Excessive bleeding and inflammatory response (IR) after cardiopulmonary bypass (CPB) are common complications of cardiac surgery [1]. Although tranexamic acid (TA), a lysine analog competitive inhibitor of plasmin and plasminogen [2], is effective in reducing bleeding after cardiac surgery, its anti-inflammatory effect in fibrinolysis blockade has been less studied.

In vitro studies suggest that the exposure of D-dimer fragments to monocytes initiates the synthesis and release of IL-6 [3]. Excessive plasmin activity and/or D-dimer formation may play an important role in proinflammatory cytokine and cellular response activation during CPB [4]. In patients undergoing elective CPB, prophylactic blockade of fibrinolysis decreases molecular expression of IR [5]. This IR may contribute to postoperative complications, including myocardial dysfunction, respiratory failure, renal and neurologic dysfunction, bleeding disorders, altered liver function, and ultimately, multiple organ failure (MOF) [4]. In a previous study using TA before and after CPB we observed reduced bleeding and IR [6]. These results were encouraging but left unanswered the question of whether the postoperative dose, coinciding with greatest fibrinolytic activation, was determinant in decreasing postoperative IR and bleeding. Nevertheless newer data indicate that, in a dose-dependent fashion, TA is associated with an increase of adverse events, particularly the observation of seizures [7]. The incidence of this adverse effect varies from 2.7% to 4.6% in any major surgical procedure, independently of dosing schedule [8–10]. For this reason, we also focused on the safety profile of TA dosing schedule as a part of the study.

The study was approved by our university hospital ethics committee and conducted according to the Helsinki Declaration. We report this trial in accordance with the CONSORT statement revised recommendations [11].

From December 2005 to January 2007, we performed a randomized, double-blind, dose-dependent parallel-groups study with 160 consecutive adult patients undergoing elective CPB, in a 24-bed intensive care unit (ICU) at a university hospital.

Exclusion criteria included: no informed consent, age < 18 years, emergencies, off-pump cardiac surgery, chronic coagulopathy (prothrombin time [PT] <50% or international normalized ratio (INR) >2 and platelets <50,000/mm³ or aggregation dysfunction), renal failure (creatinine >2 mg/dL), gross hematuria, TA hypersensibility, chronic hepatopathy (Child-B or higher), immunosuppression, endocarditis and post-operative sepsis within 24 h.

Before CPB, all participants had normal bleeding time, platelet collagen/epinephrine and collagen/ADP closure time, PT, activated partial thromboplastin time, and thrombin time. No patients received anti-inflammatory or immunosuppressive agents during 5 days before and 24 h post-CPB.

After obtaining written informed consent, the surgeon requested treatment group assignment immediately before CPB.

Demographic variables, comorbidity, perioperative clinical data, and postoperative IR, mechanical ventilation time, ICU and hospital stay, and mortality, were recorded. Core body temperature, laboratory data (hematology, inflammation, coagulation, and fibrinolysis), and hemodynamic parameters were recorded before intervention (baseline), on ICU admission after surgery (0 h), and at 4 h and 24 h post-CPB, once hemodynamic stability was confirmed. IR was clinically defined as core body temperature >38°C (100.4°F), systemic vascular resistance index <1,600 dyne · sec/cm⁵ per m², and cardiac index >3.5 L/minute per m² at 4 h, as we reported in a previous study [6]. We also recorded blood loss (chest-tube drainage and hemoderivatives) at the above time points and on chest tubes removal. In cases of reintervention due to bleeding, the post-hoc classification of bleeding into "surgical" and "non-surgical" was applied to distinguish between bleeding due to mechanical-surgical causes (where pharmacological measures are not effective) and coagulopathy (which TA was expected to reduce). Surgical risk was calculated by Euroscore and Parsonnet score.

