Date 1999-02-01
Reference number ISO/TC 194 /SC N318

ISO/TC 194/sc
Biological evaluation of medical devices
Secretariat DIN

Circulated to P- and O-members, and to technical committees and organizations in liaison for:

x discussion at next WG 9 meeting on 99-05-19 [Venue/date of meeting]
x comments by 1999-05-01 [date]
x approval for registration as a DIS in accordance with 2.5.6 of part 1 of the ISO/IEC Directives, by 1999-05-01 [date]

(P-members vote only: ballot form attached)
P-members of the technical committee or subcommittee concerned have an obligation to vote.

Title (English)
Biological evaluation of medical devices - Part 4: Selection of tests for interactions with blood

Reference language version: English

Introductory note
ISO/TC 194 decided at the Plenary meeting held on 1998 05-14 in Washington by resolution number 214 to distribute the enclosed document as Committee Draft.

Contents Page


ISO (the International Organization for Standardization) is a worldwide federation of national standards bodies (ISO member bodies). The work of preparing International Standards is normally carried out through ISO technical committees. Each member body interested in a subject for which a technical committee has been established has the right to be represented on that committee. International organizations, governmental and non-governmental, in liaison with ISO, also take part in the work. ISO collaborates closely with the International Electrotechnical Commission (IEC) on all matters of electrotechnical standardization.

Draft International Standards adopted by the technical committees are circulated to the member bodies for voting. Publication as an International Standard requires approval by at least 75% of the member bodies casting a vote.

International Standard ISO 10993-4 was prepared by Technical Committee ISO/TC 194, Biological evaluation of medical devices.

ISO 10993 consists of the following parts, under the general title Biological evaluation of medical devices:

Future parts will deal with other relevant aspects of biological testing. Annexes A, B and C, and Annex D of this part of ISO 10993 are for information only.


The initial source for developing this part of ISO 10993 was the publication, Guidelines for blood/material interactions: Report of the National Heart, Lung, and Blood Institute working group [26]; chapters 9 and 10. This publication is being revised [29].

1 Scope

This part of ISO 10993 gives guidance to agencies, manufacturers, research laboratories and others for evaluating the interactions of medical devices with blood.

It describes:

Detailed requirements for testing cannot be specified because of the limitations in knowledge and precision of tests for interactions of devices with blood. Further, this part of ISO 10993 describes biological evaluation in general terms and may not necessarily provide sufficient guidance for test methods for a specific device. Such test methods must take into consideration device design, materials, clinical utility, usage environment and risk benefit. This level of specificity can only be covered in vertical standards.

Therefore, for medical devices where a device specific standard (vertical standard) exists, the biological evaluation requirements and test methods set forth in that vertical standard shall take precedence over the general requirements suggested in this part of ISO 10993.

2 Normative reference

The following standard contains provisions which, through reference in this text, constitute provisions of this part of ISO 10993. At the time of publication, the edition indicated was valid. All standards are subject to revision, and parties to agreements based on this part of ISO 10993 are encouraged to investigate the possibility of applying the most recent edition of the standard indicated below. Members of IEC and ISO maintain registers of currently valid International Standards.

ISO 10993 - 1: 1997, Biological evaluation of medical devices-Part 1 :Evaluation and testing

3 Definitions

For the purposes of this part of ISO 10993, the definitions given in ISO 10993-1 and the following definitions apply.

3.1 blood/device interaction: Any interaction between blood or any component of blood and a device resulting in effects on the blood, or on any organ or tissue or on the device. Such effects may or may not have clinically significant or undesirable consequences.

3.2 ex vivo: Term applied to test systems that shunt blood directly from a human subject or test animal into a test chamber. If using an animal model, the blood may be shunted directly back into the animal (recirculating) or collected into test tubes for evaluation (single pass). In either case, the test chamber is located outside the body.

3.3 thrombosis/thromboembolism: An in vivo phenomenon resulting in the partial or complete occlusion of the vessel or device. Therefore, characterization of thrombosis/thromboembolism includes ex vivo, in vitro circulation systems and in vivo methods, in either animals or the clinical setting. A thrombus is composed of a mixture of red cells, aggregated platelets, fibrin and other cellular elements.

3.4 coagulation: A measurement of factors of the coagulation and fibrinolytic systems following exposure to devices either in vitro or in vivo.

3.5 platelets and platelet function: Platelet testing includes quantitation of platelet numbers as well as analysis of their struction and function. The testing can include analysis of platelet factors, or components on the platelet surface, released from platelets or adherent to the device surface.

3.6 hematology: Includes quantitation of other factors in the blood. including cellular and soluble components of the blood.

3.7 complement system: Includes the quantitation and characterization of any components of the complement system.

4 Abbreviations

Table 1 provides a list of abbreviations used in the context of this part of ISO 10993.

Table 1 - Abbreviations
Abbreviation Meaning
Bb Product of alternate pathway complement activation
b-TG Beta-thromboglobulin
C4d Product of classical pathway complement activation
C3a, C5a (active) complement split products from C3 and C5
CH 50- 50% total hemolytic complement
CT Computerised Tomography
D-Dimer Specific fibrin degradation products (F XIII cross-1inked fibrin)
ECMO Extracorporeal membrane oxygenator
E.M. Electron microscopy
FDP Fibrin/Fibrinogen degradation products
FPA Fibrinopeptide A
F1+2 Prothrombin activation fragment 1 + 2
iC3b Product of central C complement activation
IVC Inferior vena cava
MRI Magnetic resonance imaging
PAC-1 Monoclonal antibody which recognizes the activated form of platelet surface
glycoprotein IIb/IIIa
PET Positron emission tomography
PF-4 Platelet factor 4
PT Prothrombin time
PTT Partial thromboplastin time
RIA Radioimmunoassay
S-12 Monoclonal antibody which recognizes the alpha granule membrane
component GMP 140 exposed during the platelet release reaction
SC5b-9 Product of terminal pathway complement activation
TAT Thrombin-antithrombin complex
TCC Terminal complement complex
TT Thrombin time
VWF von Willebrand factor

5 Devices contacting blood

Devices contacting blood are categorised in ISO 10993-1.

5.1 Non-contact devices

See ISO 10993- 1.
An example is in vitro diagnostic devices.

5.2 External communicating devices See ISO 10993- 1.

These are devices that contact the circulating blood and serve as a conduit into the vascular system. Examples include but are not limited to those in ISO 10993-1.

5.2.1 External communicating devices that serve as an indirect blood path (see ISO 10993-1) include but are not limited to

5.2.2 External communicating devices in contact with circulating blood (see ISO 10993-1) include but are not limited to atherectomy devices, blood monitors

5.3 Implant devices

These are devices (see ISO 10993-1) that are placed largely or entirely within the vascular system. Examples include but are not limited to

6 Characterisation of blood interactions

6.1 General recommendations

Figure 1 is a decision tree that provides guidance on whether testing for interaction with blood is necessary.

Characterisation of blood interactions may be classified into five categories based on the primary process or system being measured. Table 2 is a listing of blood contacting devices and the categories of testing appropriate to the device. These are recommendations, good rationales can always be developed to justify different selections.

6.1.1 Where possible, tests should use an appropriate model or system which simulates the geometry and conditions of contact of the device with blood during clinical applications, including duration of contact, temperature, sterile condition and flow conditions. For devices of defined geometry the relation of surface area (length) to test results should be evaluated.

The selected methods and parameters should be in accordance with the current state of the art.

NOTE - Only blood-contacting parts should be tested.

6.1.2 Controls shall be used unless their omission can be justified. Where possible, testing should include a device already in clinical use or well-characterised reference materials. Several materials and configurations are available (see ISO 10993-12 [7]).

Reference materials used should include negative and positive controls. All materials tested should meet all quality control and quality assurance procedures of the manufacturer and test laboratory and should be identified as to source, manufacturer, grade and type.

6.1.3 Testing of materials which are candidates to be components of a device should be conducted for screening purposes. However, such tests do not serve as a substitute for the requirement that the complete device be tested under conditions which simulate clinical application.

Figure 1 - Decision tree that provides guidance on whether testing for interaction with blood is necessary

Table 2 - Blood contacting devices and the categories of testing appropriate *Hemolysis testing only

Device / Test Category T

Annuloplasy rings, mechanical heart valves x     x*  
Atherectomy devices       x*  
Blood monitors x     x*  
Blood storage and administration equipment,
Blood collection devices, Extension sets
  x x x*  
Cardiopulmonary bypass system,
Extracorporeal membrane oxygenators system
Intravascular oxygenators,
Hemodialysis/hemofiltration equipment
percutaneous circulatory support devices
x x x x x
intra-aortic balloon pumps, x x x x x
Catheters, Guidewires, Intravascular endoscopes,
Intravascular ultrasound, Lasers systems,
retrograde coronary perfusion catheters,
x x   x*  
Cell Savers   x x x*  
total artificial hearts, ventricular-assist devices, x     x  
Devices for absoption of specific substances from blood   x x x* x
Donor and therapeutic apheresis equipment   x x x* x
Embolization devices       x*  
Endovascular grafts (covered stents) x     x*  
Implantable defibrillators and cardioverters, x     x*  
Intravascular temporary or permanent pacemaker electrodes,
pacemakers lead insulators,
x     x*  
Leukocyte removal filter   x x x*  
Prosthetic (synthetic) vascular grafts and patches,
including arteriovenous shunts
x     x*  
Stents x     x*  
Tissue heart valves x     x*  
Tissue vascular grafts and patches,
including arteriovenous shunts
x     x*  
Vena cava filters x     x*  

6.1.4 Tests which do not simulate the conditions of a device during use may not predict accurately the nature of the blood/device interactions which may occur during clinical applications. For example, some short-term in vitro or ex vivo tests are poor predictors of long-term in vivo blood/device interactions [22], [23].

6.1.5 It follows from the above that devices whose intended use is ex vivo (external communicating) should be tested ex vivo and devices whose intended use is in vivo (implants) should be tested in vivo in an animal model under conditions simulating where possible clinical use.

6.1.6 In vitro tests are regarded as useful in screening external communicating devices or implants but may not be accurate predictors of blood/device interactions occurring upon prolonged or repeated exposure or permanent contact (see 6.3.2). Devices intended for non-contact use only do not require evaluation of blood/device interactions. Devices which come into very brief contact with circulating blood (e.g. Iancets, hypodermic needles, capillary tubes) generally do not require blood/device interaction testing.

