Efficacy and Safety of Long-Term Intravenous Tranexamic Acid Administration for Blood Management in Revision Surgery for Femoral Shaft Nonunion: A Retrospective Case-Control Study

Abstract

Femoral shaft nonunion is a complication that seriously affects physiological functions. We aimed to assess the effectiveness and safety of short- and long-term intravenous tranexamic acid (TXA) administration in the perioperative period of revision surgery for femoral shaft nonunion. In this retrospective study, 53 patients undergoing double-locking plates with channel bone grafting technology for the treatment of femoral shaft nonunion were divided into 3 groups: the patients in group A without use TXA during hospitalization, the patients in group B received intravenous (IV) 1-g TXA at 30 min before the surgery and deep soaked 1-g TXA for 5 min before closing the incision, and then 1-g TXA IV again 6 h after surgery, and the patients in group C received 1-g TXA IV before the operation, 1-g TXA topically during the operation, and subsequent long-term 1-g TXA IV until discharged. The primary outcomes were total blood loss (TBL) and hidden blood loss (HBL). The secondary outcomes included actual hemoglobin (Hb) loss values, transfusion requirement, number of units transfused, postoperative laboratory values (Hb, hematocrit, fibrinogen, and D-dimer), visual analogue scale (VAS) scores, and hospitalization time. The mean TBL was lower in group C than in group A (1168 mL vs. 2714 mL, P < 0.001) and group B (1168 mL vs. 1557 mL, P = 0.008). The differences in HBL volumes were also significant between groups A and C (P < 0.001) and between groups A and B (P < 0.01). The actual Hb loss in the 3 groups showed a consistent trend with TBL, but no significant differences between groups B and C (P = 0.23). On postoperative day (POD) 3, the Hb level was higher in group C than in group A (111.1 g/L vs. 94.6 g/L, P = 0.02). No significant differences were found in VAS, hospital stay, thromboembolic complications, incision-related complications, and TXA adverse reactions among groups. Long-term intravenous TXA during hospitalization can effectively reduce perioperative blood loss, Hb drop, and postoperative hyperfibrinolysis, but is associated with an increased incidence of adverse reactions.

Introduction

Femoral shaft fractures are common in the clinical setting, with an incidence of approximately 6% to 10% and a cure rate of approximately 99% after surgical treatment.^1–3 However, because the medical level and equipment of Chinese primary hospitals are not highly advanced, iatrogenic errors such as incorrect timing of surgery, incorrect fixation principles, and premature postoperative weight-bearing exercise time are considered accountable for bone nonunion.^4,5 Therefore, nonunion is more common in China than reported in the literature.^6,7 Femoral shaft nonunion is a highly morbid complication that aggravates pain, disability, and financial burden, and has a devastating impact on the quality of life.^4,6–8

Owing to its poor prognosis, nonunion may require multiple operations; thus, it has become a treatment dilemma for surgeons and has greatly burdened patients’ physical function and the economy.^9,10 Therefore, finding an effective treatment for femoral shaft nonunion is of great importance. Currently, surgery is still the most important and effective method for treating femoral shaft nonunion, including nail motorization, exchange nailing, retention of additional steel plates for intramedullary nails, and external fixation, along with some auxiliary methods such as ultrasound stimulation, autologous bone grafting, bone morphogenetic protein, and bisphosphonate therapy.^11–20 The double-locking plates with the channel bone grafting technology adopted by our research department have the following unique advantages over the other schemes: (1) no further expansion of the medullary cavity is needed when cleaning the fractured medullary cavity to protect the intramedullary blood supply; (2) the broken end of the fracture nourishes the blood vessels, does not affect the blood circulation of the broken end, and helps in healing; (3) the poor stability of single-plate fixation is avoided, and the broken end becomes absolutely stable; (4) maximize the biological effects of bone and soft tissue and reduces the risk of mobility impairment after surgery. This protocol has now become the first choice of treatment for femoral shaft nonunion in our department.²¹

However, when we collected clinical data on femoral shaft nonunion, we found that perioperative apparent blood loss volume was not consistent with the hemoglobin (Hb) reduction value, which may be related to the use of antifibrinolytic drugs (TXA). Subsequently, we found that the TXA regimen in our department was divided into 2 types, short- and long-term, but no study has reported on the relationship between perioperative blood loss during femoral shaft nonunion revision surgery and TXA. Therefore, the purpose of this study was to assess the benefits and risks of short- and long-term use of TXA in the treatment of femoral shaft nonunion with those using double-locking plates with channel bone graft technology, and to provide a reference for surgeons.

