Complement activation during cardiopulmonary bypass and association with clinical outcomes

Abstract In this prospective, single‐centre observational study of 30 patients undergoing cardiopulmonary bypass (CPB), the effect of unfractionated heparin (UFH), CPB surgery and protamine sulphate on complement and on post‐operative blood loss were assessed. Although C3 and C4 levels decreased significantly immediately following the administration of UFH, C3a, C5a, Bb fragment and SC5b‐9 remained unchanged. During CPB, C3 and C4 continued to fall whilst both alternative and classical pathways activation markers, Bb, C3a, C5a and SC5b‐9 increased significantly. Protamine sulphate had no effect on classical pathway components or activation markers but decreased alternative pathway activation marker Bb. Over the 12–24 h post‐surgery, both classical and alternative pathway activation markers returned to baseline, whilst C3 and C4 levels increased significantly but not to baseline values. Total drain volume 24 h after the surgery showed a moderate inverse correlation with post‐protamine C3 (r = −0.46, p = 0.01) and C4 (r = −0.57, p = 0.0009) levels, whilst a moderate positive correlation was observed with post‐protamine C3a (r = 0.46, p = 0.009), C5a (r = 0.37, p = 0.04) and SC5b‐9 (r = 0.56, p = 0.001) levels but not with Bb fragment (r = 0.25, p = 0.17). Thus, inhibition of complement activation may be a therapeutic intervention to reduce post‐operative blood in patients undergoing CPB.

produced as a result of contact activation initiated by tissue injury, has been implicated in complement activation via direct activation of C3 [10,11]. Some studies reported complement activation through the lectin pathway during CPB [12]. In a study with 185 children with congenital heart disease undergoing surgical correction with the use of CPB, changes in serum levels of mannose binding lectin and activities of MBL-MBL-associated serine protease (MASP)-1 and MBL-MASP-2 complexes were assessed immediately before, and during the surgery, throughout the first postoperative day and at discharge from the hospital. Decreases in MBL and MBL-MASP complexes were observed in all samples, correlating with a decrease in C4 and increase in C4a, confirming activation of the lectin pathway [12].
Evidence suggests pro-inflammatory cytokine release can interfere with myocardial contractility [17,18]. Furthermore, a study including 116 patients undergoing CPB showed that increased C3a levels were associated with renal and pulmonary dysfunction, abnormal bleeding and overall morbidity [15].
Several strategies have been used to reduce the impact of complement activation in patients undergoing CPB and cardiac surgery [19].
CPB circuits coated with a layer of heparin molecules have been proven to improve the biocompatibility of the bypass circuit: studies reported that C3a and SC5b-9 levels were significantly lower in patients undergoing CPB with heparin-coated circuits compared to control group, consistent with reduced activation [8]. In addition to heparin-coated circuits reducing complement activation, they have been shown to improve outcomes in terms of postoperative recovery of the patient [9]. During cardiac surgery, blood contact with the extracorporeal surface of the bypass circuit increases the risk of blood clot formation [20].
To minimise this risk, unfractionated heparin (UFH) is administered to patients during cardiac surgery [21]. At CPB termination, the effects of UFH are reversed to restore haemostasis and reduce post-operative bleeding. This is achieved by the administration of protamine sulphate, a polycationic alkaline molecule that binds to UFH producing a biologically inactive dense precipitate. Previous findings indicate that classical complement pathway activation is specifically induced by the heparinprotamine complex formation during CPB [21][22][23][24]. As heparin has antiinflammatory and anti-complement effects [25], the use of high dose UFH during CPB may reduce the complement activation in a dosedependent manner.
In this study we assessed the effect of UFH, CPB surgery and protamine sulphate on complement, complement activation markers and the effect of complement activation on post-operative blood loss and clinical outcomes.

Study design and participants
This was a prospective single-centre pilot observational study conducted at a tertiary referral centre for cardiothoracic surgery in the All patients had elective surgery. Patients with baseline platelet count of <100 10 9 /l, or platelet dysfunction, receiving heparin or antiplatelet treatment prior to surgery or patients with concomitant aortic stenosis were excluded. All participants provided written informed consent prior to surgery.

Anti-Xa assay
A chromogenic liquid anti-Xa assay (Werfen, Warrington, Cheshire, UK) was performed for quantitative determination of UFH activity in patient's plasma using an ACL TOP 500 machine (Werfen, Warrington, Cheshire, UK).

C3 and C4 level measurement
Plasma C3 and C4 were measured using an immunoturbidimetric technique on the AU680 analyser (Beckman Coulter, High Wycombe, UK).
The reagent containing latex-enhanced antibodies to C3 or C4 forms immune complexes with C3 or C4 in the plasma, which scatter light in proportion to their size, shape and concentration. The measurement of the decrease in light transmitted through particles suspended in solu-tion, as a result of complex formation, is proportional to the amount of analyte present in the sample.