6.1.7 The two recommendations in 6.1.5 and 6.1.6, together with clause 5, Figure 1 and Table 2 serve as a guide for the selection of tests listed 6.2.1.

6.1.8 Disposable laboratory equipment used for the collection of blood and performance of in vitro tests on blood should be validated to ascertain that there is no significant interference with the test being performed.

6.1.9 If tests are selected in the manner described and testing is conducted under conditions which simulate clinical applications, the results of such testing have the greatest probability of predicting clinical performance of devices. However, species differences and other factors may limit the predictability of any test.

6.1.10 Because of species differences in blood reactivity, human blood should be used where possible.
When animal models are necessary, for example for evaluation of devices used for prolonged or repeated exposure or permanent contact, species differences in blood reactivity should be considered.
Blood values and reactivity between humans and non-human primates are very similar [23].

NOTE - The use of non-human primates for blood compatibility and medical device testing is prohibited by EU law (86/906/EEC) and some national laws.

However, the use of species other than non-human primates such as the rabbit, pig, calf, sheep, or dog may also yield satisfactory results. Because species differences may be significant (for example platelet adhesion, thrombosis, [17] and hemolysis tends to occur more readily in the canine species than in the human), all results of animal studies should be interpreted with caution.

6.1.11 The use of anticoagulants in in-vivo and ex-vivo tests should be avoided unless the device is designed to perform in their presence. The choice and concentration of anticoagulant used influence blood/device interactions. Devices that are used with anticoagulants shall be assessed using anticoagulants in the range of concentrations used clinically.

6.1.12 Modifications in a clinically accepted device shall be considered for their effect on blood/device interactions and clinical functions. Examples of such modifications include changes in design, geometry, changes in surface or bulk chemical composition of materials and changes in texture, porosity or other properties.

6.1.13 A sufficient number of replications of a test including suitable controls shall be performed to permit statistical evaluation of the data. The variability in some test methods requires that those tests be repeated a sufficient number of times to determine significance. In addition repeated studies over an extended period of blood/device contact provide information about the time-dependence of the interactions.

6.2 Tests and blood interactions categories

6.2.1 Recommended tests for interactions of devices with blood
Recommended tests are organised on the basis of the type of device:

Table 3 - External communicating devices
Table 4 - Implant devices

The tests are classified into five categories based on the primary process or system being measured:

d) thrombosis/thromboembolism,

Thrombosis/thromboembolism is an in vivo phenomenon resulting in the partial or complete occlusion of the vessel or device. Therefore, characterization of thrombosis/thromboembolism includes ex vivo, in vitro circulation systems and in vivo methods, in either animals or the clinical setting. A thrombus is composed of a mixture of red cells, aggregated platelets, fibrin and other cellular elements.

d) coagulation,

Coagulation is a measurement of factors of the coagulation and fibrinolytic systems following exposure to devices either in vitro or in vivo.

d) platelets and platelet functions,

Platelets and platelet function - Platelet testing includes quantitation of platelet numbers as well as analysis of their structure and function. The testing can include analysis of platelet factors, or components on the platelet surface, released from platelets or adherent to the device surface.

e) haematology,

Hematology includes quantitation of other factors in the blood. Including cellular and soluble components of the blood.

e) complement system

Complement system - Includes the quantitation and characterization of any components of the
complement system.

The principles and scientific basis for these tests are presented in annex B. Non-contact devices

These devices generally do not require blood/device interaction testing. Disposable test kits should be validated to rule out interference of materials with test accuracy. External communicating devices

After using Table 2 to determine relevant blood interaction categories for a specific device type, blood interactions appropriate to evaluate for external communicating devices are listed in Table 3. Testing is recommended for devices intended for limited (LI, < 24 hours) and prolonged or repeated (PR, 24 hours to 30 days) exposure. See also 6.1.6. Implant devices

After using Table 2 to determine relevant blood interaction categories for a specific device type, blood interactions appropriate to evaluate for implant devices are listed in Tables 4. Testing is recommended for devices intended for prolonged or repeated (PR, 24 hours to 30 days) exposure or permanent contact (PC, > 30 days).

6.2.2 Indications and limitations

RIAs are available for human blood testing but are not generally available for other species. The human test kits usually do not cross-react with other species except for some non-human primates. Care should be taken when designing test systems to ensure that one is actually measuring activation due to the test material and not an artifact of the system.

Discrepancies in evaluating blood/device interactions may occur because of inadequate materials characterization or inappropriate handling before blood tests are performed. For example the studies may have relied on only one type of test or may have permitted the introduction of foreign material unrelated to the material or device under test. Materials to be used in a low flow (venous) environment may interact with blood quite differently when used in high flow (arterial) situations. Changes in design and/or flow conditions can alter the apparent in vivo hemocompatibility of a material.

Table 3 - External communicating devices

Test Category Evaluation Method Comment
Thrombosis Percent occlusion Pressure drop is not recommended
for devices intended for PR
( see
  flow reduction  
  gravimetric analysis (thrombus mass)  
  light microscopy ( adhered platelets,
leukocytes, aggregates, erythrocytes,
fibrin, etc.)
  pressure drop across device  
  labeled antibodies to thrombotic
  Scanning E.M.(platelet adhesion and
aggregation; platelet and leukocyte
morphology, fibrin)
Coagulation PTT (non-activated)  
  Thrombin Generation  
  Specific coagulation factor assays; FPA,
D-dimer, F1+2, PAC-1, S-12, TAT
Platelets Platelet count / adhesion  
  platelet aggregation  
  template bleeding time  
  platelet function analysis  
  PF-4, Beta-TG;thromboxane B2  
  platelet activation markers  
  platelet microparticles  
  gamma imaging of radiolabelled platelets
111In-1abeled platelet survival
In-labeling is recommended for PR
only (see
Hematology Leucocyte count with or without differential;  
  Leucocyte activation;  
  reticulocyte count; activation specific
release products of peripheral blood
cells(i.e. granulocytes)
recommended for PR only
( see
Complement system C3a; C5a; TCC; Bb; iC3b; C4d; SC5b-9;
CH50; C3 Convertase; C5 Convertase

Table 4 - Implant devices

Test category Method Comments
Thrombosis Scanning E.M.(platelet adhesion and
aggregation; platelet and leukocyte
morphology ; fibrin
  Percent occlusion  
  flow reduction  
  autopsy of devices (gross and
  autopsy of distal organs (gross and
Coagulation Specific coagulation factor assay; FPA,
D-dimer, F1+2, PAC-1, S-12, TAT
  PTT( non activated), PT, TT; Plasma fibrinogen;
Platelets PF-4, p-TG, thromboxane B2,  
  platelet activation  
  platelet microparticles  
  gamma imaging ofradiolabelled platelets
lll-In labeled platelet survival
  platelet count / adhesion  
  platelet aggregation  
Hematology Leukocyte count with or without differential;  
  Leukocyte activation  
  Reticulocyte count; activation specific
release products of peripheral blood cells
(i.e., granulocyes)
C3a, C5a, TCC, Bb, iC3b, C4b, SC5b-9,
CH 50, C3 Convertase, C5 Convertase

6.3 Types of tests

Table 5 gives a list of commercially available assays validated for use with human blood.

6.3.1 In vitro tests

Variables that should be considered when using in vitro test methods include hematocrit, anticoagulants, sample collection, sample age, aeration and pH, temperature, sequence of test versus control studies, surface-to-volume ratio, and fluid dynamic conditions (especially wall shear rate). Tests should be performed with minimal delay, usually within 4 hours, since some properties of blood change rapidly following collection.

6.3.2 Ex vivo tests

Ex vivo tests should be performed when the intended use of the device is ex vivo, for example an external communicating device. Ex vivo testing may also be useful when the intended use is in vivo, for example an implant such as a vascular graft. Such use should not however substitute for an implant test.

Ex vivo test systems are available for monitoring platelet adhesion, emboli generation, fibrinogen deposition, thrombus mass, white cell adhesion, platelet consumption, and platelet activation [17], [27], [43]. Blood flow-rates can be measured with either Doppler or electromagnetic flow probes. Alterations in flow-rates may indicate the extent and course of thrombus deposition and embolization.

May ex vivo test systems use radiolabelled blood components to monitor blood/device interactions. Radiolabelled platelets and fibrinogen are the most commonly labelled components of blood. Alteration of platelet reactivity by the labelling procedure can be minimized by strict attention to technical detail [20], [21], [32].

Advantages of ex vivo tests over in vitro tests are that flowing native blood is used (thus eliminating artifacts caused by anticoagulants as well as providing physiological flow conditions), several materials can be evaluated since the chambers can be changed, and it is possible to monitor some events in real time. Some disadvantages include variability in blood flow from one experiment to another, variable blood reactivity from one animal to the other, and the usually relatively short time intervals that can be evaluated. Positive and negative controls using the same animal are recommended in this regard.

6.3.3 In vivo tests

In vivo testing involves implanting the material or device in animals. Vascular patches, vascular grafts, prosthetic rings, heart valves and circulatory-assist devices are examples of configurations used in in vivo testing. In vivo tests are usually designed to examine hemocompatibility over a period longer than 24 hours.

Patency (of a conduit) is the most common measure of success or failure for most in vivo experiments. The per cent occlusion and thrombus mass are determined after the device is removed. The tendency of thrombi formed on a device to embolize to distal organs should be assessed by a careful gross as well as microscopic examination of organs downstream from the device. The kidneys are especially prone to trap thrombi which have embolized from devices implanted upstream from the renal arteries (for example ventricular-assist devices, artificial hearts, aortic prosthetic grafts) [16].

Methods to evaluate in vivo performance without terminating the experiment are available. Arteriograms are used to determine graft patency or thrombus deposition on devices. Radioimaging can be used to monitor platelet deposition at various time periods in vivo; platelet survival and consumption can be used as indicators of blood/device interactions and passivation due to neointima formation or protein adsorption.

In some in vivo test systems the material's properties may not be major determinants of the blood/device interactions. Flow parameters, compliance, porosity and implant design may be more important than blood compatibility of the material itself As an example, low flow-rate systems may give substantially different results when compared to the same material evaluated in a high flow-rate system. In such cases, test system performance in vivo should carry more importance than in vitro test results.

