Evaluation of the Safety and Effectiveness of Topical Intrapleural Application of Tranexamic Acid in Thoracic Surgery: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

Abstract

Objectives

Bleeding remains a common complication post-thoracic surgery. Although intravenous tranexamic acid (TXA) has been shown to decrease blood loss, its use has been associated with adverse effects. Accordingly, topical TXA has been proposed as an alternative to reduce bleeding with fewer systemic complications.

Methods

We searched Medline, Embase, and Cochrane Central databases for randomized controlled trials (RCTs) comparing topical TXA versus control (i.e., placebo) in patients undergoing thoracic procedures. The primary outcome was total postoperative blood loss at 24 hours. Secondary outcomes included were the number of red blood cell (RBC) transfusions, and hospital length of stay (LOS). Meta-analyses were pooled using mean difference with inverse-variance weighting and random-effects.

Results

Out of the 575 unique studies that were screened, we identified three randomized controlled trials (RCTs) involving 399 patients. Out of the three RCTs analyzed, two studies, accounting for 67% of the total, were found to have a low risk of bias. The primary outcome of 24-h post-operative blood loss was significantly lower in patients who received TXA (mean difference [MD] −93.6 ml, 95% CI −121.8 to −65.4 ml, I²= 45%). In addition, the need for RBC transfusion was significantly lower in the topical TXA group compared to control (MD −0.5 units, 95% CI −0.8 to −0.3 units, I²= 60%). However, there was no significant difference in the hospital length of stay (LOS) (MD −0.3 days, 95% CI −0.9 to 0.4 days, I²= 0%). These results remained consistent after several sensitivity analyses. The use of topical intrapleural tranexamic acid has also been found to be safe without any significant safety concerns.

Conclusion

Topical intrapleural TXA reduces blood loss and the need for blood transfusions during thoracic surgery. In addition, there is no evidence of the increased safety concerns associated with its use. Larger trials are necessary to validate these findings and evaluate the safety and efficacy of different dosages.

Introduction

Bleeding remains a common complication post-thoracic surgery. ¹ The risk of postoperative bleeding varies according to thrombocytopenia, acquired platelet dysfunction, clotting factor loss, free heparin, and increased fibrinolysis. ² Bleeding requiring re-exploration is associated with a poor patient outcome, increasing morbidity and mortality. ³ During the perioperative period, a surgical insult to the tissue during the procedure activates fibrinolysis, resulting in bleeding. ⁴ Accordingly, the use of antifibrinolytic therapies has been evaluated in an effort to reduce bleeding complications. ⁵

Tranexamic acid (TXA) is an antifibrinolytic that acts systemically or topically as a synthetic reversible competitive that blocks the lysine-binding sites (LBS) of plasminogen. ⁶ The occupation of the LBS by TXA prevents plasminogen from binding to fibrin, and thus from being cleaved into plasmin, ultimately stabilizing the fibrin matrix. ⁶ The use of antifibrinolytic therapy intravenously (IV) for multiple surgical procedures has been shown to decrease blood loss and the need for blood transfusion^7,8; however, IV use of TXA may be associated with adverse effects such as pulmonary embolism, deep vein thrombosis, and seizures. ⁹

Tranexamic acid (TXA) can be administered topically before surgeries as an alternative to reduce the complications associated with its systemic use. The recommended dosage is 1 to 5 g of TXA diluted in 100 mL of sterile saline solution. This topical application of TXA has been found to be effective in reducing bleeding in various surgical procedures, such as caesarian section, orthopedic surgery, cardiac surgery, and more.^10-12 However, there has been limited evidence within these reviews detailing topical TXA in thoracic surgery. Therefore, we conducted this meta-analysis of RCTs to evaluate the safety and efficacy of topical intrapleural administration of TXA in thoracic surgery.

Methods

This systematic review and meta-analysis was performed using the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines. The study was registered in the prospective international register of systematic reviews (PROSPERO, reference number CRD42022362620). Moreover, the study was approved by Institutional Review Board (IRB) - King Abdullah International Medical Research Center (KAIMRC) (Ref.# NRC23R/751/11).

Search strategy and study selection

We searched Medline, Embase, and the Cochrane Central Register of Controlled Trials. No restrictions regarding the publication status or the language were applied. We also explored the clinical trial registration database (clinicaltrials.gov) for trials relevant to this review at any completion stage. If review articles were identified during this search, their references were reviewed for relevant articles. Search strategies were designed with terms related to tranexamic acid and thoracic surgery. Full details of the search strategy are provided in Table A1 in Appendix.