Cerebrovascular events included seizure and stroke. Acute kidney injury was assessed applying RIFLE [15] criteria (>100% creatinine increase, using preoperative and highest creatinine concentration during the first week after surgery); renal failure was defined as dysfunction requiring dialysis. Myocardial infarction was considered as either new Q waves or new, persistent ST-segment or T-wave changes. Indication for re-operation was determined by clinical judgment and blood loss as >200 mL/h in three consecutive hours. Mortality was defined as death within 30 days of CPB, recorded in medical history or telephone contact with the surgeon responsible or family. These adverse effects were included as a composite variable.

IL-6 (normal range: <5.9 pg/mL; intra-assay variation: 4.5%) was measured by automatic immunoenzyme assay (IMMULITE ONE™; Diagnostic Products Corporation).

D-dimer (normal range: < 300 ng/mL; intra-assay variation: 3%) was measured by immunoturbidimetric test (D-dimer PLUS; Dade Behring).

From December 2005 to January 2007, we recruited 209 consecutive CPB patients; 49 were excluded (20 off-pump, 4 coagulation disorders, 11 emergencies, 6 endocarditis, 4 hemodialysis, and 4 immunosuppressed). Thus we studied 160 patients, 80 receiving single-dose TA before CPB and 80 receiving TA before and after CPB. Figure 1 shows a flow chart of patients enrolled.

Patients developing IR showed significantly higher levels of IL-6 at 4 hours postoperatively than those without IR, [F (2, 274) = 5.08; P = 0.013], (Figure 2).

The incident rate of IR was significantly lower in TA2 (7.5%) than in TA1 (18.8%) (P = 0.030). TA2 showed OR 0.29 (95%CI: 0.10-0.83), (P = 0.013) after adjusting for hypertension, total protamine dose and temperature after CPB. Compared with TA2, the relative risk of IR was 2.5 for TA1 (95% CI 1.02-6.12). The absolute risk difference was 11.3%. The number needed to treat with TA2 to reduce IR was 9 patients (95% CI: 5-107 patients).

In the omnibus test TA2 had significantly lower chest-tube bleeding (P = 0.014). Comparing both groups for chest bleeding at 24 hours, TA2 had lower bleeding than TA1 [978 (95% CI: 809-1147) vs 1198 (95%CI: 1017-1380)mL; P = 0.010, corrected-P significant], which remained significant after adjusting for surgery-related bleeding (P = 0.014, corrected-P significant). On summing intraoperative and 24 h postoperative bleeding, the effect of TA2 was significant: [1341 (95% CI: 1197-1486) vs 1538 (95%CI: 1391-1686) mL; P = 0.016, corrected-P significant]. Also in the omnibus test, TA2 had significantly lower D-dimer (P = 0.038) than TA1 (Figure 3). TA2 had lower D-dimer levels at 24 hours [489 (95% CI: 437-540) vs. 621(95% CI: 563-679) ng/mL; P = 0.01, corrected-P significant].

TA2 showed significant reduction of maximum CPK values [407(95%CI: 259-671) vs 487(95%CI: 359-812)U/L; P = 0.01] and TnI [2.6(95%CI: 1.7-4.7) vs 3.6(95%CI: 2.7-6.3)U/L; P > 0.01]. TA2 required lower levels of norepinephrine at 24 h [0.06 (95%CI: 0.03-0.09) vs 0.20 (95CI5%:0.05-0.35), P = 0.03, adjusted for dobutamine [F(1.40) = 6.6; P = 0.014, corrected-P significant].

We found a direct relationship between IL-6 at ICU arrival and: temperature (rho = 0.26; P < 0.01), D-dimer (rho = 0.24; P < 0.01), norepinephrine (rho = 0.33; P < 0.01), TnI (rho = 0.37; P < 0.01), CK(rho = 0.37; P < 0.01), CK-MB(rho = 0.33; P < 0.01) and lactic acid (rho = 0.46; P < 0.01) at arrival. Similarly, a direct correlation was observed between IL-6 at 4 h with temperature (rho = 0.23; P < 0.01), norepinephrine (rho = 0.25; P < 0.01), TnI (rho = 0.42; P < 0.01) and lactic acid (rho = 0.19; P = 0.03) at 4 h and chest tube bleeding at 24 h (rho = 0.18; P = 0.03) and at chest-tube withdrawal (rho = 0.20; P = 0.02), and an inverse relationship with systolic blood pressure at 4 h (rho = -0.17; P = 0.05).