Table 5 - Commercially available assays for platelet, coagulation, fibrinolytic and complement factors

Factor Type of assay
Plasminogen Colorimetric; Fluorogenic
Antithrombin III Chromogenic, fluorogenic
Protein C Chromogenic
Protein S Chromogenic
Antiplasmin Chromogenic
Prekallikrein Chromogenic
Kallikrein Fluorogenic
PF-4 RIA, ELISA, Chromogenic
Thromboxane B2 RIA, ELISA
Factor VIII-VWF Chromogenic; clotting time
Factor IX Chromogenic; clotting time
Factor Ixa Fluorogenic
Factor X Chromogenic; clotting time
Factor Xa Fluorogenic
Factor XII Chromogenic; clotting time
Factor XIIa Fluorogenic
C3a, C5a RIA
Bb, iC3b, C4d, SC5b-9 ELISA
D-dimer RIA, ELISA
P-selectin ELISA
L-selectin ELISA

NOTE -These assays have been validated for human use. Validation of their accuracy for other species shall be determined prior to use.

Annex A (informative) Preclinical evaluation of cardiovascular devices and prostheses

A.1 General considerations

A.1.1 This annex provides background for selecting tests to evaluate the interactions of cardiovascular devices with blood. Subclause 6 contains guidance on when testing is necessary, what blood interaction categories might be appropriate for specific devices, and a complete list of tests for evaluating blood/device interactions of non-contact-, external communicating-, and implant-devices. A.1.2 The following classification of blood/device interactions is provided as background.

A.1.2.1 Interactions which mainly affect the device and which may or may not have an undesirable effect on the subject are as follows:

A.1.2.2 Interactions which have a potentially undesirable effect on the animal or human are as follows:

A.1.3 Advantages and limitations of animal and in vitro testing

A.1.3.1 Animal testing

Animal models permit continuous device monitoring and systematic controlled study of important variables. However the choice of an animal model may be restricted by size requirements, the availability of certain species and cost. It is critical that the investigators be mindful of the physiological differences and similarities of the species chosen with those of the human, particularly those relating to coagulation, platelet functions and fibrinolysis, and the response to pharmacological agents such as anesthetics, anticoagulants, thrombolytic and antiplatelet agents, and antibiotics. Because of species differences in reactivity and variable responses to different devices, data obtained from a single species should be interpreted with caution. Non-human primates such as baboons bear a close similarity in hematologic values, blood coagulation mechanism and cardiovascular system to the human [27]. An additional advantage of a non-human primate is that many of the immunologic probes for thrombosis developed for humans are suitable for use in other primates. These probes include PF-4, b-TG, FPA, TAT, and F. + 2 The dog is a commonly used species and has provided useful information; however, device-related thrombosis in the dog tends to occur more readily than in the human, a difference which can be viewed as an advantage when evaluating this complication. The pig is generally regarded as a suitable animal model because of its hematologic and cardiovascular similarities to the human. The effect of the surgical implant procedure on results should be kept in mind and appropriate controls included

A.1.3.2 In vitro testing

Because of species differences in hemostatic and hematologic factors and activities it is preferable to use human blood. Thrombus formation is a dynamic process. Therefore in vitro testing is advisable to simulate as much as possible the dynamic conditions (for example shear forces of the blood - material interface) in which thrombosis occurs. Since patients with cardiovascular devices may be receiving anticoagulant or antithrombotic drugs, it is important to simulate these conditions in vitro. Static tests are useful for evaluating the interactions of blood with materials.

A.1.4 Test protocols for animal testing

The test protocols recommended follow certain general guidelines. Thrombosis, thromboembolism, bleeding and infection are the major deterrents to the use and further development of advanced cardiovascular prostheses. For devices with limited blood exposure (< 24 h), important measurements are related to the acute extent of variation of hematologic, hemodynamic and performance variables, gross thrombus formation and possible embolism. With prolonged or repeated exposure, or permanent contact (> 24 h), emphasis is placed on serial measurement techniques that may yield information regarding the time course of thrombosis and thromboembolism, the consumption of circulating blood components, the development of intimal hyperplasia and infection. In both of the above exposure and contact categories, assessment of hemolysis assessment and platelet function is important. Thrombus formation may be greatly influenced by surgical technique, variable time-dependent thrombolytic and embolic phenomena, superimposed device infections and possible alterations in exposed surfaces, for example intimal hyperplasia and endothelialization.

The consequences of the interaction of artificial surfaces with the blood may range from gross thrombosis and embolization to subtle effects such as accelerated consumption of hemostatic elements; the latter may be compensated or lead to depletion of platelets or plasma coagulation factors.

A.1.5 Recommended test protocols for in vitro testing

In vitro testing allows for the performance of sufficient number of tests for statistical evaluation without the sacrifice of animals and with relatively low costs. Measurements are related to the acute extent of variation of hematologic, hemodynamic and performance variables, gross thrombus formation and complement activation. The in vitro approach permits the study of the kinetics of various factors and activities, by varying the duration of exposure of material or devices to blood.

A.2 Cannulae

Cannulae are typically inserted into one or more major blood vessels to provide repeated blood access. They are also used during cardiopulmonary bypass and other procedures. They may be tested acutely or chronically and are commonly studied as arteriovenous (AV) shunts. The use of Cannulae induces little alteration in the levels of circulating blood cells or factors in the coagulation or complement system. Cannulae, like other indirect blood path devices (5.2.1), generally require less testing than devices in direct contact with circulating blood (5.2.2, 5.3).

A.3 Catheters and guidewires

Most of the test considered under Cannulae are relevant to the study of catheters and guidewires. The location or placement of catheters in the arterial or venous system can have a major effect on blood/device interactions. It is advised that simultaneous control studies be performed using a contralateral artery or vein. Care should be taken not to strip off thrombus upon catheter withdrawal. In situ evaluation may permit assessment of the extent to which intimal or entrance site injuries contributed to the thrombotic process. In general Doppler blood flow measurements are more informative than angiography. Kinetic studies with radiolabelled blood constituents are recommended only with chronic catheters.

A.4 Extracorporeal oxygenators, hemodialyzers, therapeutic apheresis equipment, and devices for absorption of specific substances from blood

The hemostatic response to cardiopulmonary bypass may be significant and acute. Many variables such as use of blood suction, composition of blood pump priming fluid, hypothermia, blood contact with air and time of exposure influence test values. Emboli in outflow lines may be detected by the periodic placement of blood filters ex vivo, or the use of ultrasonic radiation or other non-invasive techniques. Thrombus accumulation can be directly assessed during bypass by monitoring performance factors such as pressure drop across the oxygenator and oxygen transfer rate. An acquired transient platelet dysfunction associated with selective alpha granule release has been observed in patients on cardiopulmonary bypass [28]; the template bleeding time and other tests of platelet function and release are particularly useful.

Complement activation is caused by both hemodialyzers and cardiopulmonary bypass apparatus. Clinically significant pulmonary leucostasis and lung injury with dysfunction may result. For these reasons, it is useful to quantify complement activation with these devices.

Therapeutic apheresis equipment and devices for absorption of specific substances from the blood, because of their high surface-to-volume ratio, can potentially activate complement, coagulation, platelet and leukocyte pathways. Examination of blood/device interactions should follow the same principles as for extracorporeal oxygenators and hemodialyzers.

A.5 Ventricular-assist devices and total artificial hearts

These devices may induce considerable alteration in various blood components. Factors contributing to such effects include the large foreign surface area to which blood is exposed, the high flow regimes and the regions of disturbed flow such as turbulence or separated flow. Tests of such devices may include measurements of hemolysis, fibrinogen concentration, thrombin generation, platelet survival and activation, complement activation, and close monitoring of liver, renal, pulmonary and central nervous system function. A detailed pathologic examination at surgical retrieval is an important component of the evaluation [37], [38].

A.6 Heart valve prostheses

Invasive, non-invasive and in vitro hydrodynamic studies are important in the assessment of prosthetic valves.

2D and M mode echocardiography makes use of ultrasonic radiation to form images of the heart. Reflections from materials with different acoustic impedances are received and processed to form an image. The structure of prosthetic valves can be examined. Mechanical prostheses emit strong echo signals and the movement of the occluder can usually be clearly imaged. However the quality of the image may depend upon the particular valve being examined. Echocardiography may also be useful in the assessment of function of tissue-derived valve prostheses. Vegetations, clots and evidence of thickening of the valve leaflets are elucidated. Using conventional and color flow Doppler echocardiography, regurgitation can be identified and semi-quantified.

Measurements of platelet survival and aggregation, blood tests of thrombosis and hemolysis, pressure and flow measurements, and autopsy of the valve and adjacent tissues are also recommended.

A.7 Vascular grafts

Both porous and non-porous materials can be implanted at various locations in the arterial or venous system. The choice of implantation site is determined largely by the intended use for the prosthesis. Patency of a given graft is enhanced by larger diameter and shorter length. A rule of thumb for grafts less than 4 mm ID is that the length should exceed the diameter by a factor of 10 (i.e., 40 mm for a 4 mm graft) for a valid model. Patency can be documented by palpation of distal pulses in some locations and by periodic angiography. Ultrasonic radiation, MRI, and PET may also be useful. Results of serial radiolabelled platelet imaging studies correlate with the area of nonendothelialized graft surface in baboons [27]. Radiolabelled platelets facilitate non-invasive imaging of mural thrombotic accumulations. Serial measurements of platelet count, platelet release constituents, fibrinogen/fibrin degradation products and activated coagulation species also are recommended. Autopsy of the graft and adjacent vascular segments for morphometric studies of endothelial integrity and proliferative response can provide valuable information.

A.8 IVC filters, stents and stented grafts

These devices can be studied by angiography and ultrasonic radiation. Other techniques useful for vascular graft evaluation (see A.7) are appropriate here as well.

Annex B (informative) Laboratory tests: principles, scientific basis and interpretation

B.1 General

B.1.1 The principles and scientific basis of the tests listed in 6.2.1 are described here. Detailed methods are found in standard texts of laboratory medicine and clinical pathology. References [14] to [39] and [41] to [44] describe tests which may be useful in the evaluation of blood/device interactions. Because of both biological variability and technical limitations, the accuracy of many of these tests is limited. When possible, the tests should be repeated a sufficient number of times to determine the significance of the results.
(Ref. ICH)*** Note to WG-9 Group - Does anyone recall what this reference is supposed to be? I can't find it)

B.1.2 Principles for in vitro testing

Static systems and dynamic systems i. e. Chandler and other circulatory loop and centrifugation systems are used (Ref.***).