Studies were deemed eligible if they included adult patients aged 18 years or older who underwent thoracic surgery procedures, such as decortication surgery or pulmonary resections, and received topical application of tranexamic acid. Studies including patients who received IV tranexamic acid were excluded. Studies needed to be RCTs with a comparison group of patients who received either placebo or no intervention (referred to hereafter as the control arm). Observational studies, qualitative studies, case reports, and reviews were excluded. The studies obtained from the database search were imported into Covidence Reference Manager for organization and management ¹³ and independently screened by at least two investigators. Disagreements were resolved by consensus.

Outcomes

The prespecified primary outcome for this systematic review and meta-analysis was 24-hours (24-h) postoperative bleeding. Additional outcomes included the number of red blood cell (RBC) transfusions, and hospital length of stay (LOS). Other outcomes included 12- and 48-h postoperative bleeding. We also qualitatively synthesized safety outcomes such as thromboembolic events, adverse cardiac events, and infections whenever possible. Meta-analyses were conducted if two or more studies reported sufficient data for the same outcome.

Data extraction and risk of bias

For data extraction, we used a predefined data extraction form. First, we extracted information regarding the study methods (including the study design, location, setting, date, and enrollment period). We also pulled participant-level data (including the number of participants, age, gender, comorbidities, and surgery-related information). Finally, raw data for outcomes were extracted. Two investigators performed the data extraction independently, and a third investigator was consulted in case of disagreements.

For risk of bias assessment, the Cochrane Risk of Bias Tool ¹⁴ was used, with disagreements resolved by consulting with an additional investigator. The tool comprises a total of six bias domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessors, selective reporting, and incomplete data reporting. A study was rated as high risk of bias if at least one domain was rated as high risk.

Data synthesis

Since the available outcomes for analysis (i.e., those reported similarly in two studies at least) were reported as continuous variables, we used the mean difference to pool data in the meta-analysis. When standard deviation (SD) data were missing, they were imputed using the method recommended by the Cochrane Handbook. ¹⁵ Studies were weighted using the inverse-variance method with a fixed-effects model first, and then we ran random effect models as a sensitivity analysis. The use of fixed effect models has been recommended when there is a low number of included studies in the meta-analysis (i.e., 2–5 studies) ¹⁶ We performed post-hoc analyses, when possible, by excluding studies of high risk of bias. Heterogeneity between studies was quantified using the I² statistic. Due to the limited number of studies included in this review, subgroup analyses were not feasible. Publication bias was not performed as meta-analyses did not reach the recommended threshold of 10 studies. ¹⁵ A p-value of 0.05 was considered statistically significant. Statistical analyses were performed using R Statistical Software (version 4.2.2).

Results

Study characteristics

Of 884 records, 309 duplicates were removed, and 575 article abstracts were screened. Twelve full-text articles were retrieved, of which three were eligible for final inclusion [Figure 1].^17-19 The characteristics of the included studies are presented in Table 1, providing detailed information about each study. Patients were given 1 to 5 g of TXA that was diluted in 100 milliliters of sterile saline solution. As described in the included studies, the TXA solution was kept in the pleural cavity with chest tubes clamped during the chest closure, which took 15 to 45 min to complete; subsequently, chest tubes were connected to the underwater seal system and suction.

Figure 1.

PRISMA flow chart showing study selection process.

Table 1.

Characteristics of the Included Studies

Study	Risk of bias	Country	N	Age (years)	Gender	Surgery	Dose
Dell-Amore 2012 17	Low	Italy	89	66.7 ± 7.2	60 male (67.4%)	Pulmonary lobectomy, wedge resection, bi-lobectomy	5 g
Sivaraman 2017 18	High	India	240	Most patients 20 to 40 years old	181 male (75.4%)	Pulmonary lobectomy, drainage and evacuation, segmentectomy, decortication, pneumonectomy	1 g
Sabry 2018 19	Low	Egypt	70	43.6 ± 14.2	41 male (58.6%)	Lung decortication	3 g

Two studies were rated as having a low risk of bias^17-19 and one was warranted a high risk of bias as there was no reporting of blinding, allocation concealment, and a risk of attrition bias ¹⁸ (Figure 2). Studies were conducted in Italy, India, and Egypt. Of 399 patients in the included studies, 282 (70.7%) were male. Most surgeries were pulmonary lobectomy and wedge resection. The dose of topical intrapleural TXA was1 g in one study, ¹⁸ while 5 g, ¹⁷ and 3 g in the other two studies. ¹⁹ Outcomes data as reported in each study are presented in Table 2.

