key: cord-0005857-d0rtfa84
authors: Kapadia, F.; Valentine, S.; Smith, G.
title: The role of blood microfilters in clinical practice
date: 1992
journal: Intensive Care Med
DOI: 10.1007/bf01706469
sha: a1c139850871556df8c8b1ca0e21a3877234ade6
doc_id: 5857
cord_uid: d0rtfa84

Blood filters have been available since the 1930s. In this review we evaluate the role of microaggregate filters (MF) in certain transfusion complications, namely non-haemolytic febrile transfusion reactions (NHFTR), pulmonary injury, thrombocytopenia, fibronectin depletion and histamine release. We review the latest generation of leucocyte depleting filters and discuss their role in preventing alloimmunisation, immunosuppression and CMV transmission. Finally, we provide a rationale for the role of blood microfiltration in the present day practice of intensive care medicine.

for filtration to be effective the adsorption force of the blood debris to the filter must be greater than the force acting in the direction of flow [1] .

Eckert [1] has described the mechanism of action of these filters, highlighting that when coated with plasma the absorptive forces decrease and the filter becomes less efficient. In addition he states that other disadvantages of depth filters include prolonged contact with blood, and the possibility of unloading and channelling of particles already filtered. Wenz [3] has shown that these filters possess a larger dead space than screen filters leading to a greater red cell loss.

As a result of the advantages of screen filters over depth filters the clinical trend has been to use the former with depth filters rarely being used in modern day clinical practice.

Opinions on the role of microaggregate filters in routine clinical practice vary. Wenz [4] has stated -"All blood must be administered via a filtration device" and goes on to add "When is the microfiltration of whole blood and red cell concentrates essential? Always. When is it superfluous? A medical device which proves to have at least one advantage and no documented disadvantages is never superfluous".

Others [5, 6] have shown their support for the above view by re-quoting him in their publications.

Derrington [7] however states "the use of microfiltration is probably an unnecessary expense and, where exsanguination is a risk may be positively dangerous" and adds "their routine use for transfusions, small or large, remains to be justified".

The views expressed by Wenz and Derrington are not as divergent as they may appear. Both accept that all patients should get the benefit of doubt until the situation is further clarified. Whilst Derrington states that the expense of giving all patients the benefit of the doubt is probably unjustified on present evidence Wenz states that the benefits outweigh the potential disadvantages and justify the cost. To review this we will examine the formation of microaggregates and the complications they cause.

The early work on microaggregates concentrated on the formation of MA, and it was established that within a few hours of blood collection platelets begin to aggregate and form loosely bound structures [8] . After 24-48 h degenerating white blood cells (WBCs) get trapped in this structure and fibrin precipitation occurs resulting in a stable MA [5] , a process which occurs in blood irrespective of the storage solution used. In the first week of storage, MA formation occurs faster in citrate, phosphate, dextrose anticoagulated blood (CPD) than in adenine, citrate, dextrose anticoagulated blood (ACD) [9] . However, after this time MA numbers are similar in both mediums and it has been estimated that by 2 weeks storage there are approximately 50--250 million microaggregates per unit [10] . Most recently blood stored in optimal additive solutions (Saline, adenine, glucose, mannitol, SAG-M) has been found to contain MAs and larger macroaggregates [6, 11] . Bolton [6] examined 321 units of unopened SAG-M blood and noted that 45.5~ had particles visible to the naked eye and Robertson [11] reported that when the bags were examined by passing the blood through surgical gauze 85% of SAG-M units contained MA or macroaggregates.

When the structure, size range and the numbers of MA in stored blood had been established, attention turned to the clinical effects of infusing such large numbers of MA. Initially work concentrated on the physical effects and sequelae of entrapment of the MA within the lung micro-vasculature [12, 13] , however since then numerous other consequences have been highlighted.

As a result of increased blood matching and development of the closed bag system, fever due to haemolytic reactions or infections are now very rare. The commonest 'adverse transfusion reaction' accounting for 70% of all transfusion related problems is now the intra-or posttransfusion pyrexia [14] of which most are due to a host reaction to donor leucocytes [15] and it has been shown that these can be successfully reduced by transfusing leucocyte poor red cell concentrates [3, 16, 17] .