B. 1.3 Test conditions

In order for tests to be of use in the in vitro evaluation of blood/device interactions, anticoagulated blood or plasma collected from normal human subjects or experimental animals should first be exposed to the material or device under standardized conditions including time, temperature and flow. An aliquot of the exposed blood or plasma is then tested shortly after exposure. Conditions of exposure should be based on the intended use of the device. One proposed set of conditions is as follows:

In the preparation of the test specimens it is essential to avoid activation or release of any component from blood before testing. However, the appropriate conditions depend on the device or material being tested and its intended end-use.


When evaluating externally communicating devices and implant devices while in their in-use position, blood is collected into an anticoagulant and the test is performed as described without a prior exposure stage. The tests are classified into five categories, as defined in 6.2.1, according to the process or system being tested: thrombosis, coagulation, platelets and platelet functions, hematology and complement system.

B.2 Thrombosis

B.2.1 Percentage occlusion

Percentage occlusion is visually quantified after a device has been in use and has been removed. This is a measure of the severity of the thrombotic process in a conduit. Lack of occlusion does not necessarily eliminate the existence of a thrombotic process, since thrombi may have embolized or been dislodged before percentage occlusion is measured. Occlusion may be caused not only by thrombosis, but also by intimal hyperplasia, especially at perianastomotic sites in vascular grafts; microscopic examination is required to identify the nature of the occlusive process. Surface area covered by thrombus and thrombus free surface area are semi-quantitative tests that can be used on a comparative basis.

B.2.2 Flow reduction

Flow (rate or volume) is measured after a period of use. Measurements may be performed either during use, or before and after use. Rationale and interpretation are the same as B.2.1.

B.2.3 Gravimetric analysis (thrombus mass)

This is conducted after removal of the device from the in-use position. Rationale and interpretation are as for B.2.1.

B.2.4 Light Microscopy

By this technique, information can be obtained regarding the density of cells, cellular aggregates and fibrin adherent to materials, as well as the geographic distribution of these deposits on the materials or device. The method is semiquantitative.

B.2.5 Pressure drop across device

This is measured before and after a period of use. Rationale and interpretation are as for B.2.1.

B.2.6 Scanning E.M.

Rationale and interpretation are the same as B.2.4. This method has the advantage over B.2.4 of providing greater detail about fine structure of components being examined. Quantitative conclusions require sufficient replicate determinations to establish degree of reproducibility.

B.2.7 Antibody binding

Next to qualitative microscopic judgment of fibrin and platelet deposition on materials, a quantitative estimation is possible by measuring the amount of labeled antibody specific for fibrin(ogen) or platelet membrane receptors. For this purpose materials are washed after exposure to blood to delete nonadherent blood components prior to labeled antibody binding.

B.2.8 Autopsy of device

This method is of great importance in evaluating the biological responses to implanted devices. The distribution, size and microscopic nature of cellular deposits can best be determined by a careful and detailed autopsy. Proposed procedures have been published [37], [38].

B.2.9 Autopsy of distal organs

The rational is to examine for distal effects of implanted devices. These effects include thromboembolism, infection and embolization of components of the device.

B.2.10 Imaging techniques: Angiography, intravascular ultrasound, Doppler ultrasound, CT and MRI

Choices can be made among these methods to determine patency or degree of narrowing of a graft or other conduit and to detect thrombus deposition on devices during their in vivo performance.

B.3 Coagulation

Coagulation methods are based on the use of native (fresh, non-anticoagulated) whole blood, anticoagulated whole blood (usually citrated), platelet-rich plasma or platelet-poor plasma. Since most of the standard coagulation assays are designed to detect clinical coagulation disorders which result in delayed clotting or excessive bleeding, the protocols for evaluating blood/device interactions should be modified appropriately to evaluate accelerated coagulation induced by biomaterials. Reagents for tests based on the activated partial thromboplastin time include an activator such as kaolin, celite, or ellagic acid. Reagents with such activators should be avoided because they tend to mask the acceleration of coagulation which materials and devices cause. The material to be tested serves as the activator; controls (without the material) should be included.

Blood is exposed to test materials either in a static chamber such as a parallel plate cell or in a closed-loop system where the inner surface of the tubing is the test material. After a predetermined contact time with the test surface, tests of the surface and blood can be conducted.

B.3.1 Partial Thromboplastin Time (PTT)

The partial thromboplastin time [35] is the clotting time of recalcified citrated plasma on the addition of partial thromboplastin. Partial thromboplastin is a phospholipid suspension usually extracted from tissue thromboplastin, the homogenate from mammalian brain or lung. Shortening of the PTT following contact with a material under standard conditions indicates activation of the contact phase of blood coagulation. A prolonged PTT suggests a deficiency in any of the plasma coagulation factors I (fibrinogen), II (prothrombin), V, VIII, IX, X, XI, or XII, but not VII or XIII. Heparin and other anticoagulants also cause a prolonged PTT.

Partial thromboplastin reagents using various activating substances such as kaolin or celite are commercially available. Using these reagents, the test is called the activated partial thromboplastin time (APTT). The APTT is of no value in the in vitro evaluation of blood/device interactions because the activating substances mask any activation caused by the device or its component materials.

B.3.2 Prothrombin Time (PT)

Blood is mixed with a measured amount of citrate and the plasma is obtained by centrifugation [35]. This test is based on having an optimum concentration of calcium ions and an excess of thromboplastin, the only variable being the concentration of prothrombin and accessory factors in a carefully measured volume of plasma. Oxalate is not recommended as the anticoagulant because factor V is less stable in this solution and the PT may become prolonged as a result.

This test measures prothrombin and accessory factors. In the presence of tissue thromboplastin, the clotting time depends on the concentrations of prothrombin, factor V, factor VII and factor X (assuming fibrinogen, fibrinolytic and anticoagulant activity to be normal). A prolonged prothrombin time generally indicates a deficiency of prothrombin or factor V, VII, X or fibrinogen.

B.3.3 Thrombin Time (TT)

The thrombin time [35] is the time required for plasma to clot when a solution of thrombin is added. The thrombin time is prolonged with a deficiency in fibrinogen (below 100 mg/dl), qualitative abnormalities in fibrinogen and elevated levels of FDP or heparin. The test is useful for evaluating implant devices only.

B.3.4 Thrombin generation

Materials exposed to an intact coagulation system in the presence of phospholipids (see B.3.1) will generate thrombin which can be measured by conversion of a chromogenic substrate. This method has a much lower variability than the conventional coagulation tests.

B.3.5 Fibrinogen

Dysfibrinogenemia, afibrinogenemia and hypofibrinogenemia cause prolonged PT, PTT and TT results [18]. The screening test most sensitive to fibrinogen deficiency is the TT. If the exact level of fibrinogen is needed, a commercially available modified TT is recommended. The test is useful for evaluating implant devices only.

B.3.6 Fibrinogen and fibrin degradation products (FDP)

Normal physiological fibrinolysis yields the FDPs X, Y. C, D and E in concentrations below 2 mg/ml of plasma. The normally low level of FDPs is maintained by the low rate of the degradation reaction and the high rate of clearance of FDPs from the circulation. Pathologic degradation of fibrin and fibrinogen, a result of increased plasminogen activation, yields FDP of 2 mg/ml to 40 mg/ml or more. The test is useful for evaluating implant devices only. The use of ELISA is recommended.

B.3.7 Specific coagulation factor assays

Significant reduction (e.g. to less than 50% of the normal or control level) of coagulation factors following exposure of blood to a material or device under standard conditions suggests accelerated consumption of those factors by adsorption, coagulation or other mechanisms.

B.3.8 FPA, D-dimer, F1+2, TAT

Elevated levels of FPA, D-dimer, or F1 + 2 indicate activation of the coagulation mechanism. FPA and F1+2 indicate an activation of Prothrombin to thrombin. Elevated TAT complexes indicate activation of blood coagulation and formation of a complex between thrombin being generated and circulating antithrombin. Dimers are plasmin digested degradation products of F XIII cross linked fibrin (coagulation and fibrinolysis). The use of ELISA and RIA is recommended.

B.4 Platelets and platelet functions

It is essential to avoid activation in the preparation of platelet suspensions.

B.4.1 Platelet count

It is important to determine the platelet count [18], [44] because of the key role of platelets in preventing bleeding. A significant drop in platelet count of blood exposed to a device may be caused by platelet adhesion, platelet aggregation, platelet sequestration (for example in the spleen), or blood coagulation on materials or devices. A reduction in platelet count during use of an implanted device may also be caused by accelerated destruction or removal of platelets from the circulation. Platelet count is performed using an EDTA suspension medium.

Blood collection techniques should be reproducible. Platelets can become hyperactive under a variety of conditions, including improper blood collection. Tests to verify normal platelet reactivity are usually performed with an aggregometer. Platelet preparations with reduced reactivity are easily detected using this method, but hyperactive platelets are not normally detected. Platelet aggregation tests can be modified (by appropriately reducing the concentration of platelets or aggregating agents) to determine if platelets become hyperactive following exposure to a material or device.

B.4.1.1 Manual platelet count

Platelet counts are performed manually in some clinical laboratories despite the wide availability of highly accurate platelet counting instruments (B.4. 1.2).

B.4.1.2 Automated platelet count using whole blood or platelet-rich plasma

Automated platelet counts may be performed using either well mixed whole blood or platelet-rich plasma (PRP). The use of whole blood is preferred because using PRP may yield inaccurate results. Instrument counts are made on a greater number of particles than manual counts. Consequently, the coefficient of variation for automated counts may be as small as 4%, whereas the best achievable for the phase microscope manual count is 11%. Most instruments also contain circuitry that distinguishes between platelets and small nonbiological particles, eliminating the need for visual recognition and differentiation of debris from platelets.

B.4.2 Platelet aggregation

Platelet aggregation [35] is induced by adding aggregating agents to PRP that is being stirred continually (e.g. ADP, epinephrine, collagen, thrombin, etc.). As the platelets aggregate, the plasma becomes progressively clearer. An optical system (aggregometer) is used to detect the change in light transmission and a recorder graphically displays the variations in light transmission from the baseline setting. Delayed or reduced platelet aggregation may be caused by platelet activation and release of granular contents, increased FDP or certain drugs (e.g. aspirin, nonsteriod anti-inflammatory drugs). It is important to bear in mind that platelet aggregation using some agents varies or may be absent in some animal species. Spontaneous platelet aggregation, occurring in the absence of added agonists, is an abnormal condition indicating activation of platelets. Platelet aggregates can also be screened automatically by the WU/HOAK method (Ref***). Note: Need WG-9 members help in determining what this reference should be.

