Complement Activation in Coronary Artery Bypass Grafting Patients Without Cardiopulmonary Bypass^*

Materials and Methods

Patients

Twenty-two patients with isolated stenosis of the left anterior descending (LAD) coronary artery were prospectively studied. Patients were divided into two groups according to the surgical approach: a small incision via an anterolateral thoracotomy representing a minimized tissue injury (lateral group, n = 8) and a median sternotomy similar to conventional coronary surgery representing a large tissue injury (median group, n = 14). Informed consent was obtained from all patients and the investigation was approved by the local ethics committee. Inclusion criteria were the presence of type B2 and C stenosis of the LAD coronary artery with documented myocardial ischemia and a normal or moderately depressed ventricular function with the ejection fraction > 35%. Exclusion criteria were the presence of any other associated cardiac diseases, such as left ventricular aneurysm and valvular disease, myocardial infarction within 2 weeks, and IV heparin infusion before operation. The choice of incision was made by patients following a general consensus of feasibility made by a panel of cardiologists and cardiac surgeons. Patients’ demographic information, including age and sex, and preoperative conditions are listed in Table 1 .

Anesthesia was induced and maintained by IV infusion of sufentanil citrate (1 to 3 μg/kg) and midazolam (0.05 to 0.1 mg/kg). Muscle relaxation was achieved by pancuronium bromide (100 to 140 μg/kg). Anticoagulation during surgery was achieved by IV administration of 100 IU/kg bovine lung heparin to increase the activated clotting time to twice the baseline value. None of the patients received dexamethasone during surgery, and heparin was not neutralized by protamine.

Small Anterolateral Thoracotomy

The lateral thoracotomy was performed with a skin incision of approximately 8 to 10 cm made at the level of the fifth intercostal space, with the medial edge of the incision 3 to 4 cm lateral to the left internal mammary artery (LIMA). A wound spreader (IMA Retractor; CardioThoracic Systems; Portola Valley, CA) was secured in place and the LIMA was harvested as a pedicle from the first rib down to the seventh rib space under direct view. After the LIMA had been harvested, a coronary artery stabilizer (Stabilizer; CardioThoracic Systems) was placed. The LAD coronary artery was surrounded by two looping 5–0 polypropylene sutures, proximal and distal to the chosen site, for the anastomosis. The mammary-to-coronary anastomosis was performed with a running 7–0 or 8–0 polypropylene suture. After the anastomosis was completed, the two looping sutures were cut and the mammary pedicle was secured in place by means of two 5–0 polypropylene stitches. The small thoracotomy wound was closed in layers and one pleural drain was left in place.

Median Sternotomy

For median sternotomy, patients were prepared and draped as for a conventional coronary surgery down to the transverse umbilical line. Below this line, patients were covered with a warm air-cushion rewarming blanket in order to avoid heat dispersion and consequent hypothermia during the procedure. The skin incision was begun 2 to 3 cm caudal to the sternal notch and was extended to a point 2 to 3 cm cranially to the xyphoid. A full midline sternotomy was performed. The LIMA was harvested as a pedicle with electrocautery from its origin to its distal bifurcation. Next, a sternal spreader carrying a coronary stabilizer (MV Access Platform; CardioThoracic Systems) was positioned in the sternal wound. The preparation of LAD and the mammary-to-coronary anastomoses were similar to the techniques described above. After anastomoses were completed, the median sternotomy was closed in layers and three drains were left in place, one pleural, one pericardial, and one mediastinal.

Laboratory Measurements

Blood samples were taken after anesthesia before surgical incision as a baseline (T1), after surgical incision but before anastomoses (T2), at the end of operation (T3), and 24 h after operation from the indwelling radial artery catheter of the patient (T4). Immediately after sampling, 2 mL of whole blood anticoagulated with 0.1 mM ethylenediaminetetraacetic acid was used for the measurement of complement activation. Another 2 mL of whole blood anticoagulated with 3.06% sodium citrate was prepared for determination of leukocyte β-glucuronidase and IL-6. Plasma was obtained by centrifugation of whole blood at 1,100 g for 10 min. After centrifugation, plasma was divided into aliquots and stored at −80°C for further analysis. From the ethylenediaminetetraacetic acid–anticoagulated blood sample, leukocyte and hematocrit were measured by an automatic cell counter (Cell-Dyn 610; Sequoia-Turner Corp, Abbott Diagnostics; Santa Clara, CA). Complement C3a and C5b-9 were determined by enzyme-linked immunosorbent assays (Quidel Corp; San Diego, CA), whereas C4a was determined by radioimmunoassay (Amersham International Ltd; Bucks, UK). β-Glucuronidase was assayed by conversion of the β-glucuronidase subtract (Merck, E; Darmstadt, Germany) and read at 410 nm by a spectrophotometer (3550-UV; Bio-Rad Laboratories, Life Science Group; Hercules, CA). IL-6 was assessed with an enzyme-linked immunosorbent assay (Amersham International Ltd).

