Intravenous single dose of tranexamic acid safely reduces blood loss and the need for transfusion in elderly patients with hip fracture. A randomized double-blinded controlled trial at 1-year follow-up

,


Intravenous single dose of tranexamic acid safely reduces blood loss and the need for transfusion in elderly patients with hip fracture. A randomized double-blinded controlled trial at 1-year follow-up Introduction
Perioperative anaemia requiring blood transfusion in elderly patients with hip fractures is a severe health problem.
Recently, several randomized studies [1,2] have reported the effi cacy and safety of Tranexamic Acid (TXA) in reducing blood loss in patients with hip fractures who underwent surgery, with a follow-up between four postoperative days and three months [1][2][3][4] and only one randomised study had 6-month followup [5].The risk of Venous Thromboembolism (VTE) after hip fracture surgery is usually described until 3 postoperative months, with the highest risk during the immediate postoperative period (one to six weeks) [6].Nevertheless, some authors have demonstrated that hip fracture patients can have a risk of VTE up to one postoperative year [7,8].Thus, the safety of routine use of TXA in hip fracture patients in the longer-term remains inconclusive.
The purposes of this study were to evaluate the effectiveness of a single dose of intravenous TXA in reducing perioperative blood loss and requirement for transfusion in elderly patients undergoing hip fracture surgery, and to assess the safety of TXA within one postoperative year.The hypothesis was that administering a single dose of intravenous TXA would decrease the perioperative bleeding and reduce the need for transfusion without increased thrombotic risk and mortality within one postoperative year.

Methods
This single-centre, randomised, placebo controlled, double blinded trial was approved by the institutional review board (PI2018-141) and included in a public registry (ClinicalTrials.gov NCT03211286).Informed consent was obtained prior to randomisation.This research was performed under the Declaration of Helsinki International Ethical Guidelines, and the protocol was conducted and reported according to the Consolidated Standards of Reporting Trials (CONSORT).
Consecutive patients with hip fractures admitted to our institution from January 2019 to September 2020 were eligible for the study.The inclusion criteria were age over 65, a hip fracture that occurred within 24 hours prior to admission, and surgical treatment within 48 hours after admission.The exclusion criteria were: 1) Groups IV-V of the American Society of Anesthesiologists (ASA) scale 2) tumoral pathologic fracture; 3) concomitant fracture; 4) refusal to receive blood products; 5) antiaggregant platelet treatment in the three days before surgery; and 6) described contraindications for TXA [8].During the pandemic of the SARS-CoV-2 virus, positive patients were excluded because venous thrombosis was a potential complication in those patients [9].

Randomization
Randomization was based on a computer-generated number list using the block method by an independent assistant.Each assignment was sealed in a consecutively numbered opaque envelope opened by the nurse who prepared the intravenous solutions in the operating room.Patients were allocated into one of two groups: 1) TXA group, whose patients received 1 gram of intravenous TXA (Amchafi brin, Rottapharm Madaus, Germany) diluted in 100 ml of saline solution; 2) Placebo group, whose patients received an equivalent volume of intravenous saline solution.Masking was ensured by preparing the same volume of solution with an identical appearance.Just before the surgical incision, both treatments were performed.The surgeons, anesthetists, and patients were blinded to the assignment until the completion of the study.The dose of intravenous TXA chosen for this study was based on previous reports [10,11].

Surgical procedure
All surgical procedures were performed under spinal anesthesia.Extracapsular fractures were treated with a trochanteric femoral nail or dynamic hip screw, and intracapsular fractures with a bipolar hemiarthroplasty (usually in patients of age over 75 years) or total hip arthroplasty patients (patients under that age), acting under widely admitted indications.Diathermy was routinely used.At the end of the operation, a deep vacuum drain was placed for 24 hours.
With the proposal of minimising the effect of isovolumetric hemodilution, postoperative fl uid therapy was standardized for the fi rst 24 hours with 1500 ml of saline solution, unless the patient needed other treatment.In a standardized manner, all patients received antibiotic prophylaxis with fi rst-generation cephalosporins for 24 hours, and thromboembolic prophylaxis with low-molecular-weight heparin for 30 days, according to the protocols of our center.All patients were postoperatively mobilized under the assistance of a physiotherapist on the fi rst postoperative day.

