• 
• 

Three Part Question

In [patients with neck of femur fractures] is [administration of tranexamic acid in the emergency department] [associated with a reduction in perioperative blood transfusions, perioperative blood loss, length of hospital stay, and mortality]

Clinical Scenario

An elderly patient presents to the emergency department (ED) with hip pain after a fall from standing. They are unable to weight bear. You notice extensive bruising around their left hip and an X-ray confirms a neck of femur fracture. You wonder if giving tranexamic acid (TXA) in the ED would improve their outcome when they later undergo surgery.

Search Strategy

EMBASE and Ovid MEDLINE(R) (ALL) databases were searched (1974-2024) utilising the Ovid interface and the following keyword strategy:
(exp *Tranexamic Acid/ or txa.mp.) and ((hip fracture.mp. or exp *Hip Fractures/ or exp *Femoral Neck Fractures/ or (hip.mp. or exp *Hip/ or (exp *Femur/ or femur.mp.))) and fracture.mp.)
No search filters were utilised, including language. A supplementary search of the Cochrane database was conducted using the same search terms: that is, a combination of ‘TXA / tranexamic*’ and ‘femo* / femu* / hip fracture /fractured hip) in the title or abstract. The Google Scholar ‘cited by’ function was then used to find studies that had referenced the papers we identified as relevant in our EMBASE, MEDLINE and Cochrane searches. Reference lists of relevant papers were screened for any studies missed by our search paradigm.

Search Outcome

386 papers were identified by our EMBASE and Medline search, 264 papers after deduplication. 258 were excluded following title and abstract review as they were not relevant to three-part question. One paper was identified by scanning the reference lists. Seven papers underwent full take review, four were not relevant to three-part question and three were retained for analysis.
251 papers were identified by our Cochrane search. 245 were excluded following title and abstract review as they were not relevant to three-part question (n = 242) or were duplicates of the EMBASE and Medline search (n = 3). Six papers underwent full text review, five were excluded as they were trial protocols (n = 1), discontinued (n = 1) or not relevant to the three-part question (n=3) and one paper was retained for analysis.
Four papers1–4 were retained for final analysis. Three were randomised controlled trials (RCTs)1,3,4 and one was a cohort study2. The key results of these studies are summarised in Table 1. No statistically significant findings were reported for length of hospital stay or mortality.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Yakel et al 2023 USA	Adults over 18 with a low energy closed intertrochanteric or subtrochanteric femur fracture. Randomised to 1x IV TXA dose at hospital admission or 1x placebo (saline) dose at hospital admission n = 100	Prospective single centre double masked RCT (Level 1b)	Receipt of perioperative blood transfusion	17.5% of the TXA group received perioperative blood transfusions compared to 36.7% of the control group, RR = 0.48 (p = 0.046).	Only included extracapsular hip fractures and excluded patients on anticoagulants or with multiple injuries. Small sample size, failing to meet authors’ power calculation for detection of the primary outcome of perioperative blood transfusions.
			Peri-operative blood loss	1181ml total peri-operative blood loss in TXA group compared to 1548ml in the control group (p = 0.01).
			Length of hospital stay
			30-day mortality
Moran et al 2022 USA	Patients admitted to a fragility hip fracture service n = 508	Prospective cohort study (Level 2b)	Receipt of perioperative blood transfusion	9.4% of patients given TXA at hospital admission received blood transfusion, compared to 25.7% of patients who received delayed TXA and 29.4% who received no TXA (p < 0.001).	Non-randomised study design with relatively few patients (n = 32) given TXA on admission.
			Incidence of DVT, PE, MI and stroke
Ma et al 2021 China	Adults over 65 presenting within 6 hours of an intertrochanteric fracture Randomised to 1x IV TXA dose at hospital admission or 1x placebo (saline) dose at hospital admission n = 125	Prospective single centre RCT (Level 1b)	Receipt of preoperative blood transfusion	11.1% of the TXA group received pre-op transfusions compared to 22.6% of the control group (p = 0.036).	Only included intertrochanteric hip fractures and excluded patients who were on anticoagulants or were unable to access hospital within 6 hours. Only pre-operative but not intra- or -post-operative transfusions recorded. Unclear whether patients, clinicians and outcomes assessors were masked. Unclear whether further doses of TXA were given to patients intra- or post-operatively.
			Serial haematocrit and haemoglobin measurements	There was 254ml of hidden blood loss in the TXA group compared to 408ml in the control group 3 days post trauma (p < 0.001). Average haemoglobin levels were 100g/L in the TXA group compared to 88g/L in the control group 3 days post trauma (p < 0.001). Average haematocrit levels were 33% in the TXA group compared to 27% in the control group 3 days post trauma (p < 0.001).
			Incidence of PE/DVT at 3 months and length of hospital stay
Owen A et al 2024 USA	Patients with a low energy isolated extracapsular peritrochanteric hip fractures Randomised to 1g IV bolus of TXA over 10 minutes and 1g infusion of TXA over 8 hours or the same regiment of a saline placebo. n = 128	Prospective single centre double masked RCT (Level 1b)	Receipt of at least one unit of perioperative blood transfusion	By day 4, 27% of patients in the TXA group had received a blood transfusion compared to 30.6% in the control group (p = 0.65).	Only included extracapsular peritrochanteric fractures and excluded patients with multiple injuries. No further does of TXA were permitted, although 22% of the TXA group received TXA prior to their operation and 9% in the control group received this. A post hoc power calculation showed the study was significantly underpowered to detect a difference in transfusion rates between the two groups. Only patients presenting to ED between 7am and midnight were included.
			Estimated perioperative blood loss	There was an estimated blood loss of 1,593 ml in the TXA group compared to 2,013 ml in the control group by day 4 post admission (p = 0.17)
			Incidence of PE/DVT, MI and stroke
			Incidence of death and readmission.	No significant difference between groups for rate of readmission.