B.4.3 Blood cell adhesion

Blood cell adhesion [31] is a measure of the blood-compatibility of a material when considered in conjunction with distal embolization or evidence of activation of one or more hematological factors.

Various methods have been designed to measure the adhesion of cells to surfaces. Most of these methods are based on the observation that a certain proportion of platelets are removed from normal whole blood as a result of passage through a column of glass beads under controlled conditions of flow or pressure. This principle has been adapted to the quantification of the adhesion of other blood cells to polymers coated on glass beads. By such a method it has been reported [31] that adhesion of canine species peripheral lymphocytes and polymorphonuclear leukocytes (PMNs) to beads coated with poly(hydroxyethyl methacrylate) (PHEMA) is lower than to beads coated with polystyrene and certain other polymers. Isolated lymphocytes and PMNs were used in this study.

An alternative method is the direct counting of platelets adherent to a test surface. Following exposure to blood or platelet-rich plasma under standardized conditions, the test surface is rinsed to remove nonadherent cells, fixed and prepared for either light or scanning electron microscopy. The number of adherent platelets per unit area is directly counted and their morphology (e.g. amount of spreading, degree of aggregate formation) is recorded. Alternatively, platelets prelabeled with 51Cr or 111In may be used [30].

B.4.4 Platelet activation

The use of certain materials or devices may cause platelet activation, which can result in; 1) the release of platelet granule substances, such as BTG ( Beta thromboglobulin), platelet factor 4 (PF 4), and serotonin, 2) altered platelet morphology and 3) cause the generation of platelet microparticles. Activated platelets are pro-thrombotic. Platelet activation can be evaluated by various means: microscopic (light and electron microscopy) examination of platelet morphology of platelets adherent to the material or device, measurement of BTG, PF4 and serotonin, and the evaluation of platelet activation by flow cytometry (for microparticle generation, Pselectin (GMP-140) expression, or activated glycoprotein Ib and JIB/ IIIa expression using monoclonal antibodies. Different epitopes of activated platelets are recognized by flow cytometry using 2 antibodies: one specific for platelets (i.e. GP Ib or GP IIb/IIIa) and one specific for platelet activation (P-Selectin).

B 4.5 Template bleeding time

The commercial availability of a sterile disposable device for producing a skin incision of standard depth and length under standard conditions has significantly improved the reproducibility and value of this test. A prolonged result indicates reduced platelet function or reduced platelet count, the latter can be determined separately (B.4.1). A prolonged bleeding time combined with a normal platelet count has been observed in association with some external communicating devices with limited exposure (e.g. cardiopulmonary bypass) [28]. The test is suitable for use with some experimental animals. In vitro bleeding time measurements are also suitable.

B.4.6 Platelet function analysis

The classical template bleeding time has been used as the principle for an automated method. Whole blood is aspirated through a collagen filter with a 150 um aperture. Platelets adhere and aggregate until the aperture is closed. Blood pressure and temperature are standardized, anticoagulation does not affect this test. The test is suitable for animal blood.

B.4.7 Gamma imaging of radiolabelled platelets

The high gamma emission of 111In enables it to be used for this purpose [20], [27]. This method enables the localization and quantification of platelets deposited in a device. The technique is useful for external communicating as well as implant devices.

B.4.8 Platelet lifespan (survival)

Platelets are obtained from the patient's blood and are labeled with 51Cr or 111In [20], [21],[32]. Both these agents label platelets of all ages present in the sample, do not elute excessively from the platelets and are not taken up by other cells or reused during thrombopoiesis. Idium-111 has the advantage of being a high gamma emitter, requiring the labeling of fewer platelets and enabling surface body counting to assess localized platelet deposition to be combined with the lifespan study. A reduced platelet lifespan indicates accelerated removal from the circulation by immune, thrombotic or other processes.

B.5 Hematology

B.5.1 Leukocytes

Leukocyte activation can be determined by the microscopic examination of the device surface or activated leukocytes and the use of flow cytometry for the evaluation of increased leukocyte markers such as L- selectin and CD 1 lb and quantitative disturbances in lymphocyte subpopulations.

B.5.2 Hemolysis

This is regarded as an especially significant screening test because an elevated plasma hemoglobin level, which if properly performed indicates hemolysis, reflects red blood cell membrane fragility in contact with materials and devices (see annex C).

B.5.3 Reticulocyte count

An elevated reticulocyte count indicates increased production of red blood cells in the bone marrow. This may be in response to reduced red blood cell mass caused by chronic blood loss (bleeding), hemolysis or other mechanisms.

B.6 Complement System - CH 50 and C3a,C5a, TCC, Bb, iC3b, C4d, SC5b-9

CH 50 decrease is an indicator of total complement consumption. Elevated levels of any of these complement components indicate activation of the complement system. Some materials activate complement, and activated complement components in turn activate leukocytes, causing them to aggregate and be sequestered in the lungs.

Measurement of complement split products have the disadvantage if species-specificity and high baseline levels when performed after in vitro testing. The classical CH-50 method appears useful with human, bovine, porcine, and lupine serum.

Another functional method of measurement of complement activation in vitro is the generation of complement C3- or C5-convertase, determined by substrate conversion.

Annex C (informative) Evaluation of hemolytic properties of medical devices

Extensive literature exists describing blood/material interactions. Unfortunately, very few methods exist which are reliable, reproducible, and predict clinical performance. This annex will review the known hemolysis test methods and discuss factors pertaining to their ability to characterize medical and dental materials and devices.

C.1 Definitions

C.1.1 anticoagulant: An agent which prevents or delays blood coagulation [45]. For example, heparin and citrate.

C.1.2 oncotic pressure (colloidal osmotic pressure): The total influence of the proteins or other large molecular weight substances on the osmotic activity of plasma [46].

C.1.3 hematocrit: The ratio of the volume of erythrocytes to that of whole blood.

C.1.4 hemolysis: The rupture of the erythrocyte membrane with the liberation of hemoglobin, which diffuses into the fluid surrounding the erythrocytes, or the premature destruction of, erythrocytes.

C.1.5 negative reference material: High Density Polyethylene, or similar validated alternative. (see ISO 10993-12).

C.1.6 packed erythrocytes: Component obtained by centrifugation from a unit of human blood following removal of plasma supernatant.

NOTE - Properties of human erythrocytes for transfusion: The erythrocyte volume fraction of the component is 0.65 to 0.80. The unit contains all of the original unit's erythrocytes, the greater part of its leukocytes (about 2.5 to 3.0 x 109 cells) and a varying content of platelets depending on the method of centrifugation.

C.1.7 washed erythrocytes: An erythrocyte suspension obtained from whole blood after removal of plasma and washing in an isotonic solution.

NOTE - Properties: This component is an erythrocyte suspension from which most of the plasma, leukocytes and platelets have been removed. The amount of residual plasma will depend upon the washing protocol. Storage time should be as short as possible after washing and certainly not longer than 24 hours at 1_C to 6_C.

C.1.8 whole blood: Whole blood is unfractionated blood drawn from a selected donor and contains citrate or heparin as an anticoagulant.

C.2 Causes of hemolysis

C.2.1 Mechanical forces - pressure

The red blood cell membrane is a semipermeable membrane. A pressure differential will occur when two solutions of different concentrations are separated by such a membrane. Osmotic pressure occurs when the membrane is impermeable to passive solute movement. These pressure differentials can cause erythrocyte swelling and cell membrane rupture with release of free hemoglobin [45].

C.2.2 Mechanical forces - rheologic

Factors which influence velocity of blood flow, shear forces and other forces that can deform the red blood cell membrane can cause membrane rupture.

C.2.3 Biochemical factors

Changes to membrane structure on a molecular level can modify the strength and elastic properties of the erythrocyte membrane. A deficiency of nutritional factors or metabolic energy (ATP) can result in loss of the discoid shape and microvesiculation of hemoglobin. Other chemicals, bacterial toxins, pH and metabolic changes induced by temperature can compromise the erythrocyte membrane [47]. These changes may cause membrane rupture at lower than expected osmotic pressures. A test to determine the pressure at which an erythrocyte membrane ruptures is termed osmotic fragility.

C.3 Clinical significance of hemolysis

C.3.1 Toxic effects

Elevated levels of free plasma hemoglobin may induce toxic effects or initiate processes which may stress the kidneys or other organs [45].

C.3.2 Thrombosis

Intravascular hemolysis may promote thrombosis by liberating phospholipids [48]. When hemolysis causes a clinically significant drop in red blood cell count, anemia and compromised oxygen carrying capacity with its subsequent effects on the brain and other organs or tissues may result.

C.4 Determining a Pass/Fail assessment for hemolysis

Hemolysis is a function of time and material properties such as surface energy, surface morphology, and surface chemistry. Hemolysis is also a function of shear stress, cell-wall interaction, character of adsorbed protein layers, flow stability, air entrainment, and variations of blood source, age and chemistry [49], [50], [51]. These variables need to be adequately controlled for comparisons of hemolytic potential among materials and medical devices. The spectrum of methods for evaluating hemolysis varies from simplified to highly complicated models. Specific in vitro and in vivo models with flowing blood have been published. Studies of hemolytic potential are relative comparisons against materials or medical devices tested in the same model by a specific laboratory rather than absolute measures. In vitro test methods are able to quantify small levels plasma hemoglobin which may not be measurable under in vivo conditions (e.g. due to binding of plasma hemoglobin to heptoglobin and rapid removal from the body). Hence, the measurement of in vivo plasma hemoglobin levels as an indicator of overall blood damage is not as sensitive as in vitro models, which therefore are possibly of no clinical consequence.

It is not possible to define a universal level for acceptable and unacceptable amounts of hemolysis for all medical devices and applications. The effect of a device on hemolysis may be masked in the short term by the trauma of the surgical procedure. A device may cause a substantial amount of hemolysis, but be the only treatment available in a life-threatening situation. Intuitively, a blood compatible material is non-hemolytic. In practice, many devices cause hemolysis, but their clinical benefit outweighs the risk associated with the hemolysis. Therefore when a device causes hemolysis, it is important to confirm that the device provides a clinical benefit and that the hemolysis is within acceptable limits clinically. Acceptance criteria should be justified based on some form of Risk and Benefit assessment. The following questions are suggestions for developing such an assessment:

C.5 Hemolysis testing - general considerations

C.5.1 Methods

C.5.1.1 Total blood hemoglobin - Classical methods

Due to many different factors, there are currently about twenty different assays in use today for measuring plasma hemoglobin as an indicator of hemolysis, but no one method is widely accepted. The concentration of hemoglobin in plasma is significantly less than the total blood hemoglobin concentration. The free plasma hemoglobin concentration is normally 0 - 10 mg/dL in vivo and whereas the normal range of total blood hemoglobin concentration is 11,000 to 18,000mg/dL. Classically, three analytical methods have been used to determine total blood hemoglobin (Hb) concentrations [52].