The incidence of the non-haemolytic febrile transfusion reaction in the "normal" hospital population has been estimated as being between 0.5% [15] and 6.6% [16] . It is higher in the chronically transfused patient [17] and can be as high as 45~ in thalassaemia patients [3, 18] . In an attempt to control the high frequency of reaction many methods have been developed to prepare leucocyte poor red cell concentrates.

Unfortunately malay of these are expensive, result in significant red cell wastage, decrease the shelf life of the unit, disrupt the sterile seal of the unit and are therefore unsuitable for routine hospital transfusion [19] . As a result, simple methods to produce leucocyte poor red cell concentrates have been recently described. These include the spin and filter method [14] and the spin cool and filter method [18] utilising the MF to remove the leucocyte containing MA.

Wenz [3] and later Parravicini [181 demonstrated that these methods significantly reduced the rate of NHFTR in chronically transfused patients. In a series of trials they demonstrated that MF alone caused a fivefold decrease in the rate of the NHFTR while centrifugation and filtration decreased the incidence by 98%. This further decrease in the incidence of NHFTRs is believed to occur as granulocytes, the cells implicated in causing these reactions, are incorporated into MAs during centrifugation and the resulting MAs can be removed by filtration.

Microemboli associated with blood transfusion or otherwise have been implicated as a causative factor in the development of Adult Respiratory Destress Syndrome (ARDS) [12] . Robb et al. [20] demonstrated in an endotoxic albino rat model that progressive occlusion of the pulmonary circulation caused by microemboli occurred with stable systemic arterial and venous blood pressures.

In another animal study, Davidson et al. [21] showed, that in dogs, the rise in pulmonary artery pressure and total body oxygen consumption associated with massive exchange transfusion could be prevented by using a MF, suggesting that exogenous microemboli are the causative agent. In 1979 Dhurandhuer et al. [22] also using an animal model, confirmed the efficacy of MF in preventing pulmonary microemboli and histopathological changes secondary to the microemboli. Connell et al. [23] obtained similar results and demonstrated the MF were effective by showing that transfusion of washed red cells gave no additional protection to the lung.

Human studies have supported these results. Eckert [1] demonstrated histopathologically that MA were present in the lungs of brain dead patients following a 2000 ml blood transfusion and Ruel et al. in two separate studies [12, 131 found that traumatised patients, receiving massive blood transfusion with the MA removed by 40 Ixm filtration, demonstrated a much lower incidence of clinical ARDS than those who received blood through a giving set containing the standard 170 gm clot screen.

Despite these studies the definitive role of MF in prevention of human lung injury needs further clarification. The animal studies show both the role of MA in lung injury and the role of MF in preventing this. However, human studies are less conclusive and the significance of the Reul studies needs to be tempered by the small numbers involved. As yet no prospective randomised human study has demonstrated improved outcome following the use of microfilters.

Thrombocytopenia associated with massive blood transfusion is well recognised [24] . This was presumed to be due to a simple dilutional effect, however recent work [25] [26] [27] have shown that a significant fail in platelet counts in excess of that expected by simple haemodilution occurs with transfusions of a few units of blood.

Examining this in detail, Lim et al. [26] showed that anaemic and thrombocytopenic patients exhibited a significant decrease (41.7~ in platelet count after receiving a 3 unit transfusion of SAG-M blood through a standard blood giving set. The same patients receiving another 3 unit transfusion, at a different time, through a MF had a resulting platelet count fall of only 4.6%. Using radio labelled platelets Bareford [27] showed that the degree of splenic sequestration of the platelets was significantly reduced when blood was transfused through a MF as compared to a standard giving set and could account for the difference in overall platelet levels. He also showed that washed red cells have no further advantage emphasising the role of MA rather than any plasma constituent in causing the splenic sequestration.

