Washing and filtering of cell-salvaged blood – does it make autotransfusion safer?

HISTORY OF AUTOLOGOUS TRANSFUSION

In a brief letter to the Lancet in 1874, William Highmore, the senior physician and surgeon at Yeatman Hospital in the UK, wrote of an ‘overlooked source of blood-supply for transfusion’.¹ He described an unfortunate case of a woman who was suffering from post-partum hemorrhage. Even though he was able to stop the bleeding, she died with ‘several pounds of blood in a vessel and in the bed’, which, he speculated, could have been used to save her life had he been able to transfuse it back into her veins. At this point in history, allogeneic blood transfusion was being practiced, having been introduced some 50 years earlier by Blundell.^2–4 Highmore was only one of a handful of physicians from the same era who thought of applying the principles of allogeneic transfusion to the patient’s own blood. William Halsted, then assistant surgeon at Roosevelt Hospital in New York City, described in 1883 a case where he removed half a liter of blood from a victim of carbon monoxide poisoning, defibrinated it, strained it, and returned about a quarter of a liter to the patient who then recovered.⁵ Halsted, one of the most prominent early American surgeons who would later train Harvey Cushing, did his own historical review of the subject of autologous transfusion and ascribed the idea to a German physician, R. Volkmann, who suggested the feasibility of it during hip surgery in 1868. The first successful autotransfusion according to Halsted occurred in 1874 when another German physician, Hueter, transfused 350 mL of autologous blood into the left tibial artery of a patient with frost-gangrene of both feet. The left foot was salvaged, whereas the right was not. Further independent successful cases of autotransfusion, this time as part of amputation procedures, were performed in 1885 by Duncan and Miller, two colleagues from the Royal Infirmary in Edinburgh, Scotland.^6,7 It was not until the 20th century that the practice went from rare case reports to larger cohorts.

H. J. Thies, a German obstetrician, reintroduced autotransfusion in the medical literature with his report in 1914 describing three cases of autotransfusion in patients with ruptured ectopic pregnancies.⁸ A little more than a decade later there were at least 188 additional cases reported related to ectopic pregnancies, splenectomies, liver lacerations, hemothoraces and intracranial operations in Europe, the USA and Australia.^9–13 The only typical processing of the blood prior to retransfusion was addition of citrate and filtration through gauze. The first large case series was in 1943 from the University of Louisville School of Medicine in Kentucky by Griswold and Ortner, who reported on 100 consecutive cases of autotransfusion related to ruptured ectopic pregnancies and trauma.¹⁴

With all the progress made in blood donation and storage in the late 1930s and 1940s and the more widespread establishment of blood banks, the ease of allogeneic transfusion relegated autotransfusion to an afterthought until the late 1960s and 1970s, when Dyer, Klebanoff and Pathak made significant progress in developing techniques and devices to facilitate the transfusion of unwashed shed blood. These investigators provided much data on clinical outcomes, contaminant filtration and reducing hemolysis.^15–24 The first widely used device for intraoperative autotransfusion was the Bentley ATS-100 (Bentley Laboratories, Santa Ana, CA, USA) (Figure 1). It was basically a modified Bentley cardiotomy reservoir and consisted of an aspirating unit activated by a roller pump, a reservoir with a filter, and a delivery end connected directly to the patient or to a storage bag. The device was probably the most popular unit up until the late 1970s when it was withdrawn from the market because of litigation related to incidences of air embolism, and supplanted by devices that washed the salvaged blood.^23,25 The first prototype machine that washed the collected blood was developed at the Mayo Clinic by Wilson and Taswell in 1968 (Figure 2),²⁶ and the technology was further developed in the 1970s, with several practical devices becoming commercially available in the same decade.^27,28

Figure 1.

Schematic of the Bentley ATS-100. It was the first widely used commercial device used for autologous transfusion. The blood was aspirated into a modified Bentley cardiotomy reservoir, passed through a filter and pumped directly back into the patient. The drive unit consisted of a roller pump which controlled aspiration and transfusion. The product was pulled from the market less than a decade after being released because of incidents of air embolism. Figure reprinted from Duncan et al.,²² Annals of Surgery by American Surgical Association. Reproduced with permission of Lippincott Williams & Wilkins.

