Hemolysis in Cardiac Surgery Patients Undergoing Cardiopulmonary Bypass: A Review in Search of a Treatment Algorithm

RED BLOOD CELLS

There are some interesting truths about RBCs that clarify the clinical complications that are the consequence of RBC damage.

The first is the fact that the mature RBC does not contain a nucleus, which makes them easier to deform and allows them to enter the smallest capillaries. Being anucleate also allows the RBC more space to carry hemoglobin and makes them lighter, therefore reducing the workload of the heart by 15% (4). Enucleation also gives these RBCs their interesting biconcave shape, which enables them to have maximal contact with the surrounding tissues, facilitating the exchange of O₂ and CO₂ (4).

A second fact is that the two major characteristics of RBCs, deformability and aggregability, greatly contribute to the viscosity of the blood (5). Thus, the visco-elastic profile of blood is not only influenced by plasma viscosity, which is mainly a function of the concentration of macromolecules, temperature, and RBC count, but additionally, a modest decrease in RBC deformability results in a significant increased viscosity. Consequently, there is a significant higher blood pressure required for these rigid RBCs to enter the microcirculation (6).

A third phenomenon is that the major RBC responsibility, which is transportation of O₂ and CO₂ between the lung and tissues, is not a function of the RBC itself but of its main constituent: hemoglobin (Hb).

Approximately 97% of a dry RBC consists of hemoglobin, which is an assembly of four protein subunits, each of them containing one heme group with one iron molecule, and each of the heme groups being capable of binding one molecule of oxygen.

In addition to the isolation of hemoglobin from the plasma by the red blood cell membrane, there is also an efficient mechanism ready to clear the free hemoglobin from the plasma upon hemolysis. This safe encapsulation suggests the toxicity of this hemoglobin molecule once freed in the plasma.

LETHALLY DAMAGED RBCs AND THE CLINICAL CONSEQUENCES

Upon destruction of the RBC membrane, PfHb and heme enter the circulation. PfHb is normally cleared by the Hb scavengers, haptoglobin and CD163. Heme, the pro-oxidant and pro-inflammatory oxygen-binding component of Hb, is transported to the liver by hemopexin and degraded through heme oxygenase into carbon monoxide, biliverdin, and iron. These products, together with the interleukin (IL)-10 release induced by the Hb clearing mechanism, exert an antioxidant, anticoagulant, and vasodilating effect in the circulation, thus compensating for the adverse effects caused by PfHb and heme.

An increasing level of PfHb circulating in the plasma always indicates hemolysis, whereas normal values are not indicative of a healthy RBC.

First, the clearing mechanisms of hemolysis-derived end products only become saturated and exhausted in case of excessive hemolysis. Second, haptoglobin, essential in scavenging PfHb, exists in three different phenotypes (Hp 1-1, Hp 1-2, and Hp 2-2), all possessing varying antioxidant capacities and varying abilities to link up with Hg, which result in different levels of circulating PfHb (7). The Hp 1-1 allele has a much stronger antioxidant and PfHb scavenging capacity than the Hp 1-2 and Hp 2-2 phenotype. The distribution of these phenotypes show geographical differences, with whites showing the highest prevalence of the Hp 2-1 and Hp 2-2 type (8). The Hp 2-2 phenotype is associated with a significantly higher degree of atherosclerotic disease and higher percentage of diabetes mellitus (9). The prevalence of the Hp 1-1 phenotype is considerably increased in the elderly, indicating a higher life expectancy for the Hp 1-1 allele carrier (10). Consequently, one can assume the highest prevalence of Hp2-2 in our adult cardiac surgery population, resulting in the lowest protection from increasing PfHb levels and the lowest antioxidant capacity.

Because haptoglobin is synthesized in the liver, patients with impaired liver function show a decreased haptoglobin production. However, low levels in these patients should not be indicative of hemolysis but a decrease in the protection from PfHb in case RBCs do lyse (11). On the other hand, haptoglobin acts as an acute-phase protein, with its plasma concentration increasing in response to a variety of stimuli, such as infection, acute myocardial infarction (AMI), and inflammatory reaction. Therefore, monitoring of hemopexin, which is a non-acute phase protein, should be preferred over haptoglobin (10).

When the capacity of the intravascular scavenging mechanisms has been saturated, levels of Hb and heme increase in plasma and urine. These are associated with adverse clinical signs and symptoms, most often not recognized as the underlying phenomenon of hemolysis.

Clinical consequences of excessive PfHb levels include dystonias involving the gastrointestinal, cardiovascular, pulmonary, and urogenital systems, as well as clotting disorders (12). It was reported that PfHb directly impairs left ventricular function and coronary blood flow in neonatal rabbit hearts, especially when ischemia and reperfusion were involved (13).