Data Analysis

All data are expressed as mean ± SEM except where otherwise stated. Data processing and statistical testing were performed with computer software (SigmaStat; SPSS Inc; Chicago, IL).Student’s t test was used to determine the differences between the two groups and the paired t test for within-group differences. A p value of < 0.05 was considered statistically significant.

Results

Complement Activation

Complement C3a concentration increased in almost all patients right after surgical incision but before the start of coronary anastomoses in both the anterolateral group (from 175 ± 15 to 295 ± 43 ng/mL; p < 0.01) and the median sternotomy group (from 227 ± 33 to 368 ± 53 ng/mL; p < 0.05). By the end of operation, C3a concentration stabilized in the lateral group (305 ± 71 ng/mL) but continued to rise in the median group (512 ± 87 ng/mL; p < 0.01; Fig 1 ). C4a concentration decreased after surgical incision from 4,543 ± 319 to 2,900 ± 350 ng/mL in the lateral group (p < 0.01), and from 4,934 ± 206 to 3194 ± 299 ng/mL in the median group (p < 0.01; Fig 2 ). The terminal complement complex C5b-9 concentration remained stable throughout the whole surgical period in patients with the lateral surgical incision (from 60 ± 15 to 69 ± 22 ng/mL), but increased significantly in the median sternotomy patient group right after surgical incision (from 82 ± 23 to 189 ± 67 ng/mL; p < 0.05) and by the end of operation (159 ± 49 ng/mL; p < 0.05). By group comparison, the C5b-9 concentration was significantly higher in the median sternotomy group than in the lateral thoracotomy group at the time points after surgical incision (p < 0.05) and by the end of operation (p < 0.05; Fig 3 ).

Other Inflammatory Parameters

Leukocyte count was rather stable in the lateral group toward the end of operation (from 4.7 ± 0.4 to 5.3 ± 0.6 × 10⁹/L), but increased in the median group (from 4.9 ± 0.4 to 6.6 × 10⁹/L; p < 0.01). On the first postoperative morning, the count increased to 13.1 ± 1.6 × 10⁹/L in the lateral group (p < 0.01) and to 13.3 ± 1.0 × 10⁹/L in the median group (p < 0.01; Fig 4 ). β-Glucuronidase, representing leukocyte activation, increased slightly after incision, from 155 ± 21 to 173 ± 22 arbitrary units in the lateral group and from 161 ± 23 to 198 ± 28 arbitrary units in the median group. There was no further increase during the first postoperative day (Fig 5 ). IL-6 concentration increased slightly by the end of operation (from 1.4 ± 0.2 to 4.8 ± 1.8 pg/mL in the lateral group, and from 3.0 ± 1.2 to 16.1 ± 6.1 pg/mL in the median group). On the first postoperative day, IL-6 increased significantly in both groups, but the increase was more pronounced in the median group (32.1 ± 9.3 pg/mL; p < 0.01) than in the lateral group (23.9 ± 9.8 pg/mL; p < 0.05; Fig 6 ).

Clinical Outcome

All patients were operated on uneventfully and had a normal postoperative recovery. The duration of the operation was 101 ± 30 min in the anterolateral thoracotomy group and 119 ± 32 min in the median sternotomy group. Postoperative blood loss was 244 ± 171 mL in the lateral group vs 771 ± 284 mL in the median group (p < 0.05). The mean duration of postoperative stay in the ICU was 1 day for both groups, and the mean duration of hospital stay was 4.2 ± 0.5 days for the lateral thoracotomy group vs 6.2 ± 0.3 days for the median sternotomy group (p < 0.01).

Discussion

Complement activation during CABG is known to be caused by contact of blood with the nonphysiologic material surfaces in the heart-lung machine,²³⁴ the blood-air interface during oxygenation,¹⁷ and the administration of protamine to neutralize systemic heparin.¹⁸ Virtually all these factors are related to the use of CPB. In this study, we demonstrate that complement activation was also evident in patients who underwent CABG without CPB. Complement split product C3a increased in almost all patients following either an anterolateral thoracotomy or a median sternotomy. However, C4a did not increase during surgery, suggesting that the activation is mainly through the alternative pathway and not through the classic pathway. The terminal complement complex C5b-9 was not increased in patients who underwent lateral thoracotomy; C5b-9 was increased only in patients who underwent median sternotomy, which involved a relatively larger tissue injury. Consistently, IL-6 was significantly higher in the median sternotomy group than the lateral thoracotomy group at several time points during and after surgery, suggesting that patients with a larger tissue injury also had a higher systemic inflammatory response to operation during the early postoperative period.