Evaluation and outcome measures
A standardized protocol for co-management between orthopaedic surgeons and geriatricians was used from admission to discharge.Comorbidity patient was categorized using the ASA scale [12] and Charlson index [13].Patient evaluation was made preoperatively and at one, three, six, and 12 postoperative months, unless death had occurred before.If the patient does not return for examination, telephone contact with patients or their relatives is performed.Two independent surgeons who were blinded to the study groups evaluated all outcomes.
The primary effectiveness outcomes were the total blood loss and transfusion rate.Patients were monitored with serial measurements of Haemoglobin (Hb) during the stay.Patients received a transfusion of packed red cells if their Hb fell below 8 g/dl, or less than 9 g/dl if they had symptomatic anemia or heart disease.The total blood loss was calculated according to widely accepted mathematical formulas based on the Hb levels and the estimated blood volume [14,15].For this purpose, fi nal Hb was the lowest level within four postoperative days [15].For men = (0,3669 x H 3 ) + (0,03219 x W) + 0,6041 For women = (0,3561 x H 3 ) + (0,03308 x W) + 0,1833

[H, height (meters). W, weight (kilograms)]
To calculate the hemoglobin lost with bleeding (Hb loss ), the following formula was applied: Hb loss = BV x (Hb i -Hb e ) x 0.001 + Hb t [BV, the estimated volume of body blood (lt).Hb i , the hemoglobin of the patient before surgery (g/dl).Hb e , the lowest point of hemoglobin during admission (g/dl).Hb t , is the total amount of allogeneic Hb transfused.A unit of banked blood was considered to contain 52 g Hb, according to measurements at our hospital's blood centre].
And fi nally, to calculate the estimated blood loss (EBL), the following formula was used: EBL (ml) = 1000 x Hb loss /Hb i The primary safety outcome was the rate of thromboembolic events.Patients were clinically monitored for these events until one postoperative year.If there was suspicion of any event, the diagnosis was confi rmed by doppler ultrasonography for Deep Venous Thrombosis (DVT), computed tomography scan for pulmonary thromboembolism, magnetic resonance imaging for cerebral stroke, and Electrocardiogram (ECG) and troponin level for myocardial infarction.Data on surgical or medical infection, readmission and death were also collected.

Statistical analysis
According to a previous study on TXA for hip fracture surgery, a reduction in blood loss of 500 ml was considered clinically relevant [1].For a power of 80% and a two-sided type I error of 5%, 60 patients per group were needed.Assuming a drop-out rate of 5%, at least 63 patients per group were required.
Statistical analysis was performed using SPSS software v. 19 (SPSS Inc, Chicago, USA).The Kolmogorov-Smirnov test was used to examine the normal distribution of continuous data.Analyses between groups were performed with the chisquare, Fisher's exact, or nonparametric Mantel-Haenszel test for categorical variables, and the t-Student test or nonparametric Mann Whitney U-test for continuous variables.The paired t-Student test or Wilcoxon signed-rank test was used to compare preoperative and postoperative data.Multivariate logistic regression analyses were planned to identify independent risk factors for main outcome variables, including in the model only the variables with a univariate p -value < 0.1.Data were shown as Odds Ratio (OR) with 95% Confi dence Interval (CI).The goodness of fi t of the logistic regression model was analysed using the Hosmer-Lemeshow test.Statistical signifi cance was considered for p values less than 0.05 in all tests.