Many of the clinical sequela of intravascular hemolysis are readily explained by Hb-mediated NO scavenging. On saturation of free Hb scavengers, PfHb binds NO derived from the endothelium in a fast and irreversible way. NO plays a major role in vascular homeostasis and has been shown to be a critical regulator of smooth muscle tone and platelet activation and aggregation. It has been shown that quantities of plasma Hb > 10 mg/dL can potently inhibit NO vasodilatation in vivo (14).

NO depletion results in elevated systemic vascular resistance (SVR); elevated pulmonary vascular resistance (PVR); increased thrombin formation, fibrin deposition, platelet activation, and aggregation; organ dysfunction; and is accompanied by an elevated mortality rate (12).

Iron, freed on hemolysis, is transported to the body’s iron storage proteins (ferritin and hemosiderin) by irontransporting proteins (transferrin and lactoferrin). The iron-binding protein transferrin is normally ˜30% loaded with iron, retaining a considerable iron-binding capacity (10). During CPB, iron release is increased, partly as a result from red cell lysis, resulting in iron overload in −20% of adult patients (15) and even a higher percentage in neonates. Children >5 years of age rarely show signs of iron overload, probably because of higher plasma transferrin levels (16). Iron is a pro-oxidant, and iron overload is accompanied with acute renal failure and increased lung permeability (17).

Early postoperative hyperbilirubinemia after CPB surgery has been associated with a high mortality rate. This hyperbilirubinemia is a multifactorial process, of which the increased bilirubin production from hemolysis plays a substantial role (18).

Considering all this, one should not be worried about the relative small loss of functional RBC mass, which has no clinically significant impact on the primary function of the RBC, which is gas transport, but should be worried about the damaging effects of the RBC constituents once the body’s natural scavenging mechanism is exhausted.

SUBLETHALLY DAMAGED RBCs AND THE CLINICAL CONSEQUENCES

Beside complete cellular destruction, damage can also be induced to the RBC membrane on a sublethal level, resulting in decreased deformability and increased aggregability. This phenomenon is reversible to a certain extend depending on the time and intensity of damage (19).

This sublethal damage of the RBC membrane is hard to measure, but can become clinically very important because these changes can result in the early removal of the RBC from the circulation by the spleen, or more seriously, in altered rheological properties of the blood. The decrease in cell deformability results in a reduced capacity to enter small capillaries and a reduced contact of the cell surface with the surrounding vessel wall, all compromising efficient microcirculation and oxygen delivery to the surrounding tissues (20). Rigidified RBCs have lower oxygen content than their healthy colleagues, as a result of a decreased oxygen uptake in the lungs (6). Thus, the clinical consequences of sublethal RBC damage are a function of the loss of rheologic properties of the RBC and of the potential of the organism to compensate for the increased need and loss of vasodilatation and the need for higher perfusion pressures. Sublethal RBC damage can result in end organ dysfunction caused by cellular ischemia.

RBC DAMAGE AND CPB

Data from different studies consistently show an increase in plasma Hb concentrations and a decrease in haptoglobin levels during CPB, indicating progressive destruction of RBCs (21–24).

Results of in vitro studies regarding the sublethal damage caused by CPB are sometimes inconsistent or even contradictory (25–27). The major problem is the absence of standard methods of measurement of RBC deformability or other parameters characterizing sublethal blood trauma. Nevertheless, a recent study did indicate that RBCs exposed to certain uniform shear stresses for prolonged time (>20 minutes) show statistically significant changes in their deformability without lysis (25).

One study showed delayed hemolysis after CPB: ˜15% of the RBCs subjected to the extracorporeal circulation were irreversibly damaged and removed during the 24-hour post-infusion period (28).

Considering the clinical consequences of both RBC destruction and decreased RBC deformability, it is obvious that all efforts should be made to minimize RBC aggression during CPB.

The first step in the reduction of hemolysis during CPB is the identification of all factors responsible for RBC damage. Second, a maximal hemo-compatible system should be composed and customized for each patient, and CPB management must be optimized toward gentle blood handling. Third, high-risk groups, such as small patients and complex surgery cases, need to be identified and paid special attention to minimize hemolysis-related morbidity.

The mechanisms of mechanical cell damage during CPB have been described by different investigators (29,30), identifying positive pressure, wall impact forces, blood non-endothelial surfaces, negative pressure, blood-air interface, and shear stress as responsible forces for blood trauma. It seems that RBCs have a high tolerance to most forces, except for shear stress (30). Physiologic values of shear stress range from 1 to 50 dynes/cm², sublethal damages are reported to appear from 210 to 430 dynes/cm² onward, lethal damage to RBCs occur from 1500 dynes/cm² onward, and the threshold for platelets and leukocyte activation is, respectively, ˜100 and 75 dynes/cm² (22,31,32).