Figure 2.

Diagram of the first prototype device for autologous transfusion that also washed the blood. It was used during transurethral resection of the prostate and collected the blood (and irrigating fluid) into a reservoir and then washed it with Ringer’s lactate solution via a continuous flow centrifuge bowl. Modern cell salvage machines use the same principle to wash the blood and produce a product consisting mostly of packed red cells. Figure reprinted from Wilson and Taswell,²⁶ Mayo Clinic Proceedings by Mayo Foundation. Reproduced with permission of Dowden Health Media. RBC, red blood cell.

WASHING

From the initial idea of salvaging and readministering shed blood, the blood was readministered in an unwashed state for almost 100 years prior to a shift in practice which led to the routine intraoperative washing of this blood. With washing, blood is collected from the surgical field through a suction tip into a container and mixed with anticoagulant. The blood is centrifuged and the supernatant removed. Large volumes of saline is counterflowed through the blood and centrifuged out with the purpose being to remove any contaminants that may lead to harm of the patient. Upon readministration the red cells are suspended in saline.

Salvaged shed blood comes into contact with both the operative field and components of the salvage device, and therefore the composition is altered from its native state in circulation. Washing is performed to mitigate these changes by (i) removing thrombogenic products in the salvaged blood and reducing the risk of inducing coagulopathies; (ii) removing cell breakdown products and potentially harmful plasma proteins that cannot be removed by filtration; and (iii) removing gross chemical and physical contaminants that may have been introduced from the surgical field. The merits of each of these reasons for washing will be reviewed individually, as some controversy exists over whether or not it is beneficial to reduce salvaged whole blood to a washed suspension of red cells.

Removal of thrombogenic products and reducing risk of inducing coagulopathies

Salvaged unwashed blood is altered when compared with blood drawn from a venipuncture in terms of its coagulation components. Fibrinogen levels are decreased and fibrin degradation product (FDP) levels are markedly increased, which indicates both activation of the clotting cascade and subsequent fibrinolysis.^29,30 Platelets are decreased in number or remain unchanged; however, they are mostly degranulated and nonfunctional.^29–33

High levels of FDPs in unwashed salvaged blood leads to a transient increase in FDP levels in the patient’s blood after transfusion. Faris and Wilson measured FDP levels in patients after receiving unwashed salvaged blood and found it to be eight to 15 times higher than normal reference levels.^32,34 Kristensen, Fuller, Blevins and Krohn all found D-dimer elevations in patients transfused with unwashed blood.^29,35–37 Bell found decreased levels of fibrinogen, which is depleted as fibrin is formed.³⁸ Duchow looked at FDP levels as well as other coagulation factors and markers of thrombin generation 15 minutes after transfusion of unwashed shed blood and found reduced amounts of factor V, factor XIII, and increased amounts of thrombin-antithrombin III, FDPs and D-dimers, although all values were trending back to normal 1 day after transfusion and no clinical adverse events were seen.³⁹ The transient nature of the elevations was also seen by Thompson and Sharp, who showed that D-dimer levels were normal within 24 hours after transfusion of unwashed salvaged blood after vascular surgery, and Blevins showed D-dimer levels returning to baseline 12–18 hours after transfusion in spine and hip surgery.^29,33,40

Despite these changes in FDP levels, coagulation factors and platelet functionality, the coagulation profiles of patients are only transiently affected when receiving unwashed blood. When instead the salvaged blood is washed, most if not all FDPs are removed.^30,41,42 Kingsley showed in a primate model that FDP titers were positive even at a 1:1500 dilution in animals receiving unwashed blood, and remained undetectable in those that received washed blood.⁴² Clinically, however, in many cases there is no difference between patients who received washed versus unwashed blood in terms of their coagulation profiles. Kingsley and Griffith showedthe same changes to prothrombin time (PT), partial thromboplastin time (PTT) and fibrinogen levels after receiving washed and unwashed cells.^41,42 In both washed and unwashed salvaged blood the changes are transient and return to baseline within 1 day.^35,36

