Tranexamic Acid in patients Undergoing Coronary-Artery Surgery

Myles PS. N Engl J Med 2017; 376:136-148. doi:10.1056/NEJMoa1606424

Clinical Question

  • In adult patients undergoing elective coronary-artery surgery does tranexamic acid (TXA) compared to a placebo increase the risk of death or thrombotic complications within the first 30 days after surgery?


  • Tranexamic acid is an anti-fibrinolytic agent that is commonly used to reduce bleeding in patients undergoing cardiac surgery
  • However, prothrombotic effects may increase the risk of peri-operative myocardial infarct, stroke, seizure or death


  • 2 x 2 factorial trial
    • The other side of the trial was aspirin administration before cardiac surgery: ATACAS Asp
  • Randomized control trial using computer generated codes
  • Treatment stratified in blocks according to site and on- or off-pump surgery
  • Blinding encouraged but given constraints surrounding availability of research personnel to prepare medications, anaesthetists were allowed to prepare the drug themselves
  • Planned for 90% power to detect a 10% vs. 7% (3% difference deemed clinically significant) difference between TXA and placebo and aspirin and placebo
    • When considering solely the TXA arm this resulted in a difference of 8.5% vs. 5.95% between the TXA and placebo groups and as such the aim was to recruit 4484 patients
  • Aspirin comparison ceased after 2127 patients enrolled
  • Modified intention-to-treat analysis
    • 32 patients were randomised but not analysed (see below)


  • 31 hospitals across 7 countries (Europe, Australia, New Zealand, Hong Kong, Canada)
  • March 2006 – October 2015


  • Adults at increased risk of major complications undergoing elective coronary artery surgery either on or off pump (other procedures such as valve replacements permitted)
  • High risk defined by any of:
    • Age > 70, at least moderate LV impairment, simultaneous valvular, aortic surgery or LV aneurysmectomy, re-do surgery, COPD, CKD (baseline creatinine > 150, CrCl < 45ml/min), BMI > 25, pulmonary hypertension (mPAP > 25 mmHg) and peripheral vascular disease
  • Exclusion criteria:
    • Poor English language, Clinician preference for antifibrinolytic therapy, need for urgent surgery, active peptic ulceration, allergy to aspirin or TXA, aspiring within 4 days, warfarin or clopidogrel within 7 days or GIIb/IIIa inhibitors within 24 hours, low platelet count or history of bleeding diathesis, severe renal impairment (Cr > 250, CrCl < 25 ml/min), recent haematuria, previous thromboembolic disease (history of post-operative or spontaneous PE or arterial thrombosis or familial hypercoaguable state), pregnancy
  • 11,253 eligible
    • 4,662 enrolled
    • 2,329 randomised to receive TXA
      • 19 excluded post randomisation (not analysed)
      • 39 did not receive study drug (analysed as intention to treat)
    • 2,333 randomised to receive placebo
      • 13 excluded post randomisation (not analysed)
      • 32 received open-label TXA (analysed as intention to treat)
  • 50% received Aspirin
  • Similar baseline characteristics in each group
    • Co-morbidities, pre-operative medications also similarly matched although higher numbers with a smoking history in placebo group (1482 patients vs. 1575 patients)
    • Intra-operative characteristics similar


  • Tranexamic Acid (TXA)
    • TXA 100mg/kg
    • Administered intravenously
    • 30 minutes after induction
    • Dose reduced to 50mg/kg in January 2012 (1392 patients enrolled, 758 given TXA) as seizures reported and this was felt to be dose related


  • Placebo
    • 0.9% NaCl given intravenously
    • 30 minutes after induction

In both groups

  • Routine surgical and peri-operative care
  • Guidelines for heparin / protamine use, excessive bleeding during surgery and transfusion thresholds provided
  • Open label TXA or other antifibrinolytic not allowed pre-bypass but was allowed if significant bleeding after protamine administration to reverse heparin
  • ECGs obtained pre-op, on the 1st, 2nd, and 3rd days after surgery and on hospital discharge
  • Cardiac biomarkers collected 12 – 24 hours and 48 – 72 hours post surgery
  • Non-fatal MI was defined as per the third universal definition in addition to solely elevated cardiac biomarkers (Troponin I > 10 ng/ml or Troponin T > 4.0 or CK-MB > 3 times upper limit of normal more than 12 hours post surgery)


  • Primary Outcome was a composite outcome of death and thrombotic events (non-fatal MI, stroke, PE, renal failure or bowel infarction) during first 30 days
    • TXA group: 16.7%
    • Placebo group: 18.1%
    • Relative Risk Ratio (as published): 0.92 (95% CI 0.81 to 1.05; P = 0.22)
    • Absolute Risk Reduction (calculated): 1.40 (95% CI -0.78% to 3.58%; P = 0.22)
  • Secondary outcomes:
    • No statistically significant difference was identified in the individual outcomes of the composite primary outcome
    • A statistically and clinically significant reduction in the rate of return to theatre was observed
      • TXA 1.4% vs. placebo 2.8%
      • ARR 1.42% (95% CI 0.59% to 2.24%; p=0.001)
      • NNT 71
      • Fragility Index 12
    • Blood products administered were also statistically significantly different
      • Total blood products: TXA 4331 vs. placebo 7994 (P < 0.001)
      • Transfused any blood product:
        • TXA 37.9% vs placebo 54.7%
        • ARR 16.8% (95% CI 14.0% to 19.6%)
        • NNT 6
        • Fragility Index 322
      • Use of TXA would prevent 57 units of blood products for every 100 patients treated.
  • Seizures were added as a safety outcome in 2012
    • TXA 0.7% vs placebo 0.1%
    • Relative Risk Ratio (as published): 7.62 (95% CI 1.77 to 68.71; P = 0.002
    • Absolute Risk Increase (calculated): 0.57% (95% CI 0.22% to 0.91%; P = 0.0012)
    • NNH 177
    • Fragility Index 4
  • Similar lengths of hospital stay 8.0 vs. 8.0 days

Authors’ Conclusions

  • The use of TXA was associated with a lower risk of bleeding post-operatively and was not associated with increased risk of death or thrombotic effects when compared to the administration of a placebo


  • Large, multi-centre trial
  • Appropriately powered to answer question
  • Varied, but balanced baseline characteristics results in good external validity for this cohort of patients undergoing elective surgery


  • 2 x 2 factorial design means some patients exposed to aspirin
    • Although no statistically significant interaction with those that received aspirin and no TXA (i.e. placebo), this group had higher rates of surgical re-exploration and need for blood transfusion within 24 hours than those that did receive both aspirin and TXA (3.6% vs. 1.3% for re-exploration and 52.7% vs. 33.9% for blood transfusion)
  • TXA dose changed half way through study – underpowered to perform post-hoc dose-effect analysis
  • Not all anaesthetists blinded to TXA or placebo
  • Only a small number of off-pump cases; this and the exclusion of emergent surgery where other medications that interfere with coagulation in the pre-operative period reduced the external validity to these groups
  • Large numbers eligible but not randomised
  • Number excluded after randomisation is higher than Fragility Index for some secondary and safety outcomes, reduces the certainty of the result

The Bottom Line

  • This trial provides valid reason to consider using TXA in elective patients undergoing elective coronary artery surgery to reduce the blood product requirements, however it may increase post-operative seizure incidence
  • Given some methodological issues around concomitant aspirin use and seizure risk, further work looking at the best dosing regime is required

External links


Summary author: George Walker
Summary date: 27 September 2017
Peer-review editor: Segun Olusanya

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.