Biocompatibility Medical device toxicology and risk assessment

1
BiocompatibilityMedical device toxicology and
risk assessment

• Ron Brown (rpb_at_cdrh.fda.gov)
• Toxicologist
• Health Sciences Branch
• U.S. Food and Drug Administration
• The views presented in this lecture are strictly
  those of the presenter and do not necessarily
  represent FDA policy

2
Study techniques
3
Biocompatibility Overview

• Endpoints of concern
• Developing a testing strategy
• Controversies/new issues

4
Risk assessment overview

• What do we mean by risk assessment? How do we
  perform a risk assessment?
• How can this approach be useful in a biological
  assessment of a device?
• Controversies/new issues

5
Biocompatibility

• State of a material in a physiological
  environment, without the material adversely
  affecting the tissue or the tissue adversely
  affecting the material
• Property of a device or specific material used in
  the device which shows no toxicity when it is
  used as intended
• Simple definition What the device does to the
  body and what the body does to the device

6
How do we test devices for biocompatibility?

• Process Described in Consensus Standards and FDA
  Guidance documents
• ISO 10993 standards
• ASTM standards
• FDA Guidance documents

7
These review papers may come in handy, because...
8
FDA/ISO Biocompatibility Testing Scheme

• FDA-modified ISO 10993-1 http//www.fda.gov/cdrh/g
  951.html
• See handout
• Is this the best way to evaluate
  biocompatibility?
• Lets look at some specific issues then come back
  to this question

9
Sample Preparation

• Many tests (cytotoxicity, irritation,
  mutagenicity, systemic toxicity) are conducted
  with extracts of the device materials.
• What do we need to know about the device?
  Material composition, device geometry, duration
  of contact, service environment
• Information will be used to devise an extraction
  protocol

10
Extraction protocol

• Choice of solvent
• Extraction of polymers should (at least) employ a
  two-solvent system
• Polar (saline) and nonpolar (e.g. vegetable oil)
  
• 10993-12, Section 10.2.1.2
• The choice of solvents shall be justified
  based on consideration of maximally extracting
  the material or device under conditions that
  mimic final use.

11
Default extraction conditions in ISO 10993-12

• 37oC for 24 hours
• 37oC for 72 hours
• 50oC for 72 hours
• 70oC for 24 hours
• 121oC for 1 hour
• Are these conditions intended to be exhaustive,
  aggressive or clinically relevant? Are they
  exhaustive for some materials but not others?
• How do you choose one default over another? Are
  these conditions intended to be equivalent?

12
What extraction conditions are appropriate for
the tests identified in 10993-1?

• 10993-12 instructs the user to match extraction
  conditions to the purpose of the test.
• Many of the tests specified in 10993-1 are
  screening tests.
• Screening tests should be sensitive, therefore
  they would require aggressive, but not
  necessarily exhaustive extraction conditions.

13
Need to be careful about semantics

• The terms exhaustive, destructive, aggressive,
  exaggerated, accelerated and rigorous are being
  used interchangeably.
• Generally - destructive, exhaustive gt
  aggressive, exaggerated, accelerated gt
  clinically relevant, simulated use.
• However, for some devices (e.g., implants),
  exhaustive may be equivalent to clinically
  relevant or simulated use. Destructive may be
  equivalent to clinically relevant and simulated
  use for absorbables.

14
Proposal on Terminology
15
Why do we need aggressive extraction conditions
for screening tests?

• No uncertainty factors are applied in screening
  approach, but results are assumed to be directly
  applicable to patients
• - assumes equivalent sensitivity between
  humans
• and experimental animals
• - uses healthy animals
• Increase sensitivity of the test when small
  sample sizes are used
• Endpoints in some tests are insensitive (e.g.,
  death, convulsions in acute systemic toxicity
  test)
• Screening approach does not account for exposure
  to multiple devices

16
Discussion about extraction conditions points out
the need for an integrated approach for
biological assessment

• Interpretation of test results is dependent on
  extraction conditions, extraction conditions
  affect downstream decision-making.
• Pass-fail criteria, where they exist, are only
  valid for specific sample preparation techniques
  and test methods. Change in the sample prep
  changes the way the test should be interpreted.
• Points out the need for overall guidance in how
  to conduct a biological evaluation

17
Sample preparation vs. chemical characterization

• Screening tests for hazard identification should
  be done using extracts obtained using aggressive,
  but not exhaustive conditions.
• Clinically relevant extraction conditions should
  be used when a chemical characterization/risk
  assessment approach is used to assess systemic
  toxicity.

18
Cytotoxicity

• Usually first step in biocompatibility testing
• Usually done with fibroblasts - Why?
• Sensitive to toxic substances, easy to
• grow in the lab
• How are the tests performed?
• 1. Direct contact - materials directly in
  contact with
• monolayer of cells
• 2. Agar overlay - material in contact with
  agar on top of
• monolayer
• 3. Extract in media or on disc in agar
  overlay

19
Cytotoxicity test results

• FDA almost always sees negative results. Why?
  Most toxic materials never make it through
  screening process and studies are conducted with
  weak extracts
• Positives seen with materials that polymerize in
  situ (e.g., bone cements)
• What are the tests predictive of? Local effects?
  Systemic effects?
• Can we make better use of cytotoxicity test
  results?