The transfusion of unwashed blood therefore leads to alterations in hematologic laboratory values, most notably an increase in FDPs, with some of these alterations mitigated or eliminated by washing the blood first. High levels of FDPs may theoretically affect the clotting mechanism by competitively binding to fibrinogen receptors on platelets, thus inhibiting platelet aggregation.⁴³ This is supported by several studies that have shown reduced platelet aggregation in samples of unwashed salvaged blood.^30,31 The only adverse events related to coagulopathy reported after autotransfusion of unwashed blood are reports of disseminated intravascular coagulopathy (DIC).⁴⁴ However, as the diagnosis of DIC depends largely on laboratory parameters, the transfusion of unwashed salvaged blood falsely indicates a coagulopathic state because of elevated FDPs due to the transfusion, and not because of active DIC.^{29,37,40–42} In conclusion, there does not appear to be any clinical significance to these changes.

It is worth noting that these aforementioned studies involve autotransfusion of only a fraction of the total blood volume of the patient. Studies with large-volume transfusions are limited to one in vascular surgery and one animal study. In Sharp’s study of patients undergoing vascular surgery, all of whom receiving at least 3.5 L of salvaged washed blood, they had a 5.1% incidence of what they refer to as ‘coagulopathy’, but no laboratory parameters or further explanation is given.³³ In an animal study involving the transfusion of twice the estimated total blood volume there were significant changes in prolonged PT and PTT, as well as decreased platelet counts. Interestingly, bleeding times returned to normal within 24 hours, but platelets showed a decreased or absent aggregation response to ADP and collagen, and continued to decline in number for at least 20 hours after the transfusion was completed.³¹ This suggests there may be a prolonged effect on platelet functionality. This, together with an essentially progressive dilutional coagulopathy with large-volume transfusions as more and more platelets are inactivated, makes transfusion of packed platelets appropriate in settings of large-volume autotransfusion, but the threshold at which this becomes necessary is not clear.

Removal of cell breakdown products and activated plasma protein products that cannot be removed by filtration

The process of cell salvage damages the cellular components of blood, most notably by shear stress from the suction device.^45–47 This leads to the production of cell breakdown products in the salvaged blood product, including cell fragments, membrane debris and intracellular organelles and enzymes. Once accumulated and retransfused back to the patient, these may activate the coagulation, complement or kinin cascade, or produce end-organ damage such as acute kidney injury. In general, suction vacuum should be limited to 100 mmHg to reduce cell breakdown products.⁴³ The significance of free hemoglobin, leukocyte enzymes, activated complement proteins and inflammatory cytokines in washed and unwashed salvaged blood will be reviewed here.

Free hemoglobin

When red blood cells (RBCs) lyse, they release large quantities of free hemoglobin, which is scavenged by haptoglobin, producing hemoglobin–haptoglobin complexes, which are safely eliminated from plasma via the reticuloendothelial system. Although it is patient-specific, generally at a free hemoglobin concentration around 100 mg/mL, the haptoglobin stores are overwhelmed and free hemoglobin accumulates.⁴⁸ Free hemoglobin is nephrotoxic, causing an acute decrease in urine output and creatinine clearance.⁴⁹ Free hemoglobin is not cross-linked and is rapidly filtered into the renal tubules, where it precipitates and produces a pigmented nephropathy.⁴³ Free hemoglobin also scavenges nitric oxide, which in turn induces contraction of smooth muscle cells. The subsequent vasoconstriction may cause pulmonary hypertension and renal ischemia, as well as lead to gastrointestinal dystonias and pain.⁵⁰ Nitric oxide scavenging also seems to be responsible for platelet activation and thrombosis seen with excessive free hemoglobin.⁵¹

