% of the studies included in the assessment. Industry involvement was identified in 27% (Table S1).

| Efficacy of tranexamic acid for on-demand treatment
Studies assessing the efficacy of TXA for on-demand treatment of HAE attacks are listed in Table 3. The corresponding SIGN level is also shown; there was one RCT, two observational studies, one questionnaire, and one case series. Overall, the evidence base for on-demand treatment was moderate-low; the majority of studies were assigned a quality grading of 2, and all studies demonstrated that TXA was effective in a subpopulation of patients and was less effective than other on-demand treatments.

| Efficacy of tranexamic acid for prophylaxis
Studies assessing the efficacy of prophylactic treatment with TXA for prevention of HAE attacks are shown in Tables 4 and 5 Figure 3 shows the proportion of case reports, case series, and observational studies that reported better/worse outcomes or no change in response to TXA LTP. The most robust data were reported in a randomized crossover study; Blohme et al 63 found that 3/5 patients responded positively to TXA prophylaxis compared with placebo.

| Long-term prophylaxis
One patient experienced abdominal pains and skin edema for the duration of treatment, and TXA had to be withdrawn due to severe vomiting. Another nonrandomized crossover study showed that TXA led to a complete or almost complete cessation of attacks in 7/18 patients (39%), 64 and TXA modestly reduced the frequency of attacks in four patients; however, these attacks were of markedly reduced severity. showed that TXA prophylaxis resulted in increased attack frequency, 50 adverse effects leading to discontinuation, 48 and recurrent laryngeal edema. 42

| Short-term prophylaxis (STP)
Four studies looked at the use of TXA in STP (
Reported side effects in response to on-demand TXA treatment included fatigue and dizziness in a limited number of patients. Most of the trials that reported safety data in response to TXA LTP did not identify any adverse events; however, abdominal discomfort, vomiting, pruritus, vertigo, and diarrhea were reported in some patients.  This review demonstrates that there is very limited data on the efficacy of TXA as an on-demand treatment. Only a single RCT was identified, and results from this study showed that TXA took six times longer to induce clinically significant relief of symptoms compared with icatibant. 62 These data were supported by evidence from Similarly for STP, the guidelines do not make a specific recommendation for TXA use, stating that its efficacy in suppressing breakthrough attacks seems to be low. 67 The conflicting evidence generated by observational studies and case reports highlights the difficulty in drawing firm conclusions regarding the efficacy of TXA.

| DISCUSSION
While there were several observational studies and case reports supporting the lack of efficacy of TXA for prophylaxis, we also identified some studies demonstrating that TXA may be effective in some patients. Gomez-Traseira et al 56  In addition, awareness of HAE among physicians in Japan is low and the limited availability of effective therapies increases the burden of HAE. 73 Diagnosis of HAE worldwide can take an average of 8.3 years; this figure is thought to be much longer in Japan. 29,73,74 While danazol is available for off-label use, it has been associated with numerous side effects, including virilization, headaches, depression, and acne. In addition, the risk of adverse events in response to androgens increases the longer the duration of treatment. 15 Our findings suggest the need for increased awareness of available treatment options and an increase in the availability of more effective options to allow Japanese HAE patients to access the best available therapies. Published guidelines are in line with our findings and conclusions. TXA is not recommended for on-demand treatment or prophylaxis. This is particularly relevant in Japan, considering that alongside danazol, TXA is the only other option available for LTP.
Neither treatment can be considered both safe and effective in the majority of patients with HAE. As such, these findings highlight a significant unmet need in Japan with regard to the paucity of clinically effective treatment options, and suggest the need for increased awareness among patients and physicians to ensure more options become available in the near future. The use of TXA is advocated in selected patients that have already been shown to benefit; however, it is likely that once more options become more widely available, the requirement for TXA in both on-demand and prophylactic treatment of HAE may be significantly reduced.
There are several limitations to this review, not least the lack of robust quantitative data generated from well-conducted RCTs.
A large number of studies, particularly in prophylaxis, failed to report numerical data, stating only that TXA was either effective or ineffective at reducing symptoms and/or the frequency of attacks. 37,39,41,49,50,57 Data generated from observational studies and case reports are inherently biased due to the lack of comparators and lack of control over confounding variables; these issues are much less likely to arise in RCTs. Finally, many studies did not report any safety data, which makes it difficult to draw firm conclusions regarding the safety of TXA in the HAE patient population.
The findings presented in this review highlight several conclusions. TXA may be more effective than no treatment, particularly for STP. However, the efficacy of TXA varies widely between patients, and in many cases, the effect of TXA is negligible. For on-demand treatment, there is clear evidence that newer therapies such as icatibant and pdC1-INH are more effective than TXA. Although there have been no direct comparisons between ecallantide and TXA, ecallantide might also be considered superior given its clear advantages when compared with no treatment. For prophylaxis, while some efficacy is observed in selected patients, TXA is inferior to many other options, including C1-INH replacement therapy. As such, there may be limited utility for TXA in STP prior to dental procedures and minor surgery; however, where available, newer, more effective therapies should be used in place of TXA for LTP.

APPROVAL OF THE RESEARCH PROTOCOL
No human participant was involved in this study.