C. Cyanmethemoglobin method

The first classical method, cyanmethemoglobin detection, is that recommended by the International Committee for Standardization in Haematology [53]. The cyanmethemoglobin (hemiglobincyanide; HiCN) analysis has the advantage of convenience, ease of automation, and the availability of a primary reference standard (HiCN). The method is based on the oxidation of Hb and subsequent formation of hemiglobincyanide which has a broad absorption maxima at 540 nm. For the total hemoglobin concentration, the spectral interference due to plasma is minimal and the sample absorbance can be compared to the HiCN standard solution directly. However, for the determination of plasma hemoglobin, unless compensation is made for endogenous plasma background interference, the plasma hemoglobin concentration of the sample will be overestimated. Lysing agents such as detergents are used which, in addition to releasing Hb from the erythrocyte, decrease the turbidity (a source of interference as false absorbance at 540 nm) from protein precipitation.

NOTE - An erythrocyte Iysing agent should be used for total hemoglobin measurements. Lysing agents should not be used for cell-free, plasma hemoglobin assays. Lower than expected levels may occur if the cells are fixed with glutaraldehyde or formaldehyde or the hemoglobin is precipitated.

The broad absorption band of HiCN in this region enables the use of simple filter type photometers as well as narrow band spectrophotometers for detection. Commercially available blood analyzers have automated quantitations based on either of the two detection strategies. The use of the HiCN reference standard provides comparability among all laboratories employing this method. The major disadvantage is the potential health risk in using the cyanide solutions. Cyano-reagents are themselves toxic by various routes of exposure, and additionally, release HCN upon acidification. Disposal of reagents and products has also become a considerable concern and expense. Spectrophotometric measurement as haemoglobincynide is only practical if the amount of Iysis is > 50mg/l. This method is not recommended.

C. Oxyhemoglobin method

The second classical method is more cumbersome and not widely used today. The oxyhemoglobin method depends on the formation of HbO2 during ammonia-hydroxide treatment, and spectrophotometric detection of this product. No stable reference preparation is available but this is not important because all that the method is required to measure is the percentage of the total hemoglobin in the original specimen which is present in the plasma. In any event, a short-term standard can be prepared from a fresh blood sample.

C. Iron method

The third classical method is based on determining the hemoglobin iron concentration in solution. Iron is first separated from Hb, usually by acid or by aching. It is then titrated with TiCl3 or complexed with a reagent to develop color that can be measured photometrically. This method is too complex for routine work, and is rarely used.

C.5.1.2 Plasma hemaglobin - Tetramethylbenzidene method

Currently, there are as many as twenty different assays for measuring plasma hemoglobin in use today. The assays can be classified into two broad categories: those which are direct optical techniques (i.e. based on quantifying the oxyhemoglobin absorbance peaks at 415, 541, or 577 nm, directly or through use of derivative spectrophotometry) and those which are added chemical techniques (i.e. quantification of hemoglobin based on a chemical reaction with reagents such as benzidine-like chromogens and hydrogen peroxide, of the formation of cyanmethemoglobin) [54]. A popular method for determining the concentration of hemoglobin is based on its catalytic effect on the oxidation of a benzidine derivative, such as tetramethylbenzidene, by hydrogen peroxide. The rate of formation of a colored product (photometrically detected at 600 nm) is directly proportional to the hemoglobin concentration. The advantages of this method are ease of automation (commercial equipment), elimination of potentially toxic and environmentally unsafe cyano-reagents, and the availability of Hb standard sets which are calibrated against the HiCN primary reference standards. The detection limits of the assay (as low as 5.0 mg/dL) are comparable to the hemoglobin cyanide method [52]. The major disadvantages are that there is still a potential health risk in using benzidine dyes and an expense associated with disposal of reagents and products. Moreover, the reported dynamic range of this method is low (5 - 50 mg/dL) [55], and possible reaction inhibition (by as much as 40% ) [56] may occur from calcium chelating anticoagulants (e.g. citrates, oxalates, EDTA) [55], albumin, [57] or other non-specific plasma components [56] which may interfere with H2O2 oxidation. For these reasons, direct optical methods, such as those by Harboe [58], Cripps [59], or Taulier [60] with comparable sensitivity and reproducibility may be substituted.

C.5.2 Blood and blood component preservation

This section presents the best demonstrated practices for the preservation of human blood components by the American Association of Blood Banks [61] and the Council of Europe [62].

Anticoagulant solutions have been developed for use in blood collection that prevent coagulation and permit storage of erythrocytes for a certain interval of time. These solutions all contain sodium citrate, citric acid, and glucose; additionally, some contain adenine, guanosine, mannitol, sucrose, sorbitol and/or phosphate, among others [63], [64], [65], [66], [67], [68].

Blood clotting is prevented by citrate binding of calcium. Red blood cells metabolize glucose during storage. Two molecules of adenosine triphosphate (ATP) are generated by phosphorylation of adenosine diphosphate (ADP) for each glucose molecule metabolized via the Embden-Myerhoff anaerobic glycolysis cycle. The ATP molecules support the energy requirements of the erythrocyte in maintenance of membrane flexibility and certain membrane transport functions. Conversion of ATP to ADP releases the energy necessary to support these functions. In order to prolong storage time, alkalinity must be reduced by addition of citric acid to the anticoagulant solution. This provides a suitably high hydrogen ion concentration at the beginning of red blood cell storage at 4_C. Increasing acidity during storage reduces the rate of glycolysis. The adenosine nucleotides (ATP, ADP, AMP) are depleted during storage and the addition of adenosine to the anticoagulant solution permits synthesis of replacement AMP, ADP, and ATP.

A considerable portion of glucose and adenine is removed with plasma when erythrocyte concentrates are prepared. Sufficient viability of the erythrocytes can only be maintained after removal of plasma if the cells are not overconcentrated. Normal citrate phosphate dextrose (CPD)-adenine erythrocyte concentrates should not have an erythrocyte volume fraction greater than 0.80. Even if more than 90% of the plasma is removed, erythrocyte viability can be maintained by addition of an additive or suspension medium. Sodium chloride, adenine and glucose are necessary for viability while mannitol or sucrose may be used to further stabilize the cell membrane and prevent hemolysis [62].

The suitability of containers for the storage of blood products is evaluated by various methods that measure the quality of the blood product [52], [65]. The container with blood product containing an appropriate anticoagulant is stored upright at 1_C to 6_C under static conditions. At predetermined intervals, the amount of cell-free, plasma hemoglobin is measured to assess the viability and quality of the stored product. The quality of the stored product may be enhanced by gentle mixing once a week. Evaluation of storage in the container indirectly evaluates the permeability of the container to waste carbon dioxide from erythrocyte metabolism in the absence of other confounding factors.

C.5.3 Protection of employees handling blood

Written procedures are necessary for protecting employees receiving, handling, and working with potentially contaminated human blood. Potentially contaminated materials include blood and other body fluids and products, equipment which has been or may have been in contact with blood or other body fluids, and materials used in the culturing of organisms causing blood borne infections [69].

C.5.4 Blood collection (phlebotomy) ..

While it is not possible to guarantee 100% sterility of the skin surface for phlebotomy, a strict, standardized procedure for preparation of the phlebotomy area should exist. It is especially important to allow the antiseptic solution to dry on the skin surface prior to venipuncture and that no further contact is made with the skin surface before the phlebotomy needle has been inserted [61].

A closed container system (i.e. one that does not contain room air) is preferred for blood collection for the prevention of microbial contamination. Needle punctures in the rubber seal of the specimen vial should be completely closed after withdrawal of the needles, otherwise the partial vacuum created following cooling may draw in contaminated air [61].

NOTE - Use of a vacuum tube has the potential to cause slight hemolysis.

Blood collected in an open system may be contaminated by exposure to room air and is not considered sterile. Microbial contamination is a known cause of hemolysis.

C.5.5 Species selection

Ideally, hemolysis testing should be done with human erythrocytes. However, several factors can make such a choice difficult or impossible. In some countries, human blood supplies are limited and must be reserved for human transfusion. Health criteria for human and animal donors should also be considered. All blood has a limited "shelf life" and it may be more difficult to obtain human blood cells on a timely basis. If animal erythrocytes are used, attention should be paid to ensure one hundred percent hemolysis to obtain total hemoglobin content due to differences in membrane stability among animal species. Negative controls should cause minimal hemolysis so that the activity of the test material is not masked. Rabbit and human erythrocytes are reported to have similar hemolytic properties whereas monkey erythrocytes are more sensitive and guinea pig erythrocytes less sensitive [70].

C.5.6 Evaluation of hemolysis - in vitro, ex vivo, and in vivo exposure to blood or blood components

Hemolysis may be evaluated by exposure of materials or devices under in vitro, in vivo, and ex vivo conditions. In vitro conditions are used to evaluate materials as well as devices. Ex vivo and in vivo conditions are used to evaluate devices, which may contain more than one material.

In vivo and ex vivo assessments in animal models or during clinical trials are possible. Justification may be made for either of the following study designs. In the first case, the test device is compared to reference control marketed devices with known acceptable levels of hemolysis. In the second case, the test subject is evaluated for clinically significant consequences of hemolysis.

The purpose of in vivo or ex vivo tests is to characterize the hemolytic potential of a medical device. The preliminary studies may be in vitro and may use fresh or outdated human blood or blood from a nonhuman specie. For medical devices indicated for ex vivo use, the general practice is to recirculate blood through the device using conditions that simulate the intended clinical usage. These investigations are followed by ex vivo simulations in an animal model for some medical devices or by limited, controlled studies in humans. The size of the medical device and the intended function influence the design of these studies.

C.5.7 Direct contact versus indirect methods

Extraction conditions to be used are outlined in ISO 10993-12. Some test methods call for direct contact of the device with red blood cells, while other methods describe the preparation of an extract which is then exposed to red blood cells. Test selection should be based upon the device itself and the conditions in which it will be used. Boundary conditions to be considered when elevated temperatures are used are outlined in ISO 10993-12.