Plasma free hemoglobin is notably elevated in salvaged unwashed blood. Sharp measured average levels of 65 mg/mL for aneurysm cases, 58 mg/mL for other non-emergent vascular cases and 78 mg/mL for emergent vascular cases.³³ When unwashed salvaged blood is transfused back to the patient, it causes plasma free hemoglobin elevations and changes in haptoglobin levels. Haptoglobin levels are lower within 24 hours after transfusion, but recover by day 3, which is similar to what is seen in patients transfused with banked homologous blood.³⁵ Washing is efficient at reducing the free hemoglobin content. Rates of clearance have been reported to be between 53% and 99%, with the concentration of free hemoglobin in the washed product ranging from 1.0 to 30.4 mg/mL.^52–56

As with changes to coagulation parameters, there is no definitive evidence that changes in free hemoglobin lead to clinical adverse events. Klodell states a 12% incidence of postoperative renal dysfunction in a series of 125 patients undergoing vascular surgery procedures, with an association between amount of free hemoglobin and subsequent renal dysfunction.⁵³ However, causality cannot be clearly established as free hemoglobin increases in proportion to the amount of blood salvaged, and increased requirements for blood in vascular surgery are itself a risk for kidney injury due to ischemia during surgery. They suggest limiting the amount of transfused salvaged blood to five units to avoid renal dysfunction, but the evidence base for this is weak. Others have shown that the effect on creatinine due to salvaged unwashed blood is similar to the effect seen after transfusion with homologous banked blood.³⁵

Leukocyte enzymes

White blood cells (WBCs) contain intracellular enzymes including elastase, lysozyme and cathepsin G, which may be toxic upon reinfusion.⁴³ The salvage process damages WBCs and these enzymes accumulate, with elastase levels as much as 10-fold higher than reference values.^30,57 Theoretically, recipient alpha 1-antitrypsin may be overwhelmed by quantities of elastase in lysed leukocytes in salvaged blood and cause damage to tissues, especially the pulmonary bed;⁵⁸ however, no clinical adverse events have been attributed to this. Washing has been shown to remove all elastase and 86% of lysozyme.^30,57,59

Activated complement proteins

The complement system, once activated, may cause vasodilation, white cell activation and aggregation, and direct tissue damage.³⁰ Contact with artificial surfaces is known to activate the complement system.⁴³ Salvaged unwashed blood contains activated complement proteins, including C3a, C5a, and terminal complement complexes.^{29,30,60–62} These proteins are successfully removed by washing.^30,63 When complement levels were measured in the patients’ blood after transfusion of salvaged unwashed blood, they either were normal, or returned to normal within 12–18 hours after transfusion, suggesting that the rise was due solely to the infused proteins and not as an indicator of in vivo complement activation.^29,60,61 Levels of complement split products remained decreased a little longer, returning to baseline within 3 days of transfusion.³⁵ The clinical effects of retransfusion of activated complement proteins are difficult to tease out among adverse events reported after autotransfusion; however, it is reasonable to assume that it contributes to febrile reactions. It may also contribute to the transient hypotension observed while receiving salvaged blood in some patients.⁶⁴ There has been one case report of acute airway edema during transfusion of unwashed salvaged blood collected from a wound drain after revision total hip arthroplasty, which was attributed to increased complement levels in the salvaged blood.⁶⁵

Inflammatory cytokines

Surgery itself causes both a local and systemic increase in inflammatory mediators.⁶⁶ It is not surprising, therefore, that salvaged unwashed blood contains inflammatory mediators: both proinflammatory cytokines like interleukins (IL) 1β, IL-6, IL-8 and tumor necrosis factor-α, and anti-inflammatory cytokines like IL-4 and IL-10.^62,66–70 Studies have shown that once unwashed salvaged blood is transfused, it causes an immediate elevation in inflammatory mediators as they are transfused along with the blood product. Washing removes a fraction of inflammatory mediators, and so it blunts the sudden increase to some degree. However, as early as 1 day after surgery there is no difference in levels of inflammatory mediators between groups of patients that received unwashed salvaged blood, washed salvaged blood or no blood. Furthermore, the levels of the mediators decline in the same fashion in all three groups over the next several weeks.⁷⁰ This means that the only effect that transfusion of salvaged blood, both washed and unwashed, has on the normal postsurgical inflammatory response is a transient (less than 24 hours) increase in plasma inflammatory mediator concentrations. The transient increase of proinflammatory cytokines may contribute to benign febrile reactions seen after transfusion of both washed and unwashed salvaged blood as IL-6 levels have been measured to be higher in both groups among patients that developed febrile reactions.⁶⁸ The transient increase of anti-inflammatory cytokines may contribute to the decreased risk of infection in patients that received salvaged blood.^71,72