Statistical analysis
Data are presented as percentages for categorical data, median and ranges for continuous data. Difference between each study time point was assessed using multiple comparison of two-way Analysis of Variance (ANOVA) (Bonferroni and Sidak multiple comparisons). Correlation between the complement markers and heparin anti-Xa levels and the duration of surgery was assessed using Spearman correlation. Analyses were performed using GraphPad Prism version 9.3.1 (GraphPad Software, Inc. La Jolla, USA). Two-tailed p < 0.05 were considered statistically significant.

RESULTS
Median age of the study participants was 62 years, and 56.7% (17/30) were male. Clinical and laboratory characteristics of 30 patients included in the study are presented in Table 2. Cell-saver was used for all patients, and the median volume of cell salvage was 2211 ml (range 1126-6255 ml). Median volume of cells reinfused was 650 ml (264-1500 ml). None of the patients had autologous blood harvest.

Changes in plasma C3 and C4 levels during CPB
Administration of UFH caused a significant fall in C3 plasma levels  Abbreviations: APTT, activated partial thromboplastin time; CPB, cardiopulmonary bypass; PT, prothrombin time.
C4 levels was noted following administration of UFH. C4 levels continued to decrease (p = 0.003) during the surgery, and protamine sulphate had no effect on C4 levels. As with C3, C4 levels gradually increased over the next 12-24 h but remained significantly lower than baseline level ( Figure 1B). Changes of C3 and C4 levels over time from baseline (pre-heparin) to 24 h post-CPB are shown in Figure 2A,B, respectively.

Changes in C3a and C5a levels during CPB
In contrast to C3 and C4, administration of UFH did not change C3a lev-  Figure 2C,D respectively.

Changes in Bb during CPB
The alternative pathway activation marker Bb also followed a similar

Changes in sC5b-9 during CPB
The sC5b-9 levels followed a similar pattern to the anaphylatoxins

Correlation between complement components, activation markers and heparin anti-Xa levels and the duration of the CPB prior to administration of protamine sulphate
As heparin has anti-complement effect, and longer duration of CPB would be expected to cause higher levels of complement activation,

Effects of complement components and activation markers on clinical outcomes and use of blood products
Complement activation and the resulting inflammatory response may be an important mechanism for multisystem organ injury in patients undergoing CPB [19]. As complement and coagulation systems are closely linked, the correlation between complement components or activation markers with 24 h blood loss was assessed.  Figure 1. There were no outliers, and as the number of patients who received blood products was small, no further analysis was performed.

Assessment of plasma factor dilution
Throughout bypass surgery, blood is considerably diluted which may influence complement levels. Plasma dilution was assessed using total protein/albumin ratio measured in samples taken at the same time points for complement assessment ( Figure 5). As expected, there was a significant dilution in plasma pre-and post-heparin (p = 0.0001), and