Results
From January to December 2020, 207 patients with a hip fracture undergoing surgery were assessed for eligibility.
Seventy-seven patients were excluded for various reasons, including failure to meet inclusion criteria (7 0 patients), decline to participate (fi ve), or language barrier (two).The remaining 130 patients were randomised into the TXA group (65 patients) or control group (65 patients).One patient in the control group was then excluded due to follow-up loss (Figure 1).Thus, 129 patients were analyzed, 76% were females with a mean age of 81.8 years (range 69 to 100).The baseline data of both groups are shown in Table 1.There were no signifi cant differences between groups, specifi cally in mean Hb level on admission (p = 0.226) or patient blood volume (p = 0.845).In the analysis by subgroups according to the fracture type, there were no signifi cant differences in the baseline characteristics, except that intracapsular fractures presented longer surgery time (p = 0.019).Postoperative follow-up was one year in all patients.

Blood loss and transfusion
The mean total blood loss during the entire admission (Table 2) was signifi cantly lower in the TXA group (962.3 ml, SD 511.7) compared with the placebo group (1328.0ml, SD 752.4) (p = 0.006).There were fi ve (7.7%) patients requiring transfusion in the TXA group and 22 (34.4%) in the placebo group (p < 0.001) (Table 2).O verall, patients with extracapsular fractures had a higher mean blood loss (1207.4ml, SD 701.2) than those with intracapsular fractures (958.7 ml, SD 515.3) but without statistical signifi cance (p = 0.052).There was no signifi cant difference between extracapsular and intracapsular fractures in transfusion rate (p = 0.169).For the risk of blood transfusion, multivariate analysis adjusted for potential factors (Table 3) revealed that only the TXA treatment (OR, 0.14; 95% CI, 0.04 to 0.4; p = 0.001) and Hb level on admission (OR, 0.52; 95% CI, 0.3 to 0.7; p = 0.001) were signifi cant predictors.Thus, TXA was a protective factor for transfusion, and inversely the placebo treatment was shown to be a risk factor (OR, 7.04; 95% CI, 2.2 to 22.4; p = 0.001).The H osmer-Lemeshow test determined a good fi t for the data from the logistic regression model (p = 0.984).

Safety outcomes
There were seven thromboembolic events within one year postoperatively, all in the placebo group (p = 0.006).Three The cumulative mortality (Table 2) was not signifi cantly different between groups at 30 days (p = 0.653) or 90 days (p = 0.509).At one year, four (6.1%) patients in the TXA group and nine (14.0%) in the placebo group had died, but this difference was not signifi cant (p = 0.115).Only postoperative infection (p = 0.034) showed a signifi cant association with 1-year mortality in the univariate analysis.No other covariate presented a p -value less than 0.1, so the multivariate analysis was rejected.In the subgroup analysis according to the fracture type, there was no signifi cant difference between extracapsular and intracapsular fractures in thromboembolic event rate (p = 0.612) or 1-year mortality (p = 0.113).

Discussion
The main fi nding of this study was that the administration of 1 g of intravenous TXA at the start of the surgery signifi cantly reduced total blood loss and transfusion rate.The use of TXA was not associated with an increased risk of thromboembolic events or 1-year mortality.There was a high rate of excluded patients, although it is well known that hip fractures in the elderly have a certain rate of loss of follow-up and mortality.