Removal of chemical and other contaminants

Salvaged blood is frequently contaminated with debris from the surgical field, by other bodily fluids, and by bacteria. Washing has been shown to eliminate or significantly reduce chemical and physical contaminants. In terms of bodily fluids, one study shows complete removal of urine from a 50% blood : urine mix by washing.⁷³ Even so, there have not been any reports of adverse clinical events due to transfusion of unwashed salvaged blood that has become admixed with urine.⁴³ Klebanoff did a worse-case scenario study in which he added to salvaged blood a mixture of ‘contaminants’ he made by taking mortar ground feces, splenic tissue, liver tissue, bile and body fat, and infusing it without washing. Although blood cultures were positive immediately post-transfusion in the dogs exposed to the mix, they were negative at the end of 24 hours.¹⁸ Even so, there are some instances in which the presence of contaminants is thought to make the blood unsafe to be salvaged, even after washing. Blood collected during resection of a pheochromocytoma may contain high levels of catecholamines and can result in significant hemodynamic changes.^74,75 Caution is also advised during cases where there is the potential for large amounts of pancreatic fluid in the field as pancreatic enzyme transfusion may cause significant intravascular hemolysis.⁷⁶

Detrimental effects of washing

There is some evidence that the added shear stresses from washing do produce some damage to the red cell membranes, causing sublethal injury and an increased mechanical fragility of the RBCs.^46,56 This does not seem to be clinically relevant, however, as the rates of survival of RBCs are the same in washed versus unwashed blood.^77,78 Another potential detrimental effect of washing is dilution of coagulation factors. However, it has been shown in an animal model that coagulopathy due to hemodilution is clinically improved by transfusing washed salvaged blood. Hemorrhagic shock and hemodilution was induced by withdrawing about two-thirds of a blood volume and then volume resuscitating with a colloid infusion. The blood removed was salvaged, washed and reinfused. Clotting time, clot formation time and maximum clot firmness were improved after retransfusion as compared with values after hemodilution.⁷⁹

FILTERING

Screen filters with a 40 μm pore size are routinely used to reduce the risk of infusing debris that could create emboli. Filtration has always been used as part of the blood salvage process, with the blood being filtered through layers of gauze before filtration sets were available.¹² Table 1 lists the specifications for common filters used for cell-salvaged blood. Filtration through microaggregate filters does cause some damage to the blood, with mild increases in FDPs, mild hemolysis, and some platelet and complement activation.^32,62 However, the clinical implications of transfusing unfiltered debris are of much greater clinical significance, and there is no controversy involving the need for microaggregate filtration. ‘Unprocessed’ salvaged blood is, by default, defined as blood that has at minimum been filtered through a microaggregate filter.

Table 1.

Specifications for common filters used for cell-salvaged blood

	Design	Material	Performance	Flow rates (average)
SQ40S	40 μm screen filter	Polyester	Reduction of microaggregates	> 150 mL/minute
LipiGuard® SB	40 μm screen filter and  depth filter	Polyester	Reduction of microaggregates, as  well as some lipid particles, some  leukocytes and some interleukins	10–20 mL/minute
LeukoGuard® RS	Depth filter	Polyester	Reduction of 99% of leukocytes,  most lipid particles, and reduce  number of tumor cells	80 mL/minute

All made by Pall Corporation, New York, NY, USA. Filtering of microaggregates is always done, while filtering for removal of fat particles and leukocytes is done as needed.