More recently Hart et al. [28] have shown that the maximal transfusion induced thrombocytopenia affect is not seen until 72 h post-transfusion. A consequence of this effect being time dependent is that a number of patients may have a platelet count within an acceptable range at 24 h post-transfusion but not at 72 h and therefore extended observation would be advisable.

This effect may also be of significance in intensive care and surgical patients whose platelet count may be falling due to sepsis, infection, operational procedure or trauma.

Fibronectin is an opsonic circulating glycoprotein which has a role in coating invading organisms and debris and presenting them to the reticulo-endothelial system (RES) for clearance from the circulation [29] . Following blood transfusion through a normal giving set there is a significant drop in febronectin levels [29, 30] , whilst transfusing blood through a MF prevents this fall [30] . The method by which this occurs is not known but is probably due to host fibronectin adhering to transfused MA and being prematurely removed from the circulation by the RES [27, 311. In fibronectin depleted septic patients replenishing fibronectin levels has been shown to be beneficial in improving gas exchange in the lung and oxygen utilisation by tissues [32] presumably as the result of an improved pulmonary and systemic microcirculation.

Histamine levels are higher than normal in stored blood [33] . There is a further rise of histamine at the time of transfusion probably due to a combination of warming and contact with the drip tubing [33] . This further rise in histamine is mediated through basophil degranulations [33] und transfusing the blood through a MF decreases the number of infused basophils and thus reduces the rise of histamine [34] . The importance of this in the general population has not been established but coupled with the other transfusion related problems the cumulative effect within some patient groups may be important.

Microfiltration of blood is essentially unphysiological. It places a varying burden on cellular elements [1] and can potentially trigger chemical reactions in response to contact with a foreign surface [7] . Five potential problems exist with MF: 9 exsanguination; 9 haemolysis; 9 release of foreign particles; 9 platelet retention; 9 complement activation.

In the bleeding, critically hypovolaemic patient, the time taken to prime the filter can delay rapid blood volume replacement. However, the priming of a filter usually takes less than 1 rain and once in place does not slow down transfusion time [2] . It has been shown that resistance to flow when a screen filter is present in line alone is not significantly different from that offered by simple giving sets [2] . However, increasing resistance to flow necessitating MF exchange may occur after 3-4 units of whole blood have passed through a depth filter and 7-8 units of whole blood through a screen filter [2] .

Screen MF (which work by filtration through a fixed size pore) appear to be free of the problem of haemolysis while depth filters (which work on the principle of adsorbitive filtration) cause a degree of haemolysis especially with older blood [2, 35] . In a case report a 20 ~tm MF was associated with massive haemolysis and resulted in a death, however subsequent analysis of this report suggests that a neonate MF was improperly used in an infant resulting in massive haemolysis [36] [37] [38] .

In a review on MF, Derrington [7] quotes a reference by Eckert [1] as stating that foreign particles can be released into the blood stream during microfiltration. Scrutiny of Eckert's paper however failed to reveal any mention of this problem.

With depth filters however a theoretical possibility of this problem does exist as it is not possible to flush individual filters during manufacture. This is not the case with screen filters.

Microaggregate filtration has been reported to remove viable platelets when fresh whole blood is transfused through them [7] . To examine this Snyder et al. [39] studied the effect of MF on donor platelets when fresh whole blood or platelet concentrates were transfused through them. They showed with platelet concentrates that there were no significant numerical, biochemical or in vitro functional changes in donor platelets passed through either depth or screen microaggregate filters. In contrast it was found that when whole blood was transfused through depth filters a significant number of platelets were trapped in the filter resulting in a fall in the platelet count reaching the transfusion recipient. This did not occur with screen filters. of a reduction in the platelet count of transfused whole blood than depth filters.

The possibility of MF causing complement induced endogenous MA formation and consequent pulmonary injury has been raised [7] . Yellon [40] demonstrated a 17~ rise in C 3 when nylon filter material was incubated with heparinised blood for 60 min. The relevance of such findings were investigated by Snyder [41] who found less than 10~ rise in C3 when filter material was incubated with heparinised blood and less than 3~ rise when incubated with citrated blood. The difference between anticoagulant types is not surprising as citrate acts by chelating Ca ++ and Mg ++ , essential ions for both the coagulation and complement cascades.

