Scoliosis is the most common spinal disorder in children and can be secondary to idiopathic, neuromuscular, or congenital etiologies. Posterior spinal fusion (PSF) is indicated for many of these patients to correct the deformity and prevent further curve progression.1 The extensive exposure and instrumentation required in these procedures can result in substantial blood loss, ranging from 800 to >2700 mL.2–4 Despite common blood conservation strategies and improved surgical technique, greater than half of all patients undergoing PSF require blood transfusion.5 Although generally considered safe, allogeneic blood transfusion carries numerous risks, including increased transfusion transmitted and hospital-acquired infections, alloimmunization, transfusion-related acute lung injury, and transfusion-associated circulatory overload.6–10
Tranexamic acid (TXA) is a lysine analog antifibrinolytic drug that has been shown to reduce bleeding in cardiac surgery,11 craniosynostosis surgery,12 total joint arthroplasty,13 obstetrical procedures, and urologic procedures.14 In the pediatric orthopaedic literature, TXA has also been shown to reduce blood loss and transfusion requirements in PSF for both neuromuscular scoliosis4,15,16 as well as idiopathic scoliosis.3,17,18 The efficacy of TXA has been postulated to be dose-dependent,2,17 but no consensus has been reached on the most appropriate dosing regimen.2–4,15 The recommended TXA dose for spinal fusion in idiopathic scoliosis patients should optimize the balance between clinical efficacy, safety, and health care costs. However, the relationship between TXA dose and clinical efficacy in this population is poorly understood. The objective of this study was to assess the impact of a high-dose TXA regimen on intraoperative blood loss and perioperative blood transfusion requirements as compared with a low-dose regimen in a cohort of pediatric idiopathic scoliosis patients undergoing PSF.
After receiving institutional review board approval, we acquired electronic medical record data from a web-based intelligence portal (IMPACT Online, Haemonetics Corp., Braintree, MA) and the corresponding billing database for all inpatients discharged from our institution between January 2009 and June 2015. Use of this database and our quality control methods have been described previously.19,20 For the purposes of risk-adjustment, collected data included the initial hemoglobin measured during the hospitalization, Charlson comorbidity index, and the weighted Medicare severity diagnosis-related group (MSDRGWt), a comprehensive surrogate marker of overall disease severity and complexity of hospital procedures also known as the casemix index. TXA dose was collected by review of electronic anesthesia records. The number of vertebral levels fused was determined by reviewing operative notes. Estimated blood loss (EBL) was determined according to usual practice accounting for irrigation, blood loss on sponges, and blood loss in suction canisters.
The 2 TXA dosing regimens investigated were as follows: (1) a 10 mg/kg loading dose and a 1 mg/kg/h maintenance dose; (2) 50 mg/kg loading dose and a 5 mg/kg/h maintenance dose—designated as low and high dosing, respectively. The loading dose was given intravenously at the start of the procedure followed by the continuous infusion for the remainder of the case. The choice of dosing regimen was based on the preference of the anesthesiologist, but the general trend over time was a gradual transition from the low-dose to the high-dose regimen. At the time of this study, an intraoperative hemoglobin transfusion trigger of 9 g/dL and postoperative trigger of 8 g/dL were standard practice at our institution.
Exclusion Criteria and Outcomes Assessment
Figure 1 illustrates the study profile of excluded and included patients, and the reasons for exclusions. From January 2009 to June 2015, 589 PSF surgeries were performed by a single surgeon (P.D.S.) at our institution. Three hundred fifty-eight PSFs were performed for idiopathic scoliosis according to principal diagnosis codes. Thirty-six patients were over 18 years of age and were excluded. Of the 322 patients, 18 years of age or younger, 229 (71%) of these patients received an antifibrinolytic intraoperatively: TXA in 206 (90%) and aminocaproic acid in 23 (10%). Seventy-two patients received the low TXA dose and 44 patients received the high TXA dose. The remaining 90 patients receiving other random doses of TXA were excluded. The patients who received no antifibrinolytic were younger, with lower body mass, having less invasive procedures (less spinal levels fused), and had lower starting hemoglobin levels. Given these dissimilarities, we chose not to include them as a control group.
