A double-blind pilot randomized controlled trial of topical tranexamic acid after sinus surgery*

Background : There has been limited study of topical tranexamic acid (TXA) in endoscopic sinus surgery (ESS). We report a pilot study, examining the effects of topical TXA with regards to recovery after ESS. Methods : A pilot double-blind randomized controlled trial was conducted in 30 patients undergoing comprehensive ESS. Patients received either topical TXA or normal saline (NS) for 60 minutes via cotton pledgets at the conclusion of ESS. Patients were followed-up for a duration of 3-months. Results : The mean (95% CI) reduction in 22-item Sino-Nasal Outcome Test (SNOT-22) score at 3 months was 39.5/110 (26.9, 52.0) for TXA and 33.4/110 (24.0, 42.9) for NS (p=0.42). The mean (SD) Modified Lund-Mackay Post-operative Endoscopic (MLMES) score at 3 months was 7.79/100 ±7.70 for TXA and 10.9/100 ±9.35 for NS (p=0.12). TXA had a mean (SD) bleeding score of 4.0 ± 2.33 on day 1 compared to 3.64 ± 2.76 in NS group when measured on a Likert scale (p = 0.89). The mean self-reported time to return to work was 4.67 ± 2.22 days for TXA and 6.87 ± 4.42 for NS (p=0.10). Zero cases of confirmed thromboembolism were seen. Conclusions : Although statistically non-significant differences were observed, data from this pilot study imply that there is merit in a larger study to further assess the effects of topical TXA following ESS. There may be a role for increasing the exposure to topical TXA via a different formulation.


Introduction
The ideal nasal dressing following endoscopic sinus surgery (ESS) is one that is haemostatic but also improves mucosal healing (1) . Tranexamic acid (TXA) is a readily available and inexpensive medication that features on the World Health Organization list of essential medications (2) . Randomised controlled trials have shown that topical TXA exhibits superior haemostatic effects and higher surgeon satisfaction when used intra-operatively during ESS (3)(4)(5)(6) . However, the use of topical TXA postoperatively following ESS has not been studied with adequate follow-up time. In vitro studies of respiratory epithelium have shown wound healing and anti-inflammatory effects of TXA (7)(8)(9)(10)(11) that could theoretically accelerate recovery and improve patient outcomes (12) . In this sense, TXA has the potential to contribute to an ideal wound dressing following ESS.
Currently, the practice of the senior author (AJW) is to request the administration of intravenous (IV) TXA at anesthetic induction (13)(14)(15)(16) prior to ESS and insert cotton pledgets containing TXA topically in the nose and sinuses at the conclusion of the surgery. Surgeons may be cautious to administer two forms of TXA due to concerns of increased risk of thromboembolism, even though, this practice is shown to be safe in multiple systematic reviews and meta-analysis of knee and total hip arthroplasties (17)(18)(19)(20) and recently in ESS (15) . A retrospective study of 177 patients who had undergone comprehensive ESS from our own practice found no cases of thromboembolism attributable to TXA where both forms of TXA were used (12) .
Given the evidence from other surgical specialties and our own retrospective study, we conducted a pilot randomised controlled trial assessing the impact of topical TXA on long-term patient outcomes using both subjective and objective outcomes measures to assess healing, recovery, and haemostasis. The primary objective of this pilot study was to facilitate a subsequent, larger study with appropriate power calculations. Ultimately, this has the potential to improve long-term post-operative outcomes for patients undergoing ESS.
Study hypotheses: The primary hypothesis was that the postoperative use of topical TXA after ESS for the treatment of CRS would result in improved long-term outcomes with regards to patient-reported symptoms.
We also hypothesised that the post-operative use of topical TXA would result in objective improvements in CRS treatment and faster recovery of normal function with improved healthcare utility. Patients were enrolled in the study following informed consent.