C.6 Recommended hemolysis test methods for medical devices and materials

The following is primarily taken from ASTM F 756-93), which is a standard that is specific for testing the hemolytic properties of materials (mainly due to chemical factors) and is not sufficient for testing whole medical devices.

C.6.1 In vitro methods

The purpose of in vitro tests is to evaluate the effect on isolated erythrocytes. Direct methods determine hemolysis due to physical interactions of red blood cells with either the specimen or the extractables from the test specimen. The dynamic method increases the physical interaction and, presumably, the amount of extraction that occurs. Indirect methods determine hemolysis due to extractables from the test specimen. Indirect methods permit separating the extraction phase from the test system, thereby offering opportunity to exaggerate extraction conditions and distinguish physical surface interactions from chemical effects. [Note: Equipment calibration can be achieved with a Reference Standard for hemoglobin analysis [52]].

C.6.2 Preparation of erythrocyte suspension

Blood from one human donor is normally sufficient. Pooled blood from no more than three animal donors preferably of the same blood type is normally sufficient to minimize variation between donors [70], [71]. A rationale for the selection of blood donor species should be provided in the report.

For the purpose of standardization, whole blood collected in a closed system containing ACD (acid citrate dextrose) or CPD (citrate phosphate dextrose) anticoagulant should be stored no more than 96 hours at 1_C to 6_C.

NOTE - Vigorous mixing of blood should be avoided to prevent air induced hemolysis.

Blood collected in an open system should be used immediately. Heparinized whole blood collected in a closed system should be stored no more than 2 hours at 1 _C to 6_C in order to avoid nutrient depletion. After dilution of whole blood or washed erythrocytes in crystalloid media, they shall be used within 2 hours because the crystalloid media does not have adequate nutrients for erythrocyte preservation. The cell-free plasma hemoglobin content of whole blood or washed erythrocytes shall be less than 1.0 mg/mL to avoid artifacts [61]. [Note: Sterile crystalloid solutions (saline or phosphate buffered saline) are not required for the assays because the duration of the test (less than 4 hours) is insufficient to allow for significant microbial growth unless the sample has been collected in an open system or grossly contaminated during phlebotomy.]

C.6.3 Method for evaluating material in direct contact with diluted blood

Dilute whole blood with isotonic saline (0.9%) to a total hemoglobin concentration of 25.0 +- 2.5 mg/mL.

NOTE - Cooling blood and saline prior to dilution may reduce inter-assay variability.

This diluted suspension shall be used within 2 hours of its preparation when kept at 1 _C to 6_C or in an ice bath.

Tests on each material and blood sample combination should be performed in triplicate for each set of test conditions. Prepare test specimens of material, as described in ISO 10993-12, that have been subjected to the same manufacturing processes and sterilization as the final product. Prepare six tubes containing strips or rods of the test specimens (three tubes for dynamic testing and three tubes for static testing) and three tubes containing strips or rods of the Negative Reference Material.

NOTE - Vigorous mixing of blood should be avoided to prevent air induced hemolysis.

The amount of test specimens and Negative Reference Material should be 1500 mm2 for samples > 0.5 mm thickness and 3000 mm2 for samples < 0.5mm thickness., which is in conformance with the recommendations of ISO 10993-12.

NOTE- The value of 15.7cm2 in ASTM F 756-93 is based on 50 rods of 10 mm x 1 mm x 3.14. The change herein is consistent with ISO 10993-12.

Add 5 mL of diluted blood to each tube. Incubate with the test specimens and Negative Reference Material. Three test preparations are incubated for 4 hours at 37 +- 2_C under static conditions and three are incubated for 1 hour at 37 +- 2_C under dynamic conditions (mixing on a rocker plate at 30 +- 6 rocks per minute at a maximum of 45 degrees from the vertical position).

NOTE - Dynamic mixing conditions should not cause frothing or protein denaturation which could contribute to erythrocyte Iysing.

At the end of the contact period, the test specimens are removed and the diluted blood samples are centrifuged at 100 to 200 G in a standard clinical centrifuge for 15 minutes. Remove and transfer the supernatant fraction to individual siliconized borosilicate glass tubes. Recentrifuge the supernatant fraction at 700 to 800 G in a standard clinical centrifuge for 5 minutes to remove any remaining erythrocytes from the supernatant. Remove and transfer the supernatant into a third borosilicate tube. Analyze the samples for free hemoglobin using the method in annex C.5.1.2

C.6.4 Method for indirect contact with diluted blood

Sufficient saline extracts of the test specimen and Negative Reference Material are prepared in order to perform each test in triplicate. Prepare sufficient replicate tubes containing 4 mL of saline extract and 5 mL of diluted blood with a hemoglobin concentration of 25 +- 2mg/mL for both static and dynamic testing of test and control specimens.

NOTE - Vigorous mixing of blood should be avoided to prevent air induced hemolysis.

Three test preparations are incubated for 4 hours at 37 +- 2_C under static conditions and three are incubated for 1 hour at 37 +- 2_C under dynamic conditions. The Negative Reference Material should also undergo static and dynamic testing.

At the end of the contact period, the test specimens are removed and the diluted blood samples are centrifuged at 100 to 200 G in a standard clinical centrifuge for 15 min. Remove and transfer the supernatant fraction to individual siliconized borosilicate glass tubes. Recentrifuge the supernatant fraction at 700 to 800 G in a standard clinical centrifuge for 5 min. to remove any remaining erythrocytes from the supernatant. Remove and transfer the supernatant into a third borosilicate tube. Analyze the samples for free hemoglobin using the method in annex C.5.1.2

NOTE - It is important that the extracts be isotonic. Non-isotonic extracts may predispose cells to hemolysis.

C6.5 Hemolysis tests for blood containers

An in vitro hemolysis assay which uses an extract of the blood container materials is recommended in the European Pharmacopeia [72] and in ISO Standard 3826 (Plastics Collapsible Containers for Human Blood and Blood Components) [73]. The extracts are prepared at 110_C for 30 min.; these are the sterilization conditions of blood containers with liquid anticoagulants based on sterility assurance testing levels). These assays are focused on the container materials and do not assess gas permeability or other quality assurance parameters essential for storage of whole blood and blood products (see annex C.5.2 for further discussion of this subject).

Annex D (informative) Bibliography

D.1 International Standards

[1] ISO 5840:1989, Cardiovascular implants - Cardiac valve prostheses.
[2] ISO 5841 - 1: 1989, Cardiac pacemakers - Part 1: Implantable pacemakers.
[3] ISO 5841-3:1992, Cardiac pacemakers - Part 3: Low-profile connectors (IS-1) for implantable pacemakers.
[4] ISO 7198-1:-1), Cardiovascular implants - Tubular vascular prostheses - Part 1: Synthetic vascular prostheses.
[5] ISO 7198-2-1), Cardiovascular implants - Tubular vascular prostheses - Part 2. Sterile vascular prostheses of biological origin - Specification and methods of test.
[6] ISO 7199:-1), Cardiovascular implants and artificial organs - Blood-gas exchangers.
[7] ISO 10993-12:-1), Biological evaluation of medical devices - Part 12: Sample preparation and reference materials.

D.2 National standards

[8] AANSI/AAMI CVP3- 1981, Cardiac valve prostheses.
[9] ANSI/AAMI VP20- 1986, Vascular graft prostheses.
[10] ASTM F 756-87, Standard Practice for assessment of hemolytic properties of materials.
[11] BS 5736-11:1990, Evaluation of medical devices for biological hazards 3/4 Part 11: Method of test for hemolysis.
[12] DIN 58 361-4: 1980, Transfusionsbehaltnisse und Zubehor, Blutbeutel aus Kunststoffen, Sicherheitstechnische Anforderungen, Prufung, Uberwachung und Kennzeichnung.
[13] NF 90-300: 1981, Materiel medico-chirurgical - Oxygenateurs.