Signifi cant reduction of blood loss and need for transfusion
were also reported by most randomised studies that used a single dose [2,4,16], two doses [1,17,18], or three doses [5] of intravenous TXA.Conversely, Zufferey, et al. found no signifi cant differences in reducing blood loss or transfusion between the TXA and placebo groups, even though they used two doses, one at skin incision, and another three hours later [19].
Given the heterogeneity of the intravenous doses used in the different studies, a comparison is diffi cult.In the only study randomising one dose, two doses and placebo, the authors found no signifi cant difference between one-dose and  two-dose groups in blood loss, but the transfusion rate was signifi cantly lower in the two-dose group [20].Nevertheless, analysing the combined data from randomised studies, the mean reduction of blood loss using one-dose of TXA compared to the placebo group was 30%, and the decrease in the transfusion rate was 55% [2,4,11,16].With two doses of TXA these data were 32% and 52%, respectively [1,3,17,18,21], and with three doses were 33% and 57%, respectively [5].The above data suggested that the effectiveness of intravenous TXA was similar regardless of the number of doses.Recent metaanalyses also found that the frequency and dosage of intravenous TXA did not infl uence its benefi cial effect [22,23].Compared to the topical administration of TXA, several metanalyses reported that intravenous TXA was more effective in decreasing the need for transfusion [22,24].
As in the present study, Nikolaou, et al. found that preoperative Hb level and TXA were the only signifi cant predictor of transfusion [2].In agreement with other authors, total blood loss was higher in extracapsular fractures than intracapsular fractures, although the difference was not statistically signifi cant [25,26].Nikolaou, et al. reported signifi cant differences in blood loss and transfusion rates between patients with extracapsular fractures who received intravenous TXA and those who received a placebo, but these differences were not signifi cant in the patients with intracapsular fractures [2].In the present study, extracapsular fractures presented greater blood losses and allogenic blood transfusions compared to intracapsular fractures, although not reaching statistical signifi cance as a risk factor for transfusion.
Regarding the safety of the TXA treatment, confl icting results have been reported.Zufferey, et al. reported nine (16%) thromboembolic events with two doses of intravenous TXA according to the patient weight, given at skin incision and three hours later, and three (6%) in the placebo group, and although that difference was not statistically signifi cant, the authors did not recommend TXA for hip fracture surgery [19].
Conversely, in another randomised study administering TXA pre-and post-operatively, Tengberg, et al. found no thromboembolic events in the TXA group [1].However, regarding 90-day mortality, they did fi nd a higher rate in the TXA group (27%) compared to the placebo treatment (10%), but although this difference seems important, it was not statistically signifi cant and the infl uence of TXA on excess mortality could not be determined [1].In addition, that study was underpowered because the inclusion of patients was stopped for administrative reasons unrelated to the results.
Contrary to those outcomes, all the remaining randomised studies with a follow-up of at least three months reported similar rates of a thromboembolic event, complication, and 90-day mortality between TXA and placebo groups using two [17,18] or one [4] doses of intravenous TXA.Using two doses, Zhang, et al. reported two (3%) thromboembolic events in the TXA group and one (2%) DVT in the placebo group, and the 90day mortality rate was 1.6% and 3.3%, respectively [17].Watts, et al. also with two doses, reported a thrombotic events rate of 7% in the TXA group and 9% in the placebo group with no signifi cant difference between them, and the 90-day mortality rate was 14% in the TXA group, and 16% in the placebo group [18].In the only randomised study with a follow-up of six months, Chen, et al. used three doses of intravenous TXA, and they found an incidence of thromboembolic events of 16% in the TXA group and 14% in the placebo group, while the 6-month mortality was 6% in the TXA group and 3.5% in the placebo group, with no signifi cant differences between groups [5].
With one dose of intravenous TXA as in the present study, Ma, et al. reported 14% of DVT in the TXA group, and 13% in the placebo group, while no other thrombotic events were observed after a follow-up of three postoperative months [4].Zhou, et al. also using one dose and a follow-up of one month, found three (3%) thromboembolic events in the TXA group, and seven (7%) in the placebo group, although that difference was not signifi cant [18].Similar fi ndings were found in other randomised studies that used only one dose of intravenous TXA, although their follow-ups were between 48 hours and four days [2,11].In the present study, there were three DVT and three strokes in the placebo group, while in most randomised studies DVT represented the most frequent vascular complication, regardless of assigned treatment.
S everal recent metaanalyses reported that the short followup of the published trials limited the drawing of defi nitive conclusions on the TXA safety for hip fractures [24,27].To our knowledge, the present study has the most extended followup to date evaluating potential adverse events of TXA for hip fracture surgery.Despite the fact that the mean duration of the effect of IV TXA is 3 hours, we considered establishing a 1-year follow-up to analyze mortality in patients with hip fractures who, in addition to the risk inherent to their traumatic process and subsequent surgery, could be associated with complications derived from the use of TXA (thromboembolic events) that could infl uence the mortality rate [7,8].Other studies also performed a 1-year follow-up, but in cardiac surgery [28,29].
Nevertheless, the study also had several limitations.Due to safety considerations, high-risk patients were excluded from the study.Thus, the results may not be generalizable, and the safety of TXA in those patients remains unproven.
In the analysis of subgroups according to the fracture type, there was a signifi cant difference in surgical time, which could compromise the homogeneity of the samples.However, we believe that this does not infl uence the effectiveness or safety of the administration of intravenous TXA.The study involved multiple surgical procedures and was dominated by internal fi xation procedures, increasing the potential for bias.There is no standard reference for recording perioperative blood loss.
For the primary effectiveness outcome, the mathematical calculation of the blood loss based on clinical measurements proposed by Good, et al. was used [15].Like others, this method is not validated and maybe a source of error with a tendency to overestimate blood loss.However, it was applied to compare two groups, with a standardized postoperative fl uid therapy used to minimize the effect of isovolumetric hemodilution.The Good's method is usually used in randomised studies of TXA [1,2,20] and it is considered the most reliable [30].Regarding the safety outcome, subclinical or asymptomatic DVT might have gone undetected.On the other hand, the sample size was based on blood loss.Given the low frequency of thromboembolic events, the study may be underpowered to detect differences in these complications.Furthermore, the optimal timing and dosage of intravenous TXA are still unclear, but the administration of a standardized single dose used in the present study has shown its effectiveness in other randomised studies [4,16] and similar to those that used more than one dose [17,18].