Fat emboli in salvaged blood are also of concern. During surgery, especially orthopedic procedures, a significant amount of fat particles are on the operative field and end up in the salvaged product.⁸⁰ Concentrations as high as 194,000 fat particles greater than 9 μm in diameter per milliliter of unwashed salvaged blood have been measured.²⁹ There is histological evidence of fat emboli in the pulmonary circulation of dogs after infusion of unwashed salvaged blood.¹⁹ Microaggregate filters only remove particles larger than the pore size of the filter, and are ineffective at removing fat particles. Leukocyte reduction filters, on the other hand, have been shown to remove 99% of fat particles.⁸¹ An entity called fat embolism syndrome has been described and is associated with orthopedic and cardiac surgery cases in which the potential transfusion of large quantities of fat emboli leads to detrimental consequences, most notably pulmonary damage and potentially acute respiratory distress syndrome.⁸² The incidence and severity of fat emboli syndrome after transfusion of salvaged blood is not known, but there are no widespread reports of pulmonary complications after autotransfusion and so the clinical impact seems to be small. Where there is gross contamination of fat in the salvaged product, however, it seems prudent to either filter with a leukocyte reduction filter or wash the blood before returning it to the patient.⁸³

There is some controversy surrounding the need for additional steps to filter out potential malignant cells in cases of cancer surgery. The concern is inducing metastatic spread of the cancer that is being operated on. It has been shown that salvaged blood from cancer surgery does contain tumor cells, and that washing alone is not enough to remove them.^73,84,85 To reduce the number of potentially malignant cells, leukocyte depletion filters have been advocated and have been shown to reduce the number of tumor cells.^85–87 Clinically, however, the theoretical concern has beenhigh enough to warrant recommendations in the past by the American Medical Association Council on Scientific Affairs against blood salvage during cancer surgery.⁸⁸ More recent clinical studies, however, have shown either no difference in outcome, or better outcomes in patients who received salvaged blood during cancer surgery as compared with those that did not, and therefore current recommendations are for the use of blood salvage in cancer surgery along with the added precaution of a leukocyte reduction filter.⁸⁹

IS WASHING OF SALVAGED BLOOD NECESSARY?

In salvaged blood the coagulation and complement cascade have been activated, red cells have been hemolyzed, platelets have been activated, activated plasma proteins have been released, and the product has been mixed with physical and chemical debris. These changes were caused by a systemic response to surgical trauma, exposure to air and other surgical interfaces, the synthetic surfaces of the machine, and the mechanical trauma of the salvage and washing process itself. It seems logical to return to the patient a product which most closely resembles what was shed in the first place, and to try to reverse some of the changes to the blood that occurred since it left the patient’s circulation. Washing produces a product more closely resembling the patient’s own blood before it was shed, with reduced FDPs and activated platelets, reduced free hemoglobin, reduced leukocyte enzymes, reduced activated complement proteins and reduced chemical and physical debris. Furthermore, there are few detrimental effects of washing and none of clinical significance. Washed salvaged blood can then be said to be a better product than unwashed salvaged blood. However, there are no convincing outcome data that support the argument that unwashed salvaged blood is not good enough. In other words, it is unclear whether the washed product, albeit superior, is really necessary. The clinical literature on cell salvage, unfortunately, does not contain adequate studies that can answer this question with certainty. Studies are hampered by serious methodological weaknesses like small sample sizes, lack of control groups and inadequate randomization. Furthermore, the results are often given using subjective outcome variables, with few clinical outcome and adverse event data being reported. The greatest focus has been on reporting laboratory parameters of the blood products. There are only four randomized controlled trials looking at cell salvage with more than 100 patients in each trial arm, and none that directly compare washed versus unwashed blood.⁷¹ Large methodologically rigorous studies are needed to compare the clinical outcomes and adverse events between washed and unwashed salvaged blood before the washing process can be recommended as a necessary step. In the meantime, the use of both washed and unwashed salvaged blood can still be recommended.