The primary outcomes were EBL (mL) and transfusion requirement (in number of units of packed red blood cells (RBCs), fresh-frozen plasma, platelets, and total blood product units). EBL was also reported per vertebral level fused, and per kilogram of body mass. Transfusion requirements were subdivided into intraoperative and perioperative (whole hospital) transfusion requirements.
Morbid events were assessed during the hospital stay using ICD-9 codes from the hospital’s billing database as we have previously described.21 Composite morbidity was defined as the occurrence of any one of the following events: (1) infection (C. difficile, sepsis, surgical-site infection, or drug-resistant infection), (2) thrombotic event (deep venous thrombosis, pulmonary embolus, or disseminated intravascular coagulation), (3) kidney injury, (4) respiratory event, and (5) ischemic event (myocardial infarction, transient ischemic attack, or cerebrovascular accident). Conditions that were flagged as present on admission were not considered to be hospital-acquired morbid events.
Patient baseline characteristics were compared using Student t test, Wilcoxon test, and Pearson χ2 where appropriate. EBL and transfusion requirements between groups were compared with Student t test. Known predictors of transfusion in spinal fusion surgery6 were entered into a univariate analysis along with the variable of interest and then compared in a multivariable linear regression model to assess TXA dose as an independent predictor of transfusion requirements, adjusted for potential confounding variables that have been associated with transfusion requirements in previous studies. Results are reported as parameter estimates with 95% confidence intervals. P<0.05 (2-tailed) was considered to be significant.
Baseline patient characteristics for the low-dose and high-dose groups are summarized in Table 1. Each group had similar preoperative demographics, including age (P=0.55), sex (P=0.81), weight (P=0.71), and MSDRGWt (P=0.77). Surgical characteristics were also similar between the 2 groups, including first hospitalized hemoglobin (a surrogate for preoperative hemoglobin) (P=0.35), nadir hemoglobin (P=0.38), discharge hemoglobin (P=0.75), number of levels fused (P=0.35), and surgical duration (P=0.69). Because of practice changes over time, the majority of the low-dose cases (69/72, 95%) were performed before the year 2014, whereas most of the high-dose cases were performed in the year 2014 or later (37/44, 84%).
The differences in EBL between the low and high TXA dose group are shown in Table 2. The low-dose group had an EBL of 968 mL compared with the high-dose group’s 695 mL, which represents a 28% decrease in EBL when the high-dose regimen was used as compared with the low-dose regimen (P=0.01). This decrease in EBL was significant when assessed per number of vertebral levels fused (P=0.001) and per kilogram of patient body mass (P=0.02).
Table 3 demonstrates the difference in transfusion requirements for the low and high TXA dose groups. The high-dose TXA regimen was associated with a nearly 60% decrease in RBC units transfused for both the entire hospitalization (P=0.04) and during intraoperative time period (P=0.01). The high-dose group also had decreased intraoperative transfusion of platelets (P=0.03) and total blood component units (P=0.03).
Lastly, we analyzed predictors of RBC transfusion requirements in both an unadjusted and risk-adjusted manner. After adjusting for age, first hemoglobin, MSDRGWt, and the number of vertebral levels fused, the variables MSDRGWt (P=0.01), number of levels fused (P<0.0001), and TXA dose (P=0.01) were independently associated with perioperative RBC transfusion requirements (Table 4). Similarly, after adjusting for the same parameters, number of levels fused (P<0.0001) and TXA dose (P=0.01) were independently associated with intraoperative RBC transfusion requirements (Table 5).
Occurrence rates for the composite morbidity outcome were 3/44 (6.8%) and 2/70 (2.8%) in the high-dose and low-dose groups, respectively (P=0.31). The rate of infection was 2/44 (4.6%) and 0/72 (0%) in the high-dose and low-dose groups, respectively (P=0.14). The rate of respiratory morbidity was 1/44 (2.3%) and 2/72 (2.8%) in the high-dose and low-dose groups, respectively (P=0.87). It is important to note that there were no thrombotic or ischemic events in either dosing group. There were also no deaths during the hospital stay in either group.
The present study represents the largest study to date comparing high and low TXA dose regimens in idiopathic scoliosis patients undergoing PSF. In our study, the high-dose TXA group had significantly less intraoperative blood loss (by ≈30%) and a decreased RBC transfusion requirement (by ≈60%), when compared with the low-dose TXA group. There were no differences between groups for morbid event rates, and no thrombotic events in either group.