Beside shear stress, time of exposure seems to be another impacting factor related to cellular damage (31).

RBC DAMAGE AND CPB: ISOLATED CPB COMPONENTS

The magnitude and duration of shear stress will be different in the various components of the CPB circuit but will always be present to some extent. A high value with short duration will be found in arterial cannula, but lower magnitudes of shear stress with longer duration can be found in oxygenators and venous reservoirs.

By analyzing the cardiopulmonary system, we can locate different areas where RBC damage can occur.

RBC DAMAGE AND CPB: CPB COMPOSITION AND MANAGEMENT

More important in respect to RBC integrity is the composition of the CPB system and CPB management.

RBC DAMAGE AND CPB: ENVIRONMENTAL INFLUENCES

The primary cause for RBC damage during CPB is indeed mechanical aggression, but the susceptibility of RBCs to this stress can be influenced by other environmental factors during CPB.

First, and important in interpreting hemolysis-related studies, there is a difference in individual RBC fragility toward mechanical stress. Fetal RBCs show a significantly higher susceptibility toward mechanical stress than adult RBCs. Male RBCs are significantly more fragile than female RBCs, and human RBCs are significant more fragile than RBCs from animals. RBCs show a shear stress-induced pre-conditioning; sublethally damaged RBCs lyse at a much lower shear stress threshold than healthy RBCs (79).

In vitro data showed a significant decrease in RBC deformability and increased fragility caused by either hypothermia or plasma dilution (80).

The effect of hypothermia is more pronounced in deep than in moderate hypothermia and was found to be transient. The reasons for the effect of hypothermia on the mechanical properties of RBCs remain to be identified.

It was shown by some investigators that the dilution of plasma with phosphate-buffered saline (PBS) or dextran solutions caused a significant increase in the susceptibility of human RBCs to mechanical stress. Their results indicate that hemodilution without plasma protein replacement may therefore elevate hemolysis during extracorporeal circulation. The mechanisms of a plasma protective effect on RBC fragility is not fully understood, but it may be assumed that the components of plasma that provide protection against hemolysis give a surface coating to RBCs (81).

Recent studies showed a protective effect from mechanical damage to the RBCs through innovative pharmacologic approaches.

DEALING WITH CPB-INDUCED RBC DAMAGE

Separated suction is a major contribution toward minimizing RBC damage during CPB (65).

Autotransfusion devices remove up to 90% of plasma Hg and therefore could be used in processing damaged blood. If possible, a separate cardiotomy reservoir should be installed in all CPB cases, and damaged blood should be processed in an autotransfusion device to eliminate lysed RBCs and PfHb (87). If the blood in the separated cardiotomy reservoir is too damaged, the blood should be discarded, because it will contain sublethal damaged RBCs that will lyse later.

In case there is no separated suction and the blood is potentially too damaged because of a long and complicated run, the pump blood should preferably not be returned to the patient unless processed through an autotransfusion device (88).

Although autotransfusion devices are capable of removing most of the PfHb, this is not the case for hemofilters used perioperatively.

Various types of membranes are traditionally used for perioperative hemofiltration, all capable of removing molecules up to a molecular weight of 20,000–50,000 da. Free hemoglobin molecules have a molecular weight of ˜68,000 da, and free hemoglobin dimers are ˜32,000 da (89). This may lead to retention of these toxic molecules during classical hemofiltration. Studies have shown a contribution of hemofilters to additional hemolysis, especially in “rinsefree” hemofilters with no crystalloid prime before use (90).

To remove PfHb as efficiently as washing and centrifugation, a super high-flux 100,000-da hemofilter must be used (89).

There are some options for a pharmacologic approach in patients with hemolysis.

In case of exhausted endogenous clearing mechanisms for PfHb, additional haptoglobin could be administered. Some data have shown that haptoglobin treatment significantly decreases the levels of renal tubular enzymes, suggesting a protective effect on renal tubular function (23).

Comparing the prophylactic administration with the therapeutic administration of haptoglobin, both methods effectively prevented an increase of PfHb level, but prophylactic administration (priming administration) was safer and more useful considering PfHb level in filtered blood and changes of serum-free haptoglobin, free hemoglobin, and creatinine clearance during and after the operation (23,24,90).

A second pharmacologic approach is the administration of indirect or direct NO donors, compensating for the decrease in NO availability caused by the binding of endogenous NO with PfHb.

A study investigating sodium nitroprusside administration during the rewarming or reperfusion period showed a decreased PfHb-induced increased pulmonary (PVR) and systemic vascular resistance (SVR). Their data also showed a protective effect on renal function (91). Another study warned of the adverse effect of sodium nitroprusside administration; it may induce accelerated free cyanide release, resulting in increased risk of cyanide toxicity.