The role of tissue injury in complement activation is rarely described in literature.¹⁰¹¹¹² In vitro, injured tissue derived from trauma patients has been found to activate complement similar to the effect of endotoxin.¹¹ In general surgery patients, 20 to 30% of complement protein was consumed within the first hour after skin incision.¹⁰ Recent work on a molecular level indicated that transcription of most complement genes was accelerated following tissue injury.¹⁹ In the field of cardiovascular surgery, Fosse and coworkers¹² have demonstrated that complement activation was much more pronounced in a group of patients operated on with CPB than in a group in whom CPB was not used. However, they were puzzled by the finding that complement activation was significantly higher in another aneurysmectomy group without CPB, and stated that factors other than CPB may affect complement activation during major cardiovascular operations without CPB. Two factors, blood transfusion and the use of synthetic material, were considered to be attributed to the increased complement activation in their aneurysmectomy group. In our study, however, none of the patients received donor blood during operation, nor was there a difference between the two groups in the use of synthetic materials for catheterization. Thus, we assume that it was the level of tissue injury, caused by the surgical incision, that resulted in the difference in complement activation. This assumption is further supported by a very recent report²⁰ that systemic inflammation was similarly stimulated in patients with or without CPB (both groups undergoing median sternotomy), implying that the inflammation is predominantly caused by the surgical procedure per se.

The mechanism by which tissue injury induces complement activation remains unclear. Tissue factor is known to be released by injured endothelial cells, but its main role is to initiate the extrinsic clotting cascade.²¹ Proinflammatory cytokine IL-6 released during acute phase response following tissue injury may activate complement,²² but it may not be a cause of complement activation in the present study because plasma concentration of IL-6 was elevated at the end of operation. Tissue-type plasminogen activator (t-PA) is a release product from the injured endothelial cell and is known to activate complement.²³ A recent observation indicated that pharmacologic activation of plasminogen by t-PA, a thrombolytic intervention for treating acute myocardial infarction, may directly activate complement, leading to the increased level of C3a.²⁴ In cardiac surgical patients, higher t-PA levels have been found in blood samples collected from the surgical wound than in blood samples obtained from the systemic circulation.²⁵ Also, C3a concentrations in wound blood are higher than in circulating blood in cardiac surgical patients.²⁶ Furthermore, increased t-PA has been observed in the systemic circulation in other types of thoracic surgical patients without CPB.²⁷ However, the exact mechanism of t-PA–mediated complement activation needs to be further elucidated.

The pathways of complement activation caused by conventional coronary surgery with CPB have been well described.¹²³⁴¹⁸ From the majority of observations, C3a increased immediately after the start of CPB without the concomitant increase of C4a, indicating that it is the alternative complement pathway that is activated by blood-material interaction. Conversely, the classic complement pathway is described to be activated by heparin-protamine complex manifested after the termination of CPB by an increase of C4a accompanying the rise of C3a.¹⁸ In the present study, however, an increase in C3a alone was found after surgical incision and during operation, with no concomitant increase in C4a; this suggests that complement activation by tissue injury is mainly through the alternative pathway. These results are consistent with the earlier observation that the surgical procedures affected only C3 but not C4.¹⁰ The common pathway involves complement proteins C5 to C9 and clinically to be determined as the terminal complement complex C5b-9.²⁸ Although increased C3a was observed in both patient groups in the present study, C5b-9 increased only in the median sternotomy group. Furthermore, the pronounced complement activation observed in the median sternotomy group was also accompanied by an increased production of IL-6, suggesting a stronger inflammatory response in these patients as a result of greater surgical trauma.

We conclude that complement activation occurs in coronary surgical patients operated on without CPB. Tissue injury caused by surgical trauma is involved in inducing this CPB-independent complement activation. Thus, clinical consideration of complement inhibition during heart operations should include the contribution of tissue injury in addition to blood-material interaction in the heart-lung machine. With regard to the surgical incision, a small anterolateral thoracotomy is associated with less complement activation and a reduced postoperative inflammatory response compared with a conventional median sternotomy. However, the benefit-risk balance of a small anterolateral thoracotomy and the choice of surgical incision for coronary surgery without CPB remain to be specified²⁹³⁰³¹ after more follow-up of patient outcomes with regard to the surgical approach.