Conclusion
The use of a single dose of intravenous TXA at the start of the surgery reduces blood loss and requirement for transfusion signifi cantly without increasing the risk of thromboembolic events or mortality within one year postoperatively in patients older than 65 years with hip fractures undergoing surgery.
To calculate the patient's blood volume (BV, measured in liters), different formulas were used according to the gender of the patient:Citation: Miralles-Muñoz FA, Martin-Grandes R, Martinez-Mendez D, Mahiques-Segura G,Lizaur-Utrilla A, et al. (2023) Intravenous single dose of tranexamic acid safely reduces blood loss and the need for transfusion in elderly patients with hip fracture.A randomized double-blinded controlled trial at 1-year follow-up.Ann Musculoskelet Med 7(2): 009-015.DOI: https://dx.doi.org/10.17352/amm.000032 patients were diagnosed with symptomatic DVT at 20, 89 and 91 postoperative days, three others had a cerebral stroke at 43, 44 and 236 postoperative days, and one patient suffered a myocardial infarction 35 days after surgery.There was no case of pulmonary embolism.Concerning the fracture type, there were no signifi cant differences in thromboembolic event rate (p = 0.612).Multivariate analysis for the risk of thromboembolic events was not performed due to the low rate.The postoperative infection rate was 4.6% (three patients) in the TXA group compared with 7.8% (fi ve patients) in the placebo group (p = 0.350).All those patients required early revision surgery.There was no signifi cant relationship between infection and surgery time (p = 0.437) or blood transfusion (p = 0.529).Except for deceased patients, there were no cases of readmission.Citation: Miralles-Muñoz FA, Martin-Grandes R, Martinez-Mendez D, Mahiques-Segura G, Lizaur-Utrilla A, et al. (2023) Intravenous single dose of tranexamic acid safely reduces blood loss and the need for transfusion in elderly patients with hip fracture.A randomized double-blinded controlled trial at 1-year follow-up.Ann Musculoskelet Med 7(2): 009-015.DOI: https://dx.doi.org/10.17352/amm.000032

Ethics approval:
This study was performed in line with the principles of the Declaration of Helsinki.The Institutional Review Board approved this randomized study.Ethical Committee of Clinical Investigation, Elda University Hospital, Miguel Hernandez University, No: PI2018-141.The study was included in a public registry (ClinicalTrials.gov NCT03211286) before beginning enrollment of patients.

Table 1 :
Baseline data in both groups.

Table 3 :
Predictors of blood transfusion in multivariate analysis.