Methods

Patients

This retrospective study included patients with femoral shaft nonunion who were treated using double-locking plates with channel bone graft technology at the Hong Hui Hospital Affiliated to Xi’an Jiaotong University, Xi’an China, from July 2015 to September 2020. This study was approved by the Ethics Committee of our hospital (No: 201606008) and conducted in accordance with the Declaration of Helsinki. All participating patients provided informed consent by signing a written informed consent form.

The patients were sequentially enrolled according to the following inclusion criteria: (i) patients aged >18 years; (ii) patients treated with internal fixation and implantation due to femoral shaft fracture with subsequently developed aseptic nonunion, which was diagnosed by the chief surgeons on the basis of imaging and physical examinations. Nonunion was defined according to the US Food and Drug Administration (FDA) criteria as a fracture that had not healed within 9 months after the last surgery or had not shown any signs of healing for 3 consecutive months.²²

The exclusion criteria were as follows: (i) patients with severe brain, heart, liver, and kidney dysfunction who could not tolerate surgery; (ii) patients with no signs of infection during the preoperative laboratory examinations, with low-toxicity infections at the fracture ends during the operation that required a change in surgical plan; (iii) nonunion of the proximal or distal femur; (iv) allergy to intravenous (IV) TXA; and (v) long-term use of anticoagulants before surgery.

Patients were divided according to the perioperative TXA treatment course. Group A included patients who were not given TXA because TXA was not introduced in our department before 2016. Group B received 1-g TXA IV 30 min before surgery, deep soaked 1-g TXA for 5 min before closing the incision, and 1-g TXA IV again 6 h after surgery. This is the routine TXA treatment plan of orthopedic physicians in our department. Lastly, group C received 1-g TXA IV before the operation, 1-g TXA topically during the operation, and subsequent long-term 1-g TXA IV until discharge. This is the empirical TXA treatment plan of some orthopedic physicians in our department.

Surgical Procedure

Preoperative management

In this study, channel bone grafting technology combined with double-locking plates was used in all patients to treat nonunion under general anesthesia. All 3 groups of patients received IV antibiotics as preventive treatment 30 min before surgery. Groups B and C received 1-g TXA IV, and the other IV administration in the 3 groups was the same. The pressure of the balloon-type tourniquet on the affected limb was set to 450 mmHg, and the mean arterial blood pressure (MAP) was maintained at 60-70 mmHg.

Original internal fixation treatment and incision position design

For patients in whom bone plates or intramedullary nails were previously used for open reduction, an incision was made along the original surgical skin incision to expose and then remove the internal fixation. If the patient had previously fixed with intramedullary nails for closed reduction, a 20-cm longitudinal skin incision was made on the anterolateral or outer side of the affected limb, and then the previous internal fixation device was removed.

Treatment of fractured ends in nonunion and placement of locking plates

The Judet periosteal peeling technique is used to expose the fractured ends of the nonunion. Under the premise of protecting the surrounding soft tissue and periosteum, the scar tissue between the fractured ends is completely removed, and some hardened dead bones are retained. The length recovery of the affected limb is determined by measuring the length of the contralateral side, while anatomical axis restoration is judged by the thickness of the distal and proximal bone cortex of the broken end, the intraoperative C-arm X-ray fluoroscopy, and limb shape. After the normal force line and length of the limbs are restored, the locking plate (Depuy Synthes, USA) is used for lateral cortical bridging and fixed, and at least 6 layers of cortex are fixed with screws.

A specific type of autologous iliac crest block graft and bone slab placement

A 3.5 mm drill is used to make holes at the front, cortical bone side of fracture end to form a slotted marking area with a width of approximately 1.0 to 1.5 cm and a length of about 3 to 10 cm. Slotting is then performed across the fractured end of the nonunion and extends beyond the fractured ends. Having obvious fresh blood seeping from the bone grooves both near and far, which is a positive paprika sign, is advisable. The length of the bone groove is measured, and the corresponding length of the uni- or bi-cortical iliac crest bone is then filled into the bone groove after trimming. Cancellous bone particles are produced from the ipsilateral iliac bone filling the posterior, lateral, and medial gaps of the femur fractured ends.