The clinical significance of the perceived rise in complement was also investigated by Snyder who related the experimental work to the clinical setting and passed citrated blood through a MF and once more observed a less than 3% rise in C3 levels and no C6 activation [41] . From this they concluded that the passage of citrated blood through a MF does not pose a clinical risk and does not represent a problem of ME Animal studies support the above conclusion [22] . Dogs sacrificed 48-144h after receiving a massive exchange transfusion through a MF showed no evidence of microemboli in the lungs. This indicates that any endogenous MA load secondary to complement activation is too small to be detected histologically in the lungs of these animals and therefore not of consequence.

In conclusion none of the potential problems suggested to be associated with MF, are of clinical significance. Screen MF cause less haemolysis of older blood and less

Filtration through a MF of blood stored for two weeks results in removal of approximately 90~ 95~ of the leucocytes providing a typical residual count in the order of 8+_3.7 x 10S/unit [14] . This process has been found to be successful in preventing many transfusion related complications, as discussed, but other problems of cytomegalovirus (CMV) transfer, alloimmunisation and immunosuppression, require a greater degree of leucocyte depletion.

Immunosuppressed patients, transplant recipients and chronically transfused patients are particularly prone to these complications.

As the CMV resides solely in the leucocyte, minimising the number of transfused leucocytes will decrease the incidence of transfusion associated CMV transmission. The evidence for the efficacy of leucocyte depletion in preventing CMV transmission is presented in Table 1 .

Transfusion of leucocytes is also believed to be central to the pathogenesis of transfusion associated alloimmunisation and immunosuppression [24] . Prevention of these problems require leucocyte depletion to less than l-5xl06/unit [42] . The evidence for the efficacy of leucocyte depletion in preventing alloimmunisation associated platelet refractoriness is reviewed in Table 2 .

To help prevent these problems in recent years a new generation of leucocyte depleting filters have been developed which allow high level removal of leucocytes from red cells or platelets at the time of transfusion [42] . These filters have been designed to deplete blood products of As a result of these constraints, use is currently restricted to the chronically transfused, those patients particularly at risk of alloimmunisation and those immunosuppressed patients at risk of developing CMV associated disease. If however the immunosuppressive effect of the foreign leucocyte is implicated in post operative infections and recovery rates, as has been recently suggested [43] , the role of leucocyte depleting filters may need to be re-evaluated.

It is the authors' view that from the evidence presented, at the present time, the beneficial effects of MF in the chronically transfused, the thrombocytopenic and those receiving massive blood transfusions, is sufficient to recommend their routine use in these patients. The authors also believe that the routine use of MF in critically ill patients at risk of developing sepsis, ARDS and Multi-Organ Failure, is also justified until such time that a definitive trial disproves the benefit in this siuation. The routine use of MF in patients not at risk of ARDS and Multi-Organ Failure for a single or two unit transfusion is more contentious. In this population the hosts ability to clear a MA load and maintain an appropriate plasma fibronectin level should be adequate. Similarly a drop in platelet levels in patients with previously normal platelet counts and a functioning bone marrow is likely to be of little significance.

Although studies on the non-haemolytic febrile transfusion reaction have been performed in chronically transfused patient, it seems reasonable to assume that the use of MF will decrease the incidence of this reaction in other hospital populations as the mechanisms are the same in all groups. It would similarly decrease the incidence of histamine mediated problems.

