Study Design Summary: Double-blind, placebo controlled, multicenter randomized control trial of 496 patients with spontaneous epistaxis that was persistent despite external pressure and topical vasoconstrictors, then randomized to either topical TXA or placebo. The primary outcome was the need for anterior nasal packing.
Notes: The study included patients who presented with spontaneous, non traumatic epistaxis. Patients were excluded if they had traumatic epistaxis, hemodynamic instability, with EMS nasal packing, sent in by ENT clinic, nasopharyngeal cancer, pregnancy hemophilia. They initially received external pressure and ice for at least 10 minutes, if bleeding persisted topical; intranasal vasoconstrictors were used. Choice of agent was left to treating physician. If after 10 minutes there was persistent epistaxis the patients were randomized into the trial and blinded to patient and physician. The decision to move onto further treatment after the TXA or placebo was at the choice of the physician. Primary outcome was need for anterior nasal packing.
The authors calculated that to detect a 10% difference in packing rate they would need 414 patients total, they then aimed for 450 to account for loss to follow-up. There were 2622 patients initially deemed eligible, 2324 patients were excluded with really poor reasoning. There were 834 excluded, due to no research nurse available, 772 due to “other,” 3325 due to no senior clinical staff and 161 due to “department too busy.” These are way too high to be comfortable with and 772 due to “other” hides a lot of sin. They ended up with 496 patients. First10EM notes that the study is technically underpowered because the power calculations were based on assumption that 95% of patients would require anterior nasal packing but it ended up only being 45%. Method section notes that patients were to be randomized 1:1 but the TXA group ended up with 254 as compared to 242 of placebo. Analyzed in intentional to treat analysis. There was 100% followup for patients that were randomized into the trial.
There were significantly more men in the placebo group, and more patient son anticoagulation in the placebo group. No p values or statical analysis numbers are provided with the demographics table. They did not include a table of the adverse events in the main paper, the authors did state that all serious adverse events were not felt to be secondary to TXA.
The authors address that the TXA dosing in this study is less than in previous studies with the idea that dosing larger than 200 mg are not likely to be fully absorbed into the cotton roll and thus not likely to be fully administered to the patient.
Anterior packing was placed in 100 patients (41%) in the placebo group as compared to 111 (44%) in the TXA group. These differences were not statically significant but it should be noted that the TXA group had more patients receive packing. The hospital admission rate was 110 (45%) in the placebo group as compared to 110 (43%) in the TXA group which seems extremely high-however per First10EM it seems this is standard practice in UK which is where this study took place. There were no difference between the groups in regard to secondary outcomes including hospital admission, blood transfusion requirements, recurrent epistaxis, thrombotic events requiring hospital hospital reattendance within 1 week.
In the discussion section they mention that other previous studies have shown potential benefit to topical intranasal TXA, though they were each looking at different primary endpoints. According to First10EM there are 6 previous RCTs which total to 692 patients, each study had fairly different methods, some using oral and others using topical TXA, which treatment in the control groups also varying. Cochran review rates these studies’ evidence as “moderate to low.” These studies seem to show TXA decreases rebreeding rates.
Given the very high number of patients excluded from the study with poor reasoning, no adverse events table listed in main study, in combination with previous studies which have shown potential benefit, I am very skeptical of the results of this study and do not feel this study can be used to state whether topical TXA should or should not be used for epistaxis. While needing anterior nasal packing is a significant endpoint it may not be the best endpoint to determine TXA effectiveness, this will need to be explored further. St. Emylyns blog notes that this study and the unexpectedly low need of anterior packing the control arm shows that simple and conservative measures work better than we may expect.
Study Conclusion: There was no statistically significant difference in the need for nasal packing in the TXA arm vs the placebo arm.
Fusion Beat Bottom Line Impression: While this study’s results alone show no significant difference in the rates of needing anterior nasal packing after topical TXA vs placebo, given the limitations of this study I feel it does not provide definitive evidence one way or another about topical TXA in epistaxis and I do not think it should alter practice until more data comes out.
A Fusion Beat 