Once adequate bone grafting is achieved within the medullary cavity, a reconstruction bone plate (Depuy Synthes, USA) is placed at the front side of the bone graft; it is advisable that the screw penetrates at least 4 layers of the cortex.

Stability testing and closure incision

After checking the anatomical axis and length of the affected limb again, stability testing is performed by flexion and extension on the surgical side to check that there is no micro-motion at the fractured end. The wound is thoroughly irrigated with saline, and then 1 g TXA is used in Groups B and C, to soak the wound for 5 min. The incision is repaired anatomically with a continuous suture (2-0 absorbable suture; Ethicon Inc. Johnson & Johnson, USA), closing the wound in layers in a standard manner. One drainage tube is embedded in the incision.

Postoperative treatment

Antibiotics (Ceftazidime, IV 2.0 g Bid) are used as a preventive measure during the operation and on postoperative day (POD) 1. All patients were treated with anticoagulants (Enoxaparin, IH 20 mg qd) and intermittent compression boots as preventive treatment for lower extremity venous thrombosis during hospitalization, and then treated with oral-anticoagulants (Rivaroxaban, OP 10 mg qd) until 35 days after discharge. In patients with a history of DVT, enoxaparin, 40 mg subcutaneously daily during hospitalization. After discharged, oral rivaroxaban 20 mg daily until 35 days. Group B received 1-g TXA IV 6 hours after surgery, while group C received long-term 1-g TXA IV (from the first day after surgery, every day at 8 am until discharge). The drainage tube is preserved for 48 h, and after it is pulled out, active muscle isometric contraction and knee joint function exercise are started. From the third day after the operation, continuous passive motion is used to restore hip and knee joint activities until discharge. During this period, no weight-bearing activities are allowed. According to the X-ray results, once an external callus is formed, the patient could use crutches to walk with partial weight-bearing. Full weight-bearing walking is initiated after complete bone union is achieved.

The red blood cell transfusion indications formulated by the guidelines of the Chinese Ministry of Health were as follows: (i) Hb level < 70 g/L; (ii) 70 g/L < Hb level < 100 g/L, when the patient has symptoms of dizziness, palpitation, asthma, and fatigue. Whether the patient needed blood transfusion was determined by a chief physician who was not involved in the study.

Outcome Measurements

Characteristics

Patient characteristics included age, sex, affected side, height, weight, body mass index (BMI), American Society of Anesthesiologists (ASA) classification, and preoperative blood volume (PBV). The laboratory examination data collected included preoperative Hb, hematocrit, D-dimer, and fibrinogen (FIB) levels. The imaging examination data collected included the nonunion location (upper and middle, middle, or middle and lower), nonunion type (atrophic/hypertrophic), and original internal fixation.

Primary outcome

Perioperative total blood loss (TBL) and hidden blood loss (HBL) were the primary outcomes in the present study. TBL was calculated by applying the Gross formula²³ as follows:

$$\text{Total blood loss = PBV × }\left({\text{Hct}}_{\text{1}}{\text{− Hct}}_{\text{2}}\right){\text{/Hct}}_{\text{ave}}+{\text{Hb}}_{\text{trans}}$$

• PBV = preoperative blood volume

• Hct₁ = the initial hematocrit level at admission

• Hct₂ = the lowest postoperative hematocrit level for patients without blood transfusion or the lowest hematocrit level prior to blood transfusion

• Hct_ave = the average of Hct₁ and Hct₂

• The preoperative blood volume (PBV) was calculated by Nadler’s equation²⁴ as follows:

$${\text{PBV=}K}_{\text{1}}{\text{×}h}^{\mathit{\text{3}}}{\text{+}K}_{\text{2}}{\text{×}w\text{+}K}_{\text{3}}\text{, }\left[\text{h:height}\left(\text{m}\right)\text{,w:weight}\left(\text{kg}\right)\right]\text{;}$$

• For male patients, K₁ = 0.3669, K₂ = 0.03219, and K₃ = 0.6041

• For female patients, K₁ = 0.3561, K₂ = 0.03308, and K₃ = 0.1833

• HBL was calculated as follows:

• HBL (mL) = TBL – IBL – TPD

Intraoperative blood loss (IBL) was estimated from the weight of the surgical sponges and the measurement of the volume of blood collected by the suction canisters. The volume of the irrigation fluids added to the surgical field and the sponge volume were then subtract from this value. The total postoperative drainage (TPD) was the volume of fluid in the drainage bag collected over 48 h.