DISCUSSION
In this prospective single-centre pilot observational study with patients undergoing CPB for correction of ASD or tissue mitral valve replacement, there was no difference in the classical or alternative pathways activation markers immediately following the administration of UFH but a significant rise in these markers (C3a, C5a sC5b-9 and Bb fragment) during CPB, with gradual reduction in the level over 12-24 h post-CPB. However, sC5b-9 remained unchanged following protamine sulphate, although C3a and C5a levels continued to rise following the administration of protamine sulphate, whilst Bb fragment levels fell.
There are substantial data on the changes in complement activation during cardiac surgery, including CPB and extracorporeal membrane oxygenation. However, these studies have largely focused on the effect of complement activation in cardiac events such as myocardial infarction or myocardial contractility and mortality following the cardiac surgery [26][27][28]. Only a few studies have specifically assessed the relationship with post-operative bleeding [29,30] and, even then, those studies have considered the role of specific complement components rather than the overall assessment of the complement system provided in this report. This study provides prospective and systematic assess- although we did not measure split of C4. However,there was no rise in C3a. Although these results may suggest that the fall in C3 and C4 levels is due to formation of complexes with heparin rather than activation, it is possible that the fall in C3 and C4 levels could be due to haemodilution and/or contact with the oxygenator and the circuit rather than due to any effect from heparin. However, results were corrected for dilution using total protein/albumin ratio as a plasma dilution factor. Additionally, immunological changes associated with CPB were already in process which could contribute to the observed lack of difference in the levels of complement activation markers following UFH (counterbalancing effects of the immunological activation of complement and anti-complement effect of UFH). Therefore, it was practically impossible to isolate the effects on C3 and C4 levels from individual components.
Due to the anti-complement effect of heparin, a negative correlation coefficient was expected when assessing the relationship between complement activation marker levels and heparin anti-Xa levels. However, no significant correlation was observed between any of the complement markers and heparin anti-Xa levels ( Figure 3). It is therefore difficult to determine whether UFH has an effect independent of the effect of CPB, and these results do not reject the possibility of a relationship between complement activation marker levels and heparin anti-Xa level which may be overshadowed by the surgery and CPB effect. As expected, there was a significant negative correlation between C3 level and C3a levels in samples taken prior to the administration of protamine sulphate. We did not observe correlation between C3 and C4 levels with other complement activation markers or especially a correlation between C3a and C5a or sC5b-9 or Bb fragment.
The observed lack of correlation between these markers could be due to variation in the complex interaction between CPB and heparin in individual patients. Interestingly, total drain volume 24 h after the surgery showed an inverse correlation with post-protamine C3 and C4 levels, whilst a moderate positive correlation was observed with postprotamine C3a, C5a and SC5b-9 levels ( Figure 4). There was no correlation between blood loss and Bb fragment levels. This is in keeping with previous studies [31] and may be clinically relevant. However, we did not observe organ dysfunction such as renal failure or pulmonary dysfunction requiring ventilatory support in this cohort of patients.
Overall, due to the complex interactions of patient factors, surgical factors and the coagulation and complement activation, it is challenging to interpret these findings.
The complement and coagulation systems are closely linked. Thrombin can activate complement C3 and C5 independently [32]. Additionally, coagulation factors FIXa, FXa, FXIa and plasmin can generate C3a and C5a with cleavage of C3 and C5 independent of the known complement activation pathways [11]. Furthermore, factor XIIa has been shown to activate the classical complement pathway via activation of C1qrs complex [33], which may be highly relevant in patients undergoing CPB due to contact activation by the circuit. Therefore, inhibition of complement activation may be a therapeutic intervention to reduce post-operative blood loss in patients undergoing CPB. Since all patients received tranexamic acid as a haemostatic agent, individual variation in the post-operative chest tube is not attributable to use of tranexamic acid. Additionally, all patients were managed using the same institutional haemostatic management protocol.
Anaphylatoxins C3a and C5a were both significantly elevated during CPB. This corroborates previous findings by Chenoweth and associates who showed C3a generation 10 min post-initiation of CPB surgery [7,34]. The data suggest the generation of anaphylatoxins may be dependent in contact of the blood with foreign surfaces of the bypass machine during cardiac surgery. Additionally, it is important to highlight that our circuits were softline coated instead of heparin. C3a, generated by C3, is activated via any one of the three complement activation pathways. Therefore, no conclusions can be drawn concerning specific pathway activation. As expected, Bb fragment levels increased ( Figure 1E) during the CPB suggesting alternative pathway is also acti- In the present study, although the magnitude of change is very small, there was a significant increase in C3a levels observed 30 min after the administration of protamine sulphate ( Figures 1C and 2C). This result is in keeping with previous findings indicating that the complement system is particularly induced by heparin-protamine complexes formed following protamine sulphate administration [25]. This activation has been closely linked with the classical immune cascade [29,33]. Specifically, in vitro data have shown that the combination of heparin with protamine sulphate elevates C3a and C4a levels, indicating classical pathway activation [21]. C3 and C4 levels did not change following protamine sulphate administration ( Figure 1A,B). This observation is consistent with limited data suggesting that slow IV protamine sulphate administration does not result in additional consumption of C3 and C4 despite the classical pathway being activated [5,35].
Anaphylatoxins (C3a, C5a) and terminal complement complex levels (SC5b-9) levels remained unchanged 30 min post-administration of UFH but increased during surgery ( Figure 1C,D,F). However, the level of Bb fragment significantly reduced 30 min post-heparin administration ( Figure 1E), a probable indication of heparin's anti-complement action. Before protamine sulphate administration Bb levels increased, reaching its peak concentrations and 15 min after the administration of protamine sulphate levels fell significantly ( Figure 1E). This fall in Bb is likely due to the loss of CPB effect (via the alternative pathway) rather than an effect of the protamine sulphate forming complexes with the heparin and activating the classical pathway. This would be consistent with the extremely short half-life of heparin-protamine sulphate complexes. Due to complex interaction of UFH, which may reduce the complement activation and surgical stimuli that increase the complement activation, it is difficult to differentiate the action of these two individual components separately during CPB.
Throughout bypass surgery, blood is considerably diluted, which may influence complement marker levels, although results were corrected for dilution factor using the total protein/albumin ratio. This is considered a limitation in this study and needs to be considered when interpreting the results. This is reflected in the significant increase in the total protein/albumin ratio pre-and post-heparin sampling, which did return to baseline (pre-heparin) 24 h post-CPB. However, there were no differences in the total protein/albumin ratio sample points between post-heparin and pre-protamine sulphate, pre-protamine sulphate and post-protamine sulphate, post-protamine sulphate and 12 h post-CPB or 12 h post-CPB and 24 h post-CPB ( Figure 4). Additionally, sample size is relatively small, and the study group is confined to very specific cardiac surgeries (ASD and mitral replacement). There are several studies assessing the use of pexelizumab a humanized, monoclonal, single-chain antibody fragment that inhibits C5 in various cardiac interventions. Only two assessed the effect of pexelizumab on blood loss following coronary artery bypass graft (CABG) [29,30], and these demon-