Comment(s)

Two of the RCTs included in our review1,3 indicate that for patients with hip fractures, early TXA administration in ED decreases subsequent perioperative blood loss and three studies demonstrated a reduction in blood transfusion1–3. However, one RCT found no difference in estimated blood loss or blood transfusion rates, although a post-hoc power calculation demonstrated the study was significantly underpowered to detect a difference in blood transfusion rates.4 None of the studies found evidence of increased adverse events, such as venous thromboembolism1–4. The results of our review should be interpreted cautiously. The three RCTs1,3,4 restricted their sample to extracapsular hip fractures, limiting the generalisability of results to intracapsular fractures, which make up the majority of hip fractures5. However, extracapsular fractures are associated with greater blood loss6 and may therefore be a more clinically relevant sub-group for a TXA study. The trials in our review variously excluded patients who were taking anticoagulants1,3, multiply injuried1,3,4, or unable to present to hospital within six hours of injury1,3. This may limit the generalisability of their findings to our frailer patients, who commonly take anticoagulants, incur multiple injuries from falls, and present to hospital late – either because they cannot self-mobilise after falling, or struggle to access healthcare services in general. Unfortunately, it is this complex sub-group of patients who are likely to be the most vulnerable to the consequences of blood loss after hip fracture. Finally, it is unclear whether any benefits demonstrated by the studies in our review lead to improved outcomes for patients as none of the RCTs were powered to detect a difference in length of hospital stay, functional outcomes or mortality.1,3,4

Clinical Bottom Line

Early TXA administration in the ED for extracapsular femoral neck fractures appears safe and may reduce perioperative blood transfusions. The associated impact on patient-centred outcomes, such as length of hospital stay and mortality, is unclear. Currently, evidence for TXA in the ED for intracapsular hip fractures is insufficient.

References

1. Stefan Yakel, DO; Justin Than, DO; Jennifer Sharp, DO; Olivia Coskey, MPH; Hiroki Den, MD, MPH; Jacqueline Krumrey, MD The Efficacy of Tranexamic Acid for Reducing Blood Transfusion Rates in Extracapsular Hip Fractures: A Single-Center Randomized Controlled Trial
2. Jay Moran, Joseph B Kahan, Jensa Morris , Peter Y Joo, Mary I O'Connor Tranexamic Acid Administration at Hospital Admission Decreases Transfusion Rates in Geriatric Hip Fracture Patients Undergoing Surgery
3. Huixu Ma, Hairuo Wang, Xiaotao Long, Zexiang Xu, Xiaohua Chen, Mingjin Li, Tao He, Wei Wang, Lei Liu and Xi Liu Early intravenous tranexamic acid intervention reduces post-traumatic hidden blood loss in elderly patients with intertrochanteric fracture: a randomized controlled trial
4. Aaron R. Owen, MD, Chelsea C. Boe, MD, Nicolas P. Kuttner, BS, Alexandra M. Cancio-Bello, BS, Kristina M. Colbenson, MD, Krystin A. Hidden, MD, Jonathan D. Barlow, MD, William W. Cross, MD Tranexamic Acid Administered at Time of Hospital Admission Does Not Decrease Transfusion Rates or Blood Loss for Extracapsular Hip Fractures: A Double-Blinded Randomized Clinical Trial