Secondary outcomes

The secondary outcomes included actual Hb decrease, transfusion requirement, number of units transfused, and postoperative laboratory values (Hb, hematocrit, FIB, and D-dimer).

Postoperative VAS pain score and hospitalization time were also assessed as secondary outcomes. Reviewing the previous literature and our own experience, the patient’s Hb level often drops to the lowest level on the third day after orthopedic surgery.²⁵ Therefore, in the current study, we defaulted to POD3 to calculate the total perioperative blood loss. Complications [e.g., deep vein thrombosis (DVT), pulmonary embolism (PE), wound complications, and TXA adverse reactions] were recorded within 4 weeks post-surgery. IV ultrasound of both lower extremities was performed at admission, POD 1, POD 3, discharge, and after 4 weeks.

Statistical Analysis

Statistical analyses were computed using GraphPad Prism 8.0. Continuous variables were reported as means and standard deviations. One-way analysis of variance (ANOVA) was used to compare the differences among multiple groups. Student’s t-test was used to compare the differences between 2 groups, and the chi-square or Fisher test was used for the analysis of categorical data. P values < 0.05 were considered to indicate a statistically significant difference.

Results

Patient Demographics

A total of 53 patients were treated with double-locking plates with channel bone grafting technology, including 15 in Group A, 15 in Group B, and 23 in Group C. No significant differences in demographic data, imaging, and laboratory data at admission were found among the 3 groups (Table 1).

Table 1.

Patient Demographic Data, Imaging Data and Laboratory Values.

Variable	Group A	Group B	Group C	P value
Patient characteristics
Age (yr)a	47.6 ± 12.7	43.8 ± 15.4	45.8 ± 15.7	0.78
Gender (male/ female)b	6/9	8/7	12/11	0.71
Height (cm)a	157.4 ± 10.1	161.1 ± 5.5	159.5 ± 9.3	0.51
Weight (kg)a	65.3 ± 7.1	60.9 ± 10.4	63.4 ± 5.8	0.30
BMI (kg/m2)a	25.1 ± 5.4	24.3 ± 3.5	25.5 ± 4.7	0.74
PBV (L)a	4.4 ± 0.5	4.7 ± 0.6	4.5 ± 0.6	0.45
Operated side (L/R)b	8/7	5/10	9/14	0.51
ASA classificationb
Ⅰ	9	7	12	0.83
Ⅱ	4	7	8
Ⅲ	2	1	3
Fracture typeb
32-A	3	4	6	0.58
32-B	6	3	10
32-C	6	8	7
Fracture siteb
Upper and middle	2	4	6	0.80
Middle	12	9	15
Middle and lower	1	2	2
Original internal fixationb
Intramedullary nail	6	4	9	0.68
Plate	9	11	14
Nonunion typeb
Hypertrophic	6	5	9	0.92
Atrophic	9	10	14
Preoperative blood testsa
Hb (g/L)	146.9 ± 14.9	139.1 ± 18.0	142.9 ± 16.4	0.44
Hct (%)	43.7 ± 3.5	42.2 ± 4.6	42.9 ± 4.1	0.61
D-dimer (mg/L)	0.7 ± 0.5	0.4 ± 0.2	0.6 ± 0.4	0.11
FIB (g/L)	2.7 ± 0.6	3.2 ± 0.9	3.4 ± 1.3	0.13

Abbreviations: BMI, body mass index; PBV, preoperative blood volume; ASA, American Society of Anesthesiologists; Hb, hemoglobin; Hct, hematocrit; FIB, fibrinogen.

Intergroup comparisons were performed using ANOVA, Chi-square test, or Fisher test (^aANOVA; ^bChi-square test or Fisher).