TXA is a synthetic antifibrinolytic that competitively inhibits the lysine binding sites on plasminogen and plasmin, thus reducing fibrinolysis. Major surgery such as scoliosis surgery, can release tissue factor, inflammatory mediators, and activate the coagulation cascade, and can lead to a biochemical switch to hyperfibrinolysis which in turn can increase bleeding. For this reason, TXA has been hypothesized to decrease bleeding during major surgery not only through its antifibrinolytic effect, but also by the inhibition of plasmin formation, which may help prevent inflammation and platelet degradation.17
Of the 11 studies investigating TXA in spinal fusion surgery for pediatric scoliosis correction to date, 4 different combinations of loading and maintenance doses were used and the magnitude of these loading and maintenance doses varied by as much as a factor of 10.2–4,15–18 Furthermore, the only study comparing different doses of TXA in pediatric spinal surgery was underpowered, with only 26 patients retrospectively analyzed.2 A recent systematic review concluded that in pediatric spinal surgery TXA reduces blood loss and transfusion requirements, but most studies have been small, single center, and retrospective, and further studies using pharmacokinetic modeling are required to determine the optimal dosing strategy. Only 1 study to date in pediatric spine patients has compared the efficacy of 2 doses of TXA, with a low-dose consisting of 10 mg/kg loading dose and a 1 mg/kg/h maintenance dose, and a high-dose consisting of 20 mg/kg loading dose and a 10 mg/kg/h maintenance dose. This retrospective study showed a trend for decreased perioperative transfusion requirement (50%) of blood products in the high TXA dosing regimen compared with the low-dose regimen (P=0.07); however, this study was underpowered with only 11 patients in the high-dose and 15 patients in the low-dose groups.2
In previous studies, mostly comparing TXA versus placebo, a wide range of TXA dosing has been utilized in patients with idiopathic scoliosis undergoing spinal fusion, and these doses are summarized in Figure 2.2–4,15,18,22–24 Five studies used a TXA dosing regimen that was 2-fold higher than that used in the high-dose regimen of our study.4,15,18,22 At this very high-dose (a loading dose of 100 mg/kg and a maintenance dose of 10 mg/kg/h), a randomized controlled trial in idiopathic scoliosis4 showed a decrease in EBL (∼40%) compared with placebo. Using the same very high TXA dosing regimen, 2 retrospective studies also showed a decrease in EBL (57% and 54%) and transfusion requirements (42% and 65%) compared with placebo.18,22 The other 2 retrospective studies’ population were different from the current study in that it also included neuromuscular scoliosis patients, who would be expected to have a higher mean blood loss.15,16 With a 20 mg/kg loading dose and a 10 mg/kg/h maintenance dose, a randomized controlled trial compared with placebo showed a decrease in both EBL (43%) and blood transfusion requirements (72%).23 Two randomized controlled trials have previously reported on the low TXA dose regimen that was used in the present study (with a loading dose of 10 mg/kg and a maintenance dose of 1 mg/kg/h). One of the studies found the low-dose TXA regimen was associated with a significant decrease in total (drain+EBL) blood loss (28%)24 and the other study found a decrease in total blood products transfused (30%) but not packed RBCs alone.3
The various TXA dosage schemes that are seemingly effective at reducing EBL and transfusion requirements compared with placebo in pediatric idiopathic scoliosis surgery, must be taken in light of limitations such as a positive publication bias. Furthermore, dosage schemes used in the aforementioned studies were not based on the pharmacokinetics of TXA, as such studies have not yet been conducted in the pediatric orthopaedic population. To maximize efficacy and minimize side effects, guidelines should be based on pharmacokinetic data and pharmacokinetic modeling to ensure the minimally effective dosage scheme is recommended. Although higher TXA dosage schemes (100 mg/kg and 10 mg/kg/h) have been shown to be associated with complications such as seizures, lower dosage schemes may be ineffective in terms of maximizing TXA’s antifibrinolytic effect and anti-inflammatory properties. However, previously, a higher TXA dose has not been adequately compared with lower TXA doses. The dose ambiguity is not unique to pediatric idiopathic scoliosis surgery. In adult spinal fusion surgery, the major studies demonstrating the efficacy of TXA use the following highly variable dosing protocols: a 10/1 dosing regimen,25 1 g loading dose with a 1 mg/kg/h maintenance dose,26 a 2 g loading dose27 and local/topical administration of TXA.28 Upon completion, an ongoing pharmacokinetic/pharmocodynamic trial in adolescent idiopathic scoliosis surgery will provide better guidelines of the efficacy and safety profile of TXA as well as the most effective dosage scheme (Goobie and colleagues clincaltrials.gov # NCT01813058).