D.3 Scientific papers

[14] BOSCH, T., SCHMIDT, B., BLUMENSTEIN, M. and GURLAND, H.J., Thrombogenicity markers in clinical and ex vivo assessment of membrane biocompatibility. Contr Nephrol, 59: 90-98, 1987.
[15] CHENOWETH, D.E., Complement activation produced by biomaterials. Trans Am Soc Artif Int Organs 32: 226232, 1986.
[16] BURNS, G.L. PANTALOS, G.M. and OLSEN, D.B., The calf as a model for thromboembolic events with the total artificial heart. Trans Am Soc Artif Int Organs 33: 398-403, 1987.
[17] COOPER, S.L., FABRIZIUS, D.J. and GRASEL, T.G., Methods of assessment of thrombosis ex vivo. In: Leonard E.F., Turitto V.T., and Vroman L.(Eds): Blood in contact with natural and artificial surfaces. Ann N.Y. Acad Sciences 516:572-585, 1987.
[18] Corriveau, D.M. and Fritsma, G.A. (Eds): Hemostasis and thrombosis in the clinical laboratory. J.B. Lippincott, Philadelphia, 1988, 443 pp.
[19] DAWIDS, S et al. (Eds): Test procedures for the blood compatibility of biomaterials. In preparation, 1992.
[20] DEWANJEE, M.K., Methods of assessment of thrombosis in vivo, In: Leonard E.F., Turitto V.T. and Vroman L. (Eds): Blood in contact with natural and artificial surfaces. Ann N.Y. Acad Sciences, 516:541-571, 1987.
[21] DEWANJEE, M.K., KAPADVANJWALA, M. and SANCHEZ, A., Quantitation of comparative thrombogenicity of dog, pig, and human platelets in a hemodialyzer. Am Soc Artif Int Organs Journal 38: 88-90, 1992.
[22] DIDISHEIM, P., DEWANJEE, M.K., FRISK, C.S., KAYE, M.P. and FASS, D.N., Animal models for predicting clinical performance of biomaterials for cardiovascular use. In: Boretos J.W. and Eden M. (Eds): Contemporary Biomaterials, Noyes Publications, Park Ridge, NJ, USA, pp. 132-179, 1984.
[23] DIDISHEIM, P., DEWANJEE, M.K., KAYE, M.P., FRISK, C.S., FASS, D.N., TIRRELL, M.V. and ZOLLMAN, P.E., Nonpredictability of long-term in vivo response from short-term in vitro or ex vivo blood/material interactions. Trans Am Soc Artif Int Organs 30: 370-376, 1984.
[24] DIDISHEIM, P., OLSEN, D.B., FARRAR, D.J., PORTNER, P.M., GRIFFITH, B.P., PENN1NGTON, D.G., JOIST, J.H. , SCHOEN, J.F. , GRIST1NA, A. G. and ANDERSON, J.M., Infections and thromboembolism with implantable cardiovascular devices. Trans Am Soc Artif Int Organs 35: 54-70, 1989.
[25] DIDISHEIM, P., STROPP, J.Q., BOROWICK, J.H. and GRABOWSKI, E.F., Species differences in platelet adhesion to biomaterials: investigation by a two-stage technique, Trans Am Artif Int Organs 2: 124-132, 1979.
[26] Guidelines for blood/material interactions. Report of the National Heart, Lung and Blood Institute Working Group, U.S. Department of Health and Human Services, Public Health Service, National Institutes of Health. NIH Publication No. 85-2185, revised September 1985. Available from Biomaterials Program, Devices and Technology Branch, Division of Heart and Vascular Diseases, NHLBI, 312 Federal Building, 7550 Wisconsin Avenue, Bethesda, MD 20892, USA; phone 301-496-1586, FAX 301-480-6282.
[27] MARKER, L.A., KELLY, A.B. and HANSON, S.R., Experimental arterial thrombosis in nonhuman primates. Circulation 83. Supplement IV, 41-55,1991.
[28] MARKER, L.A., MALPASS, T.W., BRANSON, H.E., HESSEL, E.A. II and SLIGHTER, S.J., Mechanism of abnormal bleeding in patients undergoing cardiopulmonary bypass: Acquired transient platelet dysfunction associated with selective alpha granule release. Blood 56: 824834, 1980.
[29] MARKER, L.A., RATNER. B.D., and DIDISHEIM, P. (Eds): Cardiovascular Biomaterials and Biocompatibility. Supplement to Cardiovascular Pathology. Elsevier, New York, in press 1992.
[30] KARWATH, R., SCHuRER, M. and WOLF, H., Measurement of platelet adhesiveness onto artificial surfaces using Cr-51 and In-111 labeled platelets. Studia Biophysica 131: 117-123, 1989.
[31] KATAOKA, K., MAEDA, M., NISHIMURA, T., NITADORI, Y., TSURUTA, T., AKAIKE, T., and SAKURAI, Y., Estimation of cell adhesion on polymer surfaces with the use of "column method". J. Biomed Mat Res 14: 817-823, 1980.
[32] KAY, L.A., Essentials of Hemostasis and Thrombosis. Churchill Livingstone, Edinburgh, 1988, 290 pp.
[33] LEWIS, J.L., SWEENEY, J., BALDINI, L., FRIEDLAND, G.H., and SALZMAN, E.W., Assessment of thromboresistance of intravenous cannulae by 125 I-fibrinogen scanning. J Biomed Mat Res 19: 99, 1985.
[34] MAHIOUT, A., MEINHOLD, H., JORRES, A., KRIEG, R., KESSEL., M., TRETZEL, J. and BAURMEISTER, U., Ex vivo model for preclinical evaluation of dialyzers containing new membranes. Life Support Systems 3: suppl 1, 448-452, 1985.
[35] MIALE, J.B., Laboratory medicine hematology. Sixth edition, CV Mosby, St. Louis, 1982.
[36] PALATIANOS, G.M., DEWANJEE, M.K., and ROBINSON, R.P., et al, Quantitation of platelet loss with Indium-111 labeled platelets in a hollow-fiber membrane oxygenator and arterial filter during extracorporeal circulation in a pig model. Trans Am Soc Artif Int Organs 35: 667-67O, 1989.
[37] SCHOEN, F.J., ANDERSON, J.M., DIDISHEIM, P., DOBBINS, J.J., GRISTINA, A.G., HARASAKI, H., and SIMMONS, R.L., Ventricular assist device (VAD) pathology analyses: guidelines for clinical studies. J Applied Materials 1: 49-56, 1990.
[38] SCHOEN, F.J., Interventional and surgical cardiovascular pathology. Appendix: Pathologic analysis of the cardiovascular system and prosthetic devices, pp 369-396. W.B. Saunders Co., Phildelphia, 1989.
[39] SPENCER, P.C., SCHMIDT, B., SAMTLEBEN, W., BOSCH, T. and GURLAND, H.J., Ex vivo model of hemodialysis membrane biocompatibility. Trans Am Soc Artif Int Organs 31: 495-498, 1985.
[40] Tripartite Biocompatibility Guidance for Medical Devices. Prepared by toxicology sub-group of the Tripartite Subcommittee on medical devices, September 1986.
[41] WARD, R.A., SCHMIDT, B., BLUMENSTEIN, M., and GURLAND, H.J. Evaluation of phagocytic cell function in an ex vivo model of hemodialysis. Kidney International 37: 776782, 1990.
[42] WHITE, R.A., Diagnosis and therapy of emergent vascular diseases. In: Shoemaker, W.C., Ayres, S., Holbrook, P.R., and Thompson, W.L., Texthook of critical care, second edition. W.B. Saunders, Philadelphia, 1989, pp. 447-452.
[43] ZINGG, W., IP, W.F., SEFTON, M.V., and MANGER, K., A chronic arteriovenous shunt for the testing of biomaterials and devices in dogs. Life Support Systems 4: 221 -229, 1986.
[44] WHO/LAB/88.3, Recommended methods for the visual determination of white cell and platelet count, S.M. Lewis, R.M. Rowan and F. Kubota, J. Clin. Pathol. 43: 1990 932-936.

D.4 Scientific Papers referenced in Amendment I, Annex C - Evaluation of hemolytic properties of medical devices

[45] Taber's Cyclopedic Medical Dictionary, 17th Edition, F.A. Davis Company, Philadelphia, PA, 1993.
[46] Taber's Cyclopedic Medical Dictionary, 17th Edition, F.A. Davis Company, Philadelphia, PA, 1993.
[47] Mueller, M.R., Schima, H., Engelhardt, H., Salat, A., Olsen, D.B., Losert, U., Wolner, E., In Vitro Hematological Testing of Rotary Blood Pumps: Remarks on Standardization and Data Interpretation, Artificial Organs, 17(2): 102-110, 1193.
[48] Hoch, J.R., Silver, D., Hemostasis and Thrombosis, in "Vascular Surgery: A Comprehensive Review," 3rd edition, W.S. Moore, ea., W.B. Saunders, Philadelphia, 63-79, 1991.
[49] Offeman, R.D., Williams, M.C., Material effects in shear-induced hemolysis, Biomat. Med. Dev. Art. Org., 7:359391, 1979.
[50] Lampert, R.H., Williams, M.C., Effect of surface materials on shear-induced hemolysis, J. Biomed. Mater. Res., 6:499-532, 1972.
[51] Obeng, E.K., Cadwallader, D.E., In vitro dynamic method for evaluating the hemolytic potential of intravenous solutions. J. Parenteral Sci. Technol., 43: 167- 173, 1989.
[52] Henry, J.B., Hematology and Coagulation, In: Clinical Diagnosis & Management By Laboratory Methods, 18th Ed., W.B. Saunders Co., Philadelphia, PA, USA, 556-603, 1991.
[53] International Committee for Standardization in Haematology (ICSH): Recommendations for reference method for hemoglobinometry in human blood. (ICSH Standard EP 6/2: 1995) and specifications for international hemiglobin cyanide standard (4th edition) J. Clin. Path., 49:279-274, 1996.
[54] Malinauskas, R.A. Plasma hemoglobin measurement techniques for the in vitro evaluation of blood damage caused by medical devices (Currently under journal review).
[55] Sigma Diagnostics - Plasma Hemoglobin: Quantitative, colorimetric determination in plasma at 600nm (Procedure No. 527, April 1991) Sigma Diagnostics, St. Louis, MO.
[56] Standefer, J.C., Vanderjagt, D. Use of tetramethylbenzidine in plasma hemoglobin assay. Clin. Chem. 1977; 23 :749-751.
[57] Fairbanks, V.F., Ziesmer, S.C., O'Brien, P.C. Methods for measuring plasma hemoglobin in micromolar concentration compared. Clin. Chem. 1992; 38: 132-140.
[58] Harboe, M., A method for determination of hemoglobin in plasma by near-ultraviolet spectrophotometry, Scand. J. Clin. Lab Invest., 1959; 11 :66-70.
[59] Cripps, C.M. Rapid method for the estimation of plasma haemoglobin levels. J. Clin. Path., 1968: 21:110-112.
[60] Taulier, A. [Value of derivative spectrophotometry for the determination of plasma and urinary hemoglobin. Comparison with the method using Allen's correction]. Annales de Biologie Clinique,(Paris) 44:242-248 (1986). French
[61] Standards for Blood Banks and Transfusion Services, 16th ed. Bethesda, MD: American Association of Blood Banks, 1994.
[62] Guide to the Preparation, Use and Quality Assurance of Blood Components. Strasbourg: Council of Europe Publishing and Documentation Service, 1992, pp. 37-38.
[63] Anticoagulant Citrate Dextrose Solution. U.S. Pharmacopeia 23:119 (1995).
[64] Anticoagulant Acid Citrate Dextrose Solutions (ACD). European Pharmacopoeia 2:209-210 (1989)
[65] Anticoagulant Citrate Phosphate Dextrose Solution. U.S. Pharmacopeia 23: 119-120 (1995).
[66] Anticoagulant Citrate Phosphate Dextrose Adenine Solution, U.S. Pharmacopeia 23:121-122 (1995)
[67] Anticoagulant Heparin Solution. U.S. Pharmacopeia 23: 122 (1995).
[68] Anticoagulant Sodium Citrate Solution. U.S. Pharmacopeia 23:122 (1995).
[69] 29 CFR, Code of Federal Regulations - 1910. 1030, "Bloodborne Pathogens."
[70] Wennberg, A., Hensten-Pettersen, A., Sensitivity of erythrocytes from various species to in vitro hemolyzation, J. Biomed. Mater. Res., 15:433-435, 1981.
[71] A Color Atlas of Comparative, Diagnostic and Experimental Hematology, Mosby-Year Book Europe Limited, London, England, 1994.
[72] European Pharmacopeia VI., Sterile Plastic Containers for Human Blood and Blood Components.
[73] ISO Standard 3826 (Plastics Collapsible Containers for Human Blood and Blood Components).