Figure 2.

Risk of bias of the included randomized controlled trials.

Table 2.

Outcomes Data as Reported in the Included Studies

Study and outcomes	TXA	Placebo/control	p-value
Dell-Amore 2012 17	N = 44	N = 43
12 h blood loss	298.2 ± 73	458.1 ± 89	0.02
24 h blood loss	589.5 ± 109.6	633.72 ± 184.4	0.06
Transfused RBC (units)	1.5	1.9	0.04
Hospital stay	8.5 ± 2.6	8.7 ± 2.7	>0.05
Sivaraman 2017 18	N = 120	N = 120
Transfused RBC (units)	2	5	<0.01
Re-infection	3%	17%	<0.01
Sabry 2018 19	N = 35	N = 35
Transfused RBC (units)	0.69 ± 0.9	1.54 ± 1.09	0.001
24 h blood loss	267 ± 87	385 ± 87	0.001
48 h blood loss	382.5 ± 120	632.5 ± 164	0.001
Hospital stay	6.11 ± 1.51	6.40 ± 1.59	0.444

Efficacy outcomes: Bleeding, transfusion, lung drains, length of stay

The pre-specified primary outcome of 24-h post-operative blood loss was significantly lower in patients who received TXA (mean difference [MD] −93.6 mL, 95% CI −121.8 to −65.4 mL, I²= 45%; Figure 3). The analysis was repeated using random effect models, and the results remained similar (MD −89.2 mL, 95% CI −128.3 to −50.1 mL, I²= 45%). We also repeated the analysis by excluding the study that required imputation of the SD, which was also the study labeled of high risk of bias, and the results remained consistent.

Figure 3.

Forest plot for meta-analysis of blood loss 24 h after lung surgery showing a significant reduction in patients randomized to TXA.

The need for RBC transfusion was reported in two studies, and there were significantly less transfused units in the topical TXA group compared to the control (MD −0.5 units, 95% CI −0.8 to −0.3 units, I²= 60%; Figure 4). The results remained similar when the random effect model was used (MD −0.6 units, 95% CI −1.0 to −0.2 units, I²= 60%). One study, Sivaraman and colleagues, ¹⁸ reported the outcome of blood transfusion in units but included the use of other blood products such as platelets. Therefore, we repeated the analysis of blood transfusion by including this study as well, and the results remained similar, except for heterogeneity which increased to about 81% (Figure 5).

Figure 4.

Forest plot for meta-analysis of need for red blood cell transfusion showing significant reduction in patients randomized to TXA.

Figure 5.

Forest plot for meta-analysis of need for red blood cell transfusion, including all studies regardless of the type of blood product transfused, showing significant reduction in patients randomized to TXA.

In terms of hospital stay, there was no difference in the length of stay between TXA and the control groups (MD −0.3 days, 95% CI −0.9 to 0.4 days, I²= 0%; Figure 6). The results remained similar using the random effect model.

Figure 6.

Forest plot for meta-analysis of hospital length of stay showing no difference between TXA versus control

Safety outcomes: Thromboembolic events, infections, cardiac events

Dell’Amore and colleagues ¹⁷ reported no operative deaths, toxicity, or thromboembolic events attributed to the topical use of TXA. Bacterial pneumonia occurred in two patients with in the TXA arm, and one patient with control arm. Respiratory failure occurred in one patient in the TXA group only. Prolonged air leaks occurred in four patients with TXA and three patients in the control group. Supraventricular arrhythmias occurred in seven patients with TXA and eight patients in the control group. One patient in the control group had an acute myocardial infarction. None of these safety outcomes reached statistical significance.

Sivaraman and colleagues ¹⁸ reported re-infection in 3% in TXA patients and 17% in control patients; the article reports this contributed to a statistically significant reduction in ICU stay, however, it was unclear whether the re-infection outcome was also statistically significant. This study reported that outcomes such as clot formation and poor lung compliance were less in patients who received TXA, however, the event rates for these outcomes were not reported. There were no transfusion-related complications reported in either group in this study. Sabry and colleagues ¹⁹ did not report on safety outcomes (e.g., thromboembolic events, infections, cardiac events) reported.