There are certain limitations that should be recognized in our study. First, about 43% of our patients received random TXA doses that were different from the low-dose (10 mg/kg loading dose and a 1 mg/kg/h maintenance dose) and the high-dose (consisting of 50 mg/kg loading dose and a 5 mg/kg/h maintenance dose). These patients were excluded from analysis, but this is unlikely to introduce any systematic bias. One reason for the inconsistent dosing is that giving TXA to reduce perioperative blood loss is an “off-label” use of the drug, and thus there is no official recommended dose according to the package insert. In addition, our patients underwent surgery over a 6-year period during which surgical techniques and transfusion protocols may have changed. Including only 1 surgeon helps to protect against this bias. The retrospective nature of our study can introduce inherent biases but we attempted to adjust for these by controlling for known predictors of transfusion in our multivariable model. Another limitation is that intraoperative blood loss is truly an estimated value, and can be inaccurate due to irrigation and unaccounted blood loss. In a retrospective study, however, there is no reason for bias toward either TXA dose group in the way blood loss is measured. The fact that most low-dose cases were done early and the high-dose cases were done later during the study time period is a concern, but the patient characteristics and hemoglobin levels were very similar in the 2 groups over time, suggesting minimal confounding. Lastly, even with our relatively liberal transfusion triggers, only 29% of the patients in our cohort required a transfusion, which may have limited our ability to detect differences in RBC transfusion. The impact on TXA dose may therefore be different in centers with transfusion rates that are either higher, or lower, than ours.
Determining the optimal TXA dosing regimen in patients with idiopathic scoliosis undergoing PSF will require a nuanced understanding of the interplay between clinical efficacy, complications, patient-specific factors, and economic implications. A critical examination of each of these elements is beyond the scope of this study, but our data suggest that high-dose TXA regimens decrease blood loss and transfusion requirements as compared with low-dose regimens. These findings elicit the need for a well-designed, larger, prospective randomized controlled trial to determine the optimum TXA dosage scheme as it relates to the pharmacodynamic and pharmacokinetic profile in pediatric patients. Until such a study becomes available, practitioners may choose the higher dose TXA regimen in this patient population to minimize blood loss and transfusion requirements.
1. Hresko MT. Clinical practice. Idiopathic scoliosis in adolescents. N Engl J Med. 2013;368:834–841.
2. Grant JA, Howard J, Luntley J, et al. Perioperative blood transfusion requirements in pediatric scoliosis surgery: the efficacy of tranexamic acid. J Pediatr Orthop. 2009;29:300–304.
3. Neilipovitz DT, Murto K, Hall L, et al. A randomized trial of tranexamic acid to reduce blood transfusion for scoliosis surgery. Anesth Analg. 2001;93:82–87.
4. Sethna NF, Zurakowski D, Brustowicz RM, et al. Tranexamic acid reduces intraoperative blood loss in pediatric patients undergoing scoliosis surgery. Anesthesiology. 2005;102:727–732.
5. McLeod LM, French B, Flynn JM, et al. Antifibrinolytic use and blood transfusions in pediatric scoliosis surgeries performed at US Children’s Hospitals. J Spinal Disord Tech. 2015;28:E460–E466.
6. Basques BA, Anandasivam NS, Webb ML, et al. Risk factors for blood transfusion with primary posterior lumbar fusion. Spine (Phila Pa 1976). 2015;40:1792–1797.
7. Woods BI, Rosario BL, Chen A, et al. The association between perioperative allogeneic transfusion volume and postoperative infection in patients following lumbar spine surgery. J Bone Joint Surg Am. 2013;95:2105–2110.
8. Janssen SJ, Braun Y, Wood KB, et al. Allogeneic blood transfusions and postoperative infections after lumbar spine surgery. Spine J. 2015;15:901–909.