Primary Outcome

The mean TBL was 2714 ± 999 mL (range: 1236 mL-4363 mL) in group A, 1557 ± 418 mL (range: 835 mL-2235 mL) in group B, and 1168 ± 415 mL (range from 423 mL-2191 mL) in group C, with significant differences among the 3 groups (P < 0.001). The pairwise comparisons and statistical analyses between the groups showed statistically significant differences. Similar results were also obtained for HBL, with HBL values of 1566 ± 864 mL, 724 ± 210 mL, and 548 ± 200 mL in groups A, B, and C, respectively (Table 2 and Figure 1).

Table 2.

Comparison of the Primary and Secondary Results of the 3 Groups.

Variable	Group A	Group B	Group C	P value	Intergroup comparison
					P 1	P 2	P 3
Primary outcomesa
TBL (mL)	2714 ± 999	1557 ± 418	1168 ± 415	<0.001	<0.001	<0.001	0.008
HBL (mL)	1566 ± 864	724 ± 210	548 ± 200	<0.001	0.001	<0.001	<0.01
Secondary outcomes
Actual Hb loss (g/L)a	52.3 ± 20.9	35.3 ± 13.0	29.4 ± 15.5	<0.001	<0.01	<0.001	0.23
Patients requiring transfusionb	12 (80.0)	4 (26.7)	5 (21.7)	0.001	0.003	<0.001	0.73
Units transfused (IU)	3.3	2.5	3.2	NA	-	-	-
Postop. laboratory values
Hb (g/L)a
POD#1	116.7 ± 13.5	113.1 ± 14.5	115.9 ± 22.0	0.84	-	-	-
POD#3	94.6 ± 21.1	103.8 ± 14.5	111.1 ± 18.9	0.03	0.18	0.02	0.21
Hct (%)a
POD#1	34.3 ± 3.6	33.9 ± 4.1	34.2 ± 6.1	0.97
POD#3	28.0 ± 5.9	30.2 ± 4.1	33.0 ± 5.4	0.02	0.25	<0.01	0.10
D-dimer (mg/L)a
POD#1	6.3 ± 2.2	4.6 ± 2.6	4.6 ± 1.9	0.04	0.06	0.02	1.00
POD#3	5.1 ± 1.4	3.8 ± 1.1	3.3 ± 0.9	<0.001	0.01	<0.001	0.13
FIB (g/L)a
POD#1	8.7 ± 2.8	5.4 ± 1.9	5.9 ± 2.5	<0.001	<0.001	<0.001	0.51
POD#3	7.2 ± 1.9	5.1 ± 2.1	5.8 ± 1.3	<0.01	<0.01	0.01	0.21
VAS pain scorea
POD#1	4.3 ± 1.4	3.8 ± 0.9	3.6 ± 1.1	0.19	-	-	-
POD#3	2.5 ± 0.9	2.4 ± 0.9	2.4 ± 0.7	0.92	-	-	-
Hospitalization durationa (d)	6.7 ± 1.4	5.9 ± 1.5	5.6 ± 0.8	0.03	0.14	<0.01	0.43

Abbreviations: TBL, total blood loss; HBL, hidden blood loss; Hb, hemoglobin; Hct, hematocrit; FIB, fibrinogen; POD#1, first postoperative day; POD#3, third postoperative day.

Intergroup comparisons were performed using ANOVA, Chi-square test, or Fisher test (^aANOVA; ^bChi-square test or Fisher).

Figure 1.

Comparison of TBL and HBL among the 3 groups. *** indicated P < 0.001 (Group B or Group C vs Group A); ^## indicated P < 0.01 (Group B vs Group C). Intergroup comparisons were performed using ANOVA.

Secondary Outcomes

Actual Hb loss

The actual Hb loss was 52.3 ± 20.9 g/L in group A, 35.3 ± 13.0 g/L in group B, and 29.4 ± 15.5 g/L in group C (P < 0.001), with a significant difference between Groups A and B (P < 0.01) and between groups A and C (P < 0.001). We found no statistically significant difference between groups B and C (P = 0.23; Table 2).