Seizure was present in 2 (1.3%) and stroke in 3 (1.9%) patients. Acute kidney injury occurred in 14 (8.8%) patients, and 7(4.4%) needed dialysis. Finally, 3 (1.9%) suffered myocardial infarction. Seven patients required re-operation for bleeding: 5 in TA1 and 2 in TA2,(P = 0.44). There were 9 deaths, without significant differences between groups: 3 (3.8%) in TA1 versus 6 (7.6%) in TA2 (P = 0.29). Statistically, there were no significant differences in expected vs. observed mortality between the two groups. According to logistic Euroscore, the rates of mortality were: TA1-observed = 3.8% (3/80) and TA1-expected = 3.25% (3/80), (P = 0.99) and TA2-observed = 7.6% (6/80)] and TA2-expected = 3.27% (3/80), (P = 0.11). There were no significant differences between groups in cerebrovascular events, renal pathology, myocardial infarction, mechanical ventilation or ICU stay after surgery, nor when comparing composite adverse effects (Table 2).

The clinical definition of IR based primarily on early hyperthermia with a hyperdynamic state is accompanied by higher levels of IL-6 at 4 h than in other patients. Peak levels of this interleukin is observed in these first hours of the postoperative period, as found by our team in a previous study [6] and by others [16, 17]. The presence of hyperthermia in the criteria used avoids the potential confounding effect of vasoactive drugs on the associated hyperdynamic state.

Our findings indicate that prolonged inhibition of fibrinolysis by a postoperative dose of TA reduces IR and bleeding in CPB patients. TA1 patients presented higher levels of D-dimer, greater bleeding and IR than TA2 patients who received a second dose after CPB.

Several mechanisms have been proposed to explain IR after CPB, including contact activation, ischemia-reperfusion, and endotoxemia. These triggers may activate numerous systems involving complement, cytokines, immune cellular response and coagulation-fibrinolytic cascades [4]. These systems are closely interconnected and provide continuous feedback, so the release of cytokines[18] or activation of the complement system [19] may amplify the fibrinolytic response. This in turn may re-activate the release of inflammatory mediators by plasmin and D-dimer [3, 20, 21]. This amplified IR is especially relevant in CPB.

In a previous study, we observed significantly higher levels of IL-6 in patients with IR and TA was effective in reducing this response [6]. Thus in the present study we focused on IL-6 and fibrinolysis. In addition to the feed-back between the release of inflammatory markers and fibrinolysis, other components must be considered in the release of these markers. After CPB (Figure 2) we observed a clear increase in the levels of IL-6 and other interleukins, which interact with each other [16, 17] and are involved in producing fever ("postperfusion syndrome") [22]. Also, oxidative stress induced by norepinephrine may release IL-6 in other settings [23]. In this study, we observed no differences in IL-6 levels between groups, probably because both them received antifibrinolytic.

Proinflammatory cytokines may contribute to inflammatory response, myocardial ischemia-reperfusión injury and hemodynamic instability after clinical CPB. This myocardial dysfunction has been highlighted in diverse studies through the association between cytokine levels (IL-6) and markers of myocardial tissue damage such as TnI y CK [24, 25]

In the present study, patients developing IR showed higher levels of IL-6 than those who did not, together with worse hemodynamic status. The association between greater cytokine (IL-6) release, activation of fibrinolysis (D-dimer), higher temperature, higher levels of TnI, CK, CK-MB, and increased hemodynamic instability (norepinephrine requirements and higher lactate levels) in the immediate postoperative period support the inter-relation between inflammation and fibrinolysis [18–20].

The protective effect of TA regarding IR, more evident in the TA2 group, was also reflected in less postoperative myocardial tissue damage.