Randomisation
Once anaesthetised, enrolled patients were randomised according to an envelope system. Kraft envelopes were prepared prior to the beginning of the trial using a randomly generated sequence in eight blocks of four and one block of two. The randomisation schedule is available at http://www.randomization. com (seed: 12141).
Research subjects were randomised to receive either topical TXA (5ml of 100mg/ml) or 5ml topical 0.9% sodium chloride (normal saline/NS) as placebo. Cotton pledgets soaked in the study medication were placed in the sinus cavities for 60 minutes at the conclusion of ESS. During surgery, routine measures were utilised for bleeding control including hypotensive anaesthesia, intravenous TXA, head-up positioning, injected 0.5% Marcaine with adrenaline for sphenopalatine blockade and topical Moffett's solution (25) . The severity of intra-operative bleeding was graded using a system established by Boezaart et al. (26) .

Surgical treatment and post-operative care
With the exception of the randomisation to a treatment arm, all Please refer to Figure 1 for a summary of the study protocol.

Statistical analysis
Descriptive statistics were used to report mean, standard deviation (SD) in normally distributed and median (range) in non-normally distributed data. For completion, differences in baseline characteristics were tested. For continuous variables that were normally distributed, Independent Samples T-test was used. Otherwise, Mann-Whitney U-test was utilised. For categorical variables, Chi-square test for Independence or Fisher's exact testing was used.
To compare the effectiveness of the interventions assessed Differences between the means were reported as mean and 95% confidence interval (95% CI). Unless otherwise stated, p<0.05 was considered statistically significant. Where appropriate, Bonferroni correction was applied. In all cases, two-tailed p-value was reported. All statistical analyses were carried out using IBM SPSS (v. 27).

Results
92 patients were assessed for eligibility. 62 patients were excluded and 30 were recruited and randomised to the experimental arms. One patient did not complete full 3-months follow-up and two patients did not have MLMES scores due to COVID-19 pandemic restrictions meaning that assessment occurred remotely.
Please refer to Figure 2 for CONSORT (29) patient flow chart.
Patient characteristics for both experimental groups were similar at baseline with no statistically significant differences. The mean age (SD) for TXA group was 50 ± 13.8 years and for NS group, 51 ±16.5 years (p=0.88). In both groups, most patients were male and of European ethnicity. Table 1 outlines a full comparison. operative visits, nasal endoscopy was performed with debridement of the sinuses as required.

Study follow-up/outcome measures
The patients were followed up 1-day, 2-weeks and at 3-months following surgery.
The primary outcome measure was to compare the absolute reduction in the SNOT-22 score (calculated from the pre-operative and 3-month scores) in the TXA and placebo groups. Secondary outcome measures included the Modified Lund-Mackay Endoscopic Score (MLMES) at 2 weeks and 3 months follow-up, which measure the cumulative inflammatory burden post-ESS to give a total score out of 100 (27) .
Patient recovery time reported as days before returning to work/ daytime roles. The severity of patients' post-operative bleeding was measured using a Likert scale scored from 0-10 and Clinical graduation of post-operative bleeding as established by Kastl et al. (28) . This was then used as a measure of short-term patientreported outcomes.
The following adverse events were measured: re-presentations to the Hospital during follow-up period as a measure of healthcare utility, undesirable effects such as ophthalmic, cardiac, neurological symptoms or nausea during study follow-up, postoperative imaging performed due to suspicion of thromboembolisms and confirmed cases of thromboembolisms such as myocardial infarction (MI), deep vein thrombosis (DVT), pulmo- For the TXA group, the mean ± SD of pre-operative score was 54.4 ± 17.8 points, compared to 3-months score of 14.9 ± 15.4.
For the NS group, the mean pre-operative score was 51.8 ±

Severity of post-operative bleeding
The severity of bleeding was measured on two scales, a Likert scale (0-10) and a clinical graduation scale (0-4). One patient did not complete the 3-month follow-up visit. Therefore, only 29

Adverse events
During Only four patients in the study population had an adverse post-operative bleeding event, two in each experimental group.
No patient required surgical management for post-operative bleeding. Fisher's exact testing found no significant differences between the two groups in incidence of re-presentation to hospital (p=0.33), imaging within follow-up period (p=0.48), or adverse symptoms (p=0.36).