This cyanide toxicity is hard to recognize and is accompanied by hypotension, encephalopathy, and metabolic acidosis (92).

Restoration of endogenous NO depletion could also be provided through iNO administration.

iNO converts plasma-free oxyHb to plasma-free methemoglobin, which is not an effective NO scavenger, and one study showed a 100% reversal of PfHb-induced PVR and SVR (84). In contrast, the formation of methemoglobin can induce oxidative damage because of formation of other radicals (93).

Methemoglobin has a very high affinity for oxygen and can clinically be detected during extracorporeal circulation by the appearance of chocolate-colored blood.

Data from a recent NIH-supported study suggested that nitroxyl generated by Angeli salt (sodium α-oxyhyponitrite, Na₂N₂O₃) preferentially reacts with cell-free hemoglobin compared with that encapsulated in the RBCs under physiologically relevant conditions. Nitroxyl oxidizes oxygenated ferrous hemoglobin to methemoglobin and can convert the methemoglobin to a more stable, less toxic species: iron-nitrosyl hemoglobin. These results support the notion that Angeli salt or a similar compound could be used to effectively treat conditions associated with intravascular hemolysis (94).

Iron chelation therapy such as recombinant human transferring or lactoferrin for neonates or desferrioxamine for adults has proven to be beneficial to avoid iron overload (17).

A recent study suggested the prophylactic alkalinization of urine with sodium bicarbonate during CPB to protect the kidney from acute kidney injury by PfHb (95).

Sublethal blood damage resulting in decreased microcirculation is harder to detect, and thus, to treat. Innovative ways of monitoring oxygen consumption in vital organs, such as non-invasive venous oxygen saturation monitoring, might help to detect hypoperfusion of the brain and abdominal organs. Once detected, efforts could be made to restore systemic vasodilation, and blood pressure could be increased to ensure microperfusion.">Sublethal blood damage resulting in decreased microcirculation is harder to detect, and thus, to treat. Innovative ways of monitoring oxygen consumption in vital organs, such as non-invasive venous oxygen saturation monitoring, might help to detect hypoperfusion of the brain and abdominal organs. Once detected, efforts could be made to restore systemic vasodilation, and blood pressure could be increased to ensure microperfusion.

DISCUSSION

In summary, it can be said that the mechanical forces and environmental damage during standard CPB are reasonably well tolerated by RBCs.

If proper component selection is made, the little RBC damage that does occur is mostly within the clearance capacity of the reticulo-endothelial system, so that the PfHb should not increase significantly during standard CPB. On the other hand, even with state of the art equipment and proper size selection, much damage can be incurred from the way a perfusionist is managing the heart lung machine.

Additionally, there are inevitable situations of more pronounced hemolysis, such as in pediatric cases and during more extensive or prolonged surgery.

This is the case for long-term support, as in VAD or ECMO support, where isolated component selection is much more important than in regular CPB. Here, mechanical RBC damage is often encouraged by the blood pump in combination with inappropriate cannula size or design. In VAD- and ECMO-supported patients, control of bleeding and coagulation is detrimental for improved outcome, low pulmonary and systemic resistance is essential for the suffering ventricle and/or lungs to be weaned off mechanical support, and compensatory mechanisms fail to compensate for the increased viscosity because of rigidified RBCs. Thus, outcome in these patients could theoretically be substantially improved by eliminating toxic hemolysis-derived end products and rheologicaltered blood cell characteristics.

In all patients with poor outcome results, potential clinical complications of sublethal and lethal RBC damage might add to increased morbidity and mortality rate, and, even more importantly, to decreased quality of life.

In patients undergoing CPB, efforts are routinely done to minimize damaging forces, although this is mainly done to protect platelets and leukocytes from activation. Perfusionists and physicians consider the detection of free hemoglobin only as a window to a greater view; because RBCs are much more resistant to mechanical stress than platelets and white blood cells, it is assumed correctly that once RBCs are damaged, other blood elements have already been affected more severely.

To summarize, besides clinical monitoring through reddening of urine or plasma in standstill blood, no further action is undertaken during or after CPB, and nowhere in the literature is there any mention of existing written protocols or algorithms toward protecting patients from this underestimated complication.

CONCLUSION

From extensive literature, it seems that hemolysis during CPB is a well-recognized, but somewhat underestimated complication. Although efforts are made to minimize RBC damage during CPB, the potential consequences of hemolysis seem to be neglected in the perioperative management of patients undergoing CPB. Nevertheless, a growing list of clinical manifestations related to hemolysis suggests the toxicity of the proinflammatory and pro-oxidant RBC-derived components.

Therefore, and especially in those patient groups where there is still room to improve outcome, there may be a need for an algorithm in the detection, prevention, and treatment of RBC damage and related complications.