Patients requiring transfusion and number of units transfused

The number of patients per group who needed blood transfusion was as follows: 12 (80.0%, 3.3 U) in group A, 4 (26.4%, 2.5 U) in group B, and 5 (21.7%, 3.2 U) in group C. Significant differences were observed among the 3 groups of patients requiring transfusion (P = 0.001). This is believed to be due to the small number of people included in groups A and B and the fact that only 9 patients required transfusion in groups B and C (23.7%, n = 38; Table 2).

Postoperative laboratory values

No significant differences in the Hb and hematocrit levels on POD 1 were found among the 3 groups, but the levels on POD 3 were highest in group C (P = 0.03 and P = 0.02, respectively). The D-dimer and FIB levels in group C were lowest on PODs 1 and 3, with significant differences (all P < 0.05; Table 2).

VAS pain score and hospitalization duration

The VAS pain scores on PODs 1 (P = 0.19) and 3 (P = 0.92) were not significantly different among the groups. The lengths of hospital stay were 6.7, 5.9, and 5.6 days in groups A, B, and C, respectively (P = 0.03). Further comparisons between the groups showed significant differences between groups A and C (P < 0.01; Table 2).

Complications

Thromboembolic event morbidity, incision-related complications, and TXA adverse reactions did not differ significantly among the 3 groups. However, the patients in group C had the highest incidence of adverse reactions such as peripheral nervous system and digestive system symptoms, but their symptoms improved after the infusion rate or interval administration were decreased (Table 3).

Table 3.

Other Complications Including Thromboembolic Events, Wound Complications, and TXA Adverse Reactions.

Variable	Group A	Group B	Group C	P value
Thromboembolic eventsa
DVT	0	1	1	0.62
PE	0	0	0	-
Incision-related complicationsa
Ecchymoses	2	1	1	0.59
Hematoma	0	0	0	-
Infection	0	0	0	-
Wound secretion	1	1	0	0.45
TXA adverse reactionsa
Nausea or vomiting	0	0	1	0.51
Dizziness or headache	0	2	5	0.15

Abbreviations: DVT, deep vein thrombosis; PE, pulmonary embolism.

^aIntergroup comparisons performed using Chi-square or Fisher test.

Discussion

With the popularization of TXA in China, an increasing number of orthopedic surgeons have started using TXA with satisfactory results in clinical practice. Therefore, some professors in my country jointly formulated guidelines on TXA application, such as an expert consensus on enhanced recovery after orthopedic surgery in China, guidelines on perioperative blood management, an expert consensus on perioperative antifibrinolytic drugs, and sequential anticoagulant application programs in hip and knee replacement surgery in China. Several studies have confirmed that perioperative anemia increases the risk of postoperative infection and mortality and length of hospital stay, and significantly affects the quality of life.^26–28 Good perioperative blood management can reduce blood loss and transfusion due to surgical trauma, reduce the incidence of anemia, and maintain a high postoperative Hb level. It is also closely related to postoperative physical function rehabilitation.^29–32 Antifibrinolytic therapy, an important aspect of blood management that is considered to be closely related to the concept of enhanced recovery after surgery (ERAS), has emerged as a particular research focus, and TXA makes this strategy a reality.

Currently, TXA is administered in IV, intramuscular, topical, and oral regimens. As the safety of intramuscular and oral regimens has not yet been supported by a large number of clinical trials, these routes of administration have not been adopted by our institution. A pharmacokinetic study of TXA showed that its half-life is approximately 3 h.³³ However, several clinical studies have shown that the hyperfibrinolytic state of the body that is caused by surgical trauma reaches its peak at 6 h after surgery and continues until 18 to 24 h.^34–36 In addition, the application of postoperative anticoagulants (e.g., enoxaparin or rivaroxaban) may cause new bleeding. These facts indicate that a single-dose application of TXA in the perioperative period does not sufficiently inhibit postoperative hyperfibrinolysis; thus, we believe that administration of multiple dosages of drugs is necessary to inhibit fibrinolysis and reduce blood loss in the perioperative period.