Discussion

Our study provides a comprehensive overview of the available evidence of topical intrapleural TXA in lung and pleural surgeries. We found a significant reduction in our primary outcome of postoperative blood loss in patients who were randomized to topical TXA. Secondary outcomes such as the need for RBC transfusion were lower in the TXA group, however, there was no difference in the hospital stay between groups. A qualitative review did not identify elevated risk of safety outcomes such as thromboembolic events, however, this was limited by poor reporting in included studies. These results suggest that intrapleural topical TXA may be considered in patients undergoing lung and pleural surgeries to reduce complications related to bleeding.

Local administration of TXA has been studied in thoracic and nonthoracic surgery with promising results in reducing the side effects of the thrombotic tendency caused by systemic infusion.^1,2,9 Thoracic surgeries, such as decortication, are more likely to be affected by fibrinolysis phenomena as a result of significant trauma to the pleura and usually result in a reasonable amount of blood loss. ²⁰ Fibrinolysis due to surgical trauma is one of the most common causes of postoperative bleeding, with the risk of chest re-exploration due to bleeding ranging between 1 to 3.7%.^3,4 Anatomical studies have suggested that the pleural space functions as an important barrier for absorption. ²⁰ This may permit hemostasis without thrombotic complications due to limiting systemic absorption.

In the current meta-analysis, all three studies included thoracic surgery cases, primarily lung resections, and decortications. The TXA dosage varied between trials, with 1 g, 3 g, and 5 g being used. Unfortunately, literature supporting the use of this route of administration is scarce. Studies evaluating the topical application of TXA to other locations such as the pericardial cavity have shown similar results to our study. For instance, Khalil and colleagues performed a randomized trial evaluating patients receiving high dose systemic TXA with and without topical TXA. This study showed that the additional topical administration reduced post-operative blood loss by 33%. ²¹ This highlights that the topical application of TXA is promising and may promote hemostasis via mechanisms other than those involved in systemic absorption. Future randomized trials with arms including various TXA dosages may inform clinicians of an optimal dosage.

This study had several strengths. First, this systematic review was conducted according to a predefined protocol, which was registered on PROSPERO. Accordingly, our significant primary outcome was selected a priori. Second, our literature search was comprehensive, including three databases. Study screening was performed independently by several authors. Third, our research question was relevant given the recent publications of randomized trials evaluating topical intrapleural TXA. A synthesis of the evidence was warranted to provide clinicians a transparent overview of the evidence.

Nevertheless, this study had several investigations. First, only three RCTs were included, each of which had few patients included. This underpowered our study, even in the case of a meta-analysis combining all available data. Second, we encountered statistical heterogeneity, particularly for the outcome of need for RBC transfusion. Reasons for this heterogeneity may be explained by differences in a study's conduction from a methodological or clinical perspective. Methodologically, Sivaraman and colleagues demonstrated a high risk of bias. ¹⁸ Clinically, some patients in this study had different indications for lung surgery, such as fungal aspergilloma, lung abscess, and empyema, which may also explain statistical heterogeneity. In addition, the type of blood transfused was defined differently in the study by Sivaraman and colleagues compared to the other studies. In fact, our post-hoc analysis excluding this study significantly reduced the heterogeneity from 81% to 60% for the outcome of blood loss. Third, given few studies, we were unable to perform subgroup analyses that may explain differences in responses to TXA, such as specific lung surgeries, patient comorbidities, or TXA dosage. Sabry and colleagues ¹⁹ suggested decortication surgeries may theoretically benefit most given the surgery's traumatic effect on the serous membrane which can promote the fibrinolysis cascade. Fourth, one included study excluded patients at higher risk of bleeding, who may arguably benefit most from TXA. ¹⁹ Finally, we were unable to meta-analyze thromboembolic outcomes, which have been a risk of systemic administration of TXA. ⁹

Conclusion

Topical intrapleural TXA reduced blood loss and the need for RBC transfusions during thoracic surgery specifically, lung resection and pleural decortication. Topical intrapleural TXA has not revealed any heightened safety concerns, including thromboembolic events. Larger trials are needed to validate these findings, as the current effect sizes are too small to evaluate secondary outcomes using the current evidence.