Discussion
In this study, we hypothesised that the post-operative use of topical TXA in ESS would result in improved long-term outcomes with regards to patient-reported and objective measures.
Intended as a pilot study, this study has, unsurprisingly yielded non-significant comparisons with regards to the major outcomes measured. Based on these data, a change in practice is not recommended. It is noted however that the SNOT-22 scores and endoscopic scores at 3 months as well as the data around return to work all show a trend towards favouring TXA, implying that our hypothesis has potential merit and further study is warranted.
There are some theories as to why topical TXA could generate superior outcomes. In a previous scoping review conducted by the authors (12) TXA demonstrated a faster and higher magnitude of wound healing compared to normal saline or no treatment in an in-vitro setting in both respiratory and other types of epithelia (7)(8)(9)(10)(11)30,31) . In addition, TXA demonstrated anti-fibrinolytic and anti-inflammatory effects which in combination with wound healing could afford a greater improvement in SNOT-22 score, improved objective endoscopic score and faster return to normal functioning.
It is worth noting that, although statistically non-significant, patients who received TXA returned to daytime functioning on average 2.2 days earlier than those who received NS. CRS is associated with an estimated 18.7 missed workdays a year (32) .
Although there is no recognised minimally clinically important difference (MCID) for return to work, we would suggest that any earlier return to work, especially if confirmed in future studies would be expected to be clinically important. It is noted that the MCID threshold for improvement after sinus surgery in SNOT-22 score has been reported to be 9.0 (33). In that context, an absolute improvement of 6.04 points between TXA and NS groups in this study could be seen to be relatively modest. However, the promising results from this study encourages us to also consider studying increased use of topical TXA from the described 60 minutes of exposure, perhaps in the form of a longer-term dressing or by regularly adding TXA to post-operative lavage solutions.
We hypothesised improved healthcare utility with the use of topical TXA. In total, only four (23%) patients re-presented to Hospital during the study period. Two of these patients re-presented due to chest symptoms and both had received topical TXA.
One of these patients developed cardiac symptoms due to new onset atrial fibrillation. For the other, the chest symptoms were associated with recurrent chest infections. Therefore, the cause of the re-presentation cannot be fully attributed to the use of TXA. Regardless, observation for complications and side-effects associated with TXA would need to be a part of any future study.
What is reassuring in this study is the zero cases of major complications in the form of thromboembolic disease.
Surprisingly, we found no significant differences between TXA and NS in patient-reported severity of post-operative bleeding.
This contrasts with most current literature in both cardiac, and to a lesser extent, ENT surgery that found superior haemostatic properties and reduced post-operative bleeding with the use of topical TXA (34)(35)(36)(37)(38) . There are a few reasons why our results may vary from published literature. The biggest limitation in our analysis was the small sample size which leads to a large probability of missing an actual effect. Although multiple factors are relevant including delivery method and renal function, the plasma half-life of TXA is generally measured around 2-3 hours (39,40) . It is possible therefore that topical application of TXA for 60 minutes only aids in the stabilization of the fibrin clot and reduction in bleeding in the initial hours, which our methodology may not have captured.

Ethics approval and consent to participate
Not applicable.

Consent for publication
Not applicable.

Availability of data and material
Not applicable.
The biggest limitation in this pilot study is the small sample size.
With only 30 patients, there is a high probability of committing a Type II error and thus concluding no difference when one such exists in the population (41) . Post-hoc power calculation showed that 30 patients only accounted for 13.1% of the study power when detecting a difference of six points in the SNOT-22 score between interventions. Thus, we are at a risk of committing a type II error 87.5% of the time when the minimum acceptable probability is 20% (41) . Therefore, to make generally applicable treatment recommendations, it would be prudent to conduct a trial that is adequately powered to find a difference. Our power calculations show that to detect a statistically significant difference of 6-points in SNOT-22 scores between interventions, with a population standard deviation of 22.7 points, as it was observed in our study, we would need approximately 222 patients in each intervention group. This would achieve alpha-level of 0.05 and 80% power (42,43) . If however we were to pursue patient recovery as the primary outcome, to detect a statistically significant difference of 2.2 days with a population standard deviation of 2.22 days, we would only require a sample size of 32 patients.

Conclusions
We performed a pilot randomised controlled trial comparing the use of topical TXA to NS in improving long-term patient outcomes after ESS. Although not statistically significant, our results encourage us to pursue hypotheses around improved subjective and objective outcomes associated with topical TXA. If proven in a larger study, being inexpensive, safe, and readily available, topical TXA has the potential to make an excellent addition to medical management after ESS.