The present study is the first to investigate the reduction of blood loss by TXA administration in the surgical treatment of femoral shaft nonunion. The results showed that with the long-term use of TXA after surgery (Group C), the TBL volumes were lower than those in no TXA group (1168 mL vs 2714 mL, P < 0.001) and short-term group (1168 mL vs 1557 mL, P = 0.008). On POD 3, the Hb level in long-term group was higher than those in no TXA group and short-term group, indicating that when we prolonged use of TXA could reduce the blood loss from POD 1 to POD 3. The calculation of HBL also confirmed the advantages of prolonged TXA use. The mean HBL volumes in long-term group, short-term group, and no TXA group were 548 mL, 724 mL, and 1566 mL, respectively. The blood coagulation test results showed that the D-dimer and FIB levels in long-term group were lower than those in no TXA group and short-term group on POD 1 and POD 3. In light of the above-mentioned, we believe that long-term empiric use of TXA after revision surgery for femoral shaft nonunion can help inhibit fibrinolysis, thereby reducing perioperative blood loss. Although routine use (short-term group) also reduced perioperative blood loss, and the mean Hb and hematocrit levels on POD 1 in short-term group were not much different from those in long-term group, surgeons who empirically administered TXA believe that the advantage of this regimen is that the patients do not need frequent IV TXA administration on the operation day after general anesthesia.

Some animal models, in vitro experiments, and clinical trials have proven that TXA can reduce the release of inflammatory factors after surgery, thereby reducing postoperative swelling and pain.^37–41 In addition, it also reduces hospitalization costs and shortens hospital stays. However, in the present study, neither short- nor long-term use of TXA showed the advantage of reducing postoperative pain. In the ERAS strategy, postoperative analgesia is administered by the cooperation of surgeons, anesthesiologists, and nursing staff. We believe that the use of TXA alone cannot achieve the desired effect. Another advantage of this empirical use (group C) is that it can shorten the length of hospital stay as compared with that in patients who did not use TXA (group A) (P < 0.01). Although our institution has strict requirements on the length of hospital stay for patients, the mean hospital stay was only 5.6 days (group C).A similar conclusion was reported by Tzatzairis et al,³⁴ Liang et al,⁴² and Xing et al.⁴³

Whether the use of TXA will increase the incidence of DVT and PE has always been the focus of surgeons; no unified conclusion has been made yet.^44–46 Thus far, most orthopedic clinical trials have been designed to test the hemostatic effect of TXA instead of its safety. For rare complications such as PE, the current clinical trial sample size cannot provide a definitive conclusion. In this study, the coagulation levels on POD 1 and POD 3 were not significantly different between groups C and B. These results confirmed from another aspect that long-term TXA IV administration is safe. Among the patients who received TXA treatment during the perioperative period (n = 38), 1 patient in group C had nausea or vomiting symptoms, and 7 patients had dizziness or headache symptoms (2 patients in group B; 5 patients in group C). Therefore, long-term IV use of TXA has a dual sword; although it reduces blood loss during the perioperative period, it may also increase the patient’s discomfort and anxiety. However, during the course of our research, all the patients still agreed to undergo TXA therapy even after being briefed about the risks (adverse reactions) and benefits (reduced blood loss and blood transfusion) of TXA use.

Despite the encouraging results of the present study, several limitations remain. First, the sample size included in the study was small because most patients did not undergoing laboratory tests (hemoglobin, hematocrit, D-dimer, and FIB) on POD#3, they were excluded from the current study. Regarding the blood transfusion rate and number of units transfused, we believe that a larger sample size will better detect the differences between the groups. Second, we did not perform long-term follow-up after surgery, so the safety of long-term TXA use cannot be accurately concluded. Lastly, we did not check the postoperative inflammatory marker levels of these patients, so the relationship between postoperative inflammation and pain cannot be determined.

Conclusion

This retrospective case-control study proved the advantages of long-term TXA IV administration in the treatment of femoral shaft nonunion involving double-locking plates with channel bone grafting technology, which decreased the blood loss volume and number of transfusion units, maintained the postoperative Hb values. However, attention should be paid to the adverse reactions caused by the TXA IV use.

Abbreviations

TXA

tranexamic acid

IV

intravenous

TBL

total blood loss

HBL

hidden blood loss

RBC

red blood cell

DVT

deep vein thrombosis

ASA

American Society of Anesthesiologists score

Hb

hemoglobin

Hct

hematocrit

FIB

fibrinogen

BMI

body mass index

VAS

visual analogue scale

PE

pulmonary embolism

PBV

preoperative blood volume

TBL

total blood loss

HBL

hidden blood loss.