In conclusion, it appears that the use of microfilters is warranted in a broad array of intensive care settings, and the authors believe that in the evidence from the liter-

Prevention of pulmonary microembolism during blood transfusion by means of fine mesh filters

Some observations on blood microfllters

Microaggregate blood filtration and the febrile transfusion reaction

When is microfiltration of whole blood and red cell concentrate essential? When is it superfluous

Blood microaggregates, their role in transfusion reactions

Microaggregates and filtration. Correspondence

The present status of blood filtration

ACO and CPO blood preservation

The causes and effects of blood cell aggregates

Filtration in IV therapy. A review of the clinical and technical aspects of intravenous fluid and blood filtration, in four parts. Part III: Clinical aspects of blood filtration

Macroaggregate formation in optimal additive red cells

Protection of the pulmonary vasculature by fine screen blood filtration

Prevention of post traumatic pulmonary insufficiency fine screen filtration of blood

Preparation of granulocyte poor red blood ceils by microaggregate filtration: a simlified method to minimise future transfusion reactions

Non haemolytic febrile transfusion reactions. Quantitative effect of blood components with emphasis on isoantigenic incompatability of leucocytes

Blood transfusion in clinical medicine, 6th edn

Transfusion reactions reported after transfusion of red blood cells and of whole blood

The preparation of leucocyte poor red ceils for transfusion by a simple cost effective techniques

The preparation of red cells depleted of leucocytes

Role of pulmonary microembolism in the haemodynamics of endotoxic shock?

Pulmonary microembolisms associated with massive transfusion

Pulmonary structural changes following microembolism and blood transfusions

Pulmonary microemboli 263 after blood transfusions. An electron microscope study

Coagulation effects associated with massive blood transfusion

Haemostasis testing during massive blood replacement

Thrombocytopenia following routine blood transfusion: micro-aggregate blood filters prevent worsening thrombocytopeuia in patients with low platelet counts

Splenic platelet sequestration following routine blood transfusion is reduced by filtered/washed blood products

Post-transfusion thrombocytopenia: its duration in splenic and asplenic individuals

Plasma fibronectin (opsonic glycoprotein): its synthesis by vascular endothelial cells and role in cardiopulmonary integrity after trauma as related to reticuloendothelial function

Effect of blood transfusion on plasma fibronectin

Effects of blood transfusion on in vivo levels of plasma fibronectin

Opsonic surface binding glycoprotein therapy during sepsis

Histamine levels in stored human blood

Effect of micro filtration on the histamine levels in stored human blood

Microaggregate blood filters: effect on filtration time, plasma haemaglobin and fresh blood platelet counts

In vitro evaluation of a paediatric microaggregate blood filter

RBC destruction caused by a micropore blood filter

Microaggregate filters and neonate patients

Effects of microaggregate blood filtration on platelet concentrated in vitro

Complement activation by blood micro-filters

Activation of complement by blood transfusion filters

Transfusion induced alloimmunisation and immunosuppression and the effect of leucocyte depletion

Blood transfusion and infectious complications following colorectal cancer surgery

Cytomegalovirns seronegative platelets and neukocyte-poor red blood ceils from random donors can prevent primary cytomegalovirus infection after bone marrow transplantation

Prevention of primary cytomegalovirus infection during induction treatment of acute leukemia using at random leucocyte poor blood products

Use of leucocyte poor blood components to prevent cytomegalovirus (CMV) infections in patients with acute leukemia

Prevention of transfusion acquired cytomegalovirus infections in infants by blood filtration to remove leucocytes

Prevention of primary CMV infection by leukocyte depletion of blood products

Prevention of primary cytomegalovirus infection in patients with haematologic malignancies by intensine white cell depletion of blood products

The role of blood product filtration in the prevention of transfusion associated cytomegalovirus infection after marrow transplant

Prevention of platelet refractoriness due to HLA antibodies by administration of leukocyte-poor blood components

A randomized trial of leukocyte-depleted platelet transfusions to modify alloimmunisation in patients with leukaemia

Use of leukocyte-poor blood components and HLA matched platelet donors to prevent alloimmunization

Prevention of refractoriness and HLA alloimmunisation using filtered blood products

Prevention of HLA immunization with leucocyte poor packed cells and platelet concentrates obtained by filtration

Alloimmunisation after leukocyte depleted multiple random donor platelet transfusions

Effective prophylaxis against platelet refractoriness in multitransfused patients by use of leucocyte-free blood components