Optimal use of inotropes or vasopressors in the perioperative period of cardiac surgery remains controversial. We usually employ a combination of dobutamine plus norepinephrine in patients who develop postoperative cardiac dysfunction when there is associated hypotension, to maintain an adequate perfusion pressure. This would justify the use of norepinephrine in patients who do not develop IR [26, 27]. Although the initial analysis revealed no differences in the need for norepinephrine in the first 24 hours postoperatively between the treatment groups, differences did appear at 24 h after adjusting for the dose of dobutamine at the time.

Although TA effectively reduces bleeding after cardiac surgery, doses vary widely [12, 13]. Most are based on an initial dose and subsequent infusion, which can result in high cumulative doses. TA half-life is 80 min, and the effect of a single dose (TA1) is nearly 100% inhibition of plasmin activity[14] in the immediate postoperative period, so a second dose administered at the end of surgery may prevent IR and bleeding complications, since enhanced fibrinolysis might be expected after heparin reversal [28].

Thus the dosing schedule used in our study ensures complete inhibition of fibrinolysis, although no pharmacokinetic information was available. With the second dose we wished to investigate the possible benefit of postoperative inhibition of fibrinolysis on IR.

It is questionable whether the statistically smaller blood loss in TA2 group is really clinically relevant, in terms of transfusion requirements. The use of hemoderivatives partly depends on the transfusion policy of each center, as in our case. However, given that numerous authors concur on the need for antifibrinolytic agents in cardiac surgery [29], we believe the protective effect of the TA2 dosing schedule used in the present study on IR and clinical parameters should be considered.

The incidence of non-ischemic clinical seizure varies from 2.7% to 4.6% in most major surgical procedures [8–10], only 0.4% in CABG [9], but as high as 6.7% [8] or 7.9% [9] in open chamber surgery. This adverse effect has been reported with different several dosing schedules and cumulative doses of 61-259 mg/kg [8–10].

The incidence of non-ischemic clinical seizure was 0% in the single-dose group versus 2.5% in the double-dose group (non-significant difference); one patient underwent CABG and the other open chamber surgery (valve replacement), and both had associated renal dysfunction. Although our incidence was below the limits reported by other authors, with slightly higher cumulative doses than those used by Martin K et al [9] and lower than Sander M et al [8] and Murkin JM et al [10], there seems to be a certain dose-dependent relationship with seizure.

Tranexamic acid has been shown to have an epileptogenic effect in animals [30]. The suspected mechanism is a γ-aminobutiric acid-driven inhibition of the central nervous system [31]. The patho-mechanism of the two seizures recorded in our study remains undetermined. Both occurred in patients with postoperative renal dysfunction which facilitates accumulation of the drug, and one patient underwent aortic valve surgery where micro-plaque dislodgement or the presence of micobbubles, which inevitably remain after closing the heart, could have triggered the seizure. This may also partly explain the higher incidence of seizure in open chamber surgery

Cerebrovascular events, renal pathology and myocardial infarction were similar to those reported previously[32]. Hospitalization and mortality were similar in both groups.

With 80 patients in each group, we empirically estimated a 13% difference between the incidence rates of IR, because in the balance between safety and effectiveness, we preferred safety. Although the empirical power for the observed proportions (18.8 vs. 7.5) was 67%, the study was sufficiently powered and significant differences were found at the end of the study.

Although the incidence of adverse events observed are consistent with that reported in other series, the study was underpowered to be able to affirm that the dosing schedule was safe since the sample size was calculated for IR.

Only IL-6 was measured as a biochemical marker of inflammatory response to CPB. However, this cytokine is widely used in this context, as shown in the literature [16, 17].

Blood levels of TA were not measured, but the pharmacokinetic profile of this drug has been previously studied by other authours and the dosing schedule was based on their findings [14].

Lastly, the experimental nature of this single-center study with CPB patients undergoing a particular surgical technique (with non-coated circuits), limits the external validity of our findings.

In conclusion, an additional postoperative dose of TA may reduce IR and postoperative bleeding (but not transfusion requirements) in CPB patients. A question which remains unanswered is whether the dose used was ideal in terms of safety, but in terms of effectiveness we have no doubt