9. Goodnough LT. Risks of blood transfusion. Crit Care Med. 2003;31:S678–S686.
10. Rohde JM, Dimcheff DE, Blumberg N, et al. Health care-associated infection after red blood cell transfusion: a systematic review and meta-analysis. JAMA. 2014;311:1317–1326.
11. Faraoni D, Goobie SM. New insights about the use of tranexamic acid in children undergoing cardiac surgery: from pharmacokinetics to pharmacodynamics. Anesth Analg. 2013;117:760–762.
12. Goobie SM, Meier PM, Pereira LM, et al. Efficacy of tranexamic acid in pediatric craniosynostosis surgery: a double-blind, placebo-controlled trial. Anesthesiology. 2011;114:862–871.
13. Hiippala ST, Strid LJ, Wennerstrand MI, et al. Tranexamic acid radically decreases blood loss and transfusions associated with total knee arthroplasty. Anesth Analg. 1997;84:839–844.
14. Ducloy-Bouthors AS, Susen S, Wong CA, et al. Medical advances in the treatment of postpartum hemorrhage. Anesth Analg. 2014;119:1140–1147.
15. Shapiro F, Zurakowski D, Sethna NF. Tranexamic acid diminishes intraoperative blood loss and transfusion in spinal fusions for duchenne muscular dystrophy scoliosis. Spine (Phila Pa 1976). 2007;32:2278–2283.
16. Dhawale AA, Shah SA, Sponseller PD, et al. Are antifibrinolytics helpful in decreasing blood loss and transfusions during spinal fusion surgery in children with cerebral palsy scoliosis? Spine (Phila Pa 1976). 2012;37:E549–E555.
17. Yagi M, Hasegawa J, Nagoshi N, et al. Does the intraoperative tranexamic acid decrease operative blood loss during posterior spinal fusion for treatment of adolescent idiopathic scoliosis? Spine (Phila Pa 1976). 2012;37:E1336–E1342.
18. Ng BK, Chau WW, Hung AL, et al. Use of Tranexamic Acid (TXA) on reducing blood loss during scoliosis surgery in Chinese adolescents. Scoliosis. 2015;10:28–33.
19. Frank SM, Savage WJ, Rothschild JA, et al. Variability in blood and blood component utilization as assessed by an anesthesia information management system. Anesthesiology. 2012;117:99–106.
20. Frank SM, Wick EC, Dezern AE, et al. Risk-adjusted clinical outcomes in patients enrolled in a bloodless program. Transfusion. 2014;54:2668–2677.
21. Johnson DJ, Scott AV, Barodka VM, et al. Morbidity and mortality after high-dose transfusion. Anesthesiology. 2016;124:387–395.
22. Lykissas MG, Crawford AH, Chan G, et al. The effect of tranexamic acid in blood loss and transfusion volume in adolescent idiopathic scoliosis surgery: a single-surgeon experience. J Child Orthop. 2013;7:245–249.
23. Xu C, Wu A, Yue Y. Which is more effective in adolescent idiopathic scoliosis surgery: batroxobin, tranexamic acid or a combination? Arch Orthop Trauma Surg. 2012;132:25–31.
24. Verma K, Errico T, Diefenbach C, et al. The relative efficacy of antifibrinolytics in adolescent idiopathic scoliosis: a prospective randomized trial. J Bone Joint Surg Am. 2014;96:e80.
25. Farrokhi MR, Kazemi AP, Eftekharian HR, et al. Efficacy of prophylactic low dose of tranexamic acid in spinal fixation surgery: a randomized clinical trial. J Neurosurg Anesthesiol. 2011;23:290–296.
26. Bednar DA, Bednar VA, Chaudhary A, et al. Tranexamic acid for hemostasis in the surgical treatment of metastatic tumors of the spine. Spine (Phila Pa 1976). 2006;31:954–957.
27. Khurana A, Guha A, Saxena N, et al. Comparison of aprotinin and tranexamic acid in adult scoliosis correction surgery. Eur Spine J. 2012;21:1121–1126.
28. Krohn CD, Sorensen R, Lange JE, et al. Tranexamic acid given into the wound reduces postoperative blood loss by half in major orthopaedic surgery. Eur J Surg Suppl. 2003:57–61.