20
Correlation between cytotoxicity and irritation

• Good correlation b/w cytotoxicity indices and
  thickness of inflammatory layer in implantation
  test
• Nakamura et al. (1990) Biomaterials 1192-94
• Tsuchiya (1993) J. Appl.Biomat 4(2) 9153-156

21
Correlation b/w cytotoxicity and systemic toxicity

• Considerable amount of work going on under
  auspices of ICCVAM, etc. to replace LD50
• May be possible to replace unnecessary systemic
  toxicity of devices testing with alternative
  methods.
• Important distinction - unnecessary testing vs.
  all testing

22
Use of cytotoxicity data to predict systemic
toxicity

• How well do cytotoxicity data predict systemic
  toxicity endpoints? See handouts.
• Other studies - Kjellstrand et al. (Cell Biol.
  Toxicol. 10137-142, 1994). 851 in vitro/in vivo
  tests, no false negatives
• Thus, among the tests on living systems, the
  cell test alone seems to be sensitive enough to
  provide sufficient information. Nothing appears
  to be gained from the in vivo tests

23
Not a new proposal

• Since the systemic and mouse safety tests
  appear to have very limited value in predicting
  acute toxicity, the bacterial luminescence test,
  when validated to the mouse safety test and USP
  toxicity tests for containers and medical devices
  on a product-by-product basis, could reduce and
  in some cases replace in vivo tests.
• Although it would be naïve to think that the
  bacterial luminescence test or another in vitro
  assay will replace every in vivo acute toxicity
  test, the use of in vitro acute toxicity testing
  should be reevaluated so that the practical
  shortcomings of several of the in vivo tests can
  be reevaluated (Burton et al., 1986).

24
Why do we want to replace the systemic toxicity
tests anyway?

• Insensitive endpoints (behavioral changes,
  dyspnea, death)
• Relatively small number of animals
• Uses young, healthy animals
• Assumes single exposure
• No safety factor
• Extraction conditions are not necessarily
  rigorous, especially when cell culture media is
  used for extraction

25
What have others said about the utility of the
acute toxicity test?

• The acute systemic toxicity test is considered to
  be of very little importance for the safety
  assessment of insoluble medical devices, since
  such materials leach only minute amounts of their
  constituents and these very seldom reach levels
  which cause any acute effects (21,22). For other
  types of devices which are partly soluble or can
  disintegrate (for example, hydrocolloidal wound
  dressings), the outcome of the test is generally
  systemic collapse of the animals due to the
  viscosity, particulate nature, or other aspects
  of the extract/solution. Nevertheless, it is
  recognised that the test may be relevant for
  certain materials.
• Svensen et al. (1996) ATLA 24659-669

26
What have others said about the utility of the
acute toxicity test?

• Compared to in vitro tests, many in vivo acute
  toxicity tests are poor indicators of toxicity.
  Data from in vivo tests tend to be imprecise and
  sometimes of limited usefulness. For example,
  the mouse systemic toxicity test is simple and
  fast but is insensitive16,24,25. The mouse
  safety test is likewise insensitive.
• Burton et al. (1986)

27
If the acute systemic toxicity test is
inappropriate, then which approach is better?

• Improve acute systemic toxicity test
• Consider alternative methods that provide the
  same or better information to assess the
  potential for a device to cause adverse systemic
  effects in patients

28
Advantages of using cytotoxicity data as a screen
for doing systemic toxicity tests

• Clear evaluation criteria
• Can be designed to assess multiple toxicological
  endpoints and the use of human tissue (if
  necessary)
• Likely to be sensitive
• Can always default to in vivo testing if
  preferred. User is not forced to use.
• Takes advantage of existing data
• More relevant data for regulatory decision making
• Fewer animals, less cost

29
Limitations of using cytotoxicity data as a
screen for doing systemic toxicity tests

• Still have questions about in vivo relevance of
  in vitro data
• Not validated for extracts obtained from devices
• Appropriate for mixtures?
• Doesnt address all systemic endpoints e.g,
  effects of particles, immunotoxicity,
  carcinogenicity

30
Implications for ISO 10993 Standards

• Supports the concept of tiered testing and
  hierarchical decision making
• Could result in reduced testing and better data
  for decision making
• Reinforces need for chemical characterization
• Reinforces the need for evaluation criteria and
  use of SRMs in standards
• Could result in dramatic changes in the format of
  10993-1

31
Implantation Test

• Assess effect of device on tissue and tissue on
  the device
• Physical characteristics (e.g., surface finish,
  charge, etc.) can affect response.
• Muscle often site of implantation, but preferable
  to implant in same location as the intended
  clinical use.
• Response compared to that elicited by material
  with long hx of safe clinical use

32
Implantation test

• Issue When should systemic toxicity tests be
  incorporated into implantation test protocol?
  (see handout)

33
Neurological implants

• No ISO standard, but there is an FDA Guidance
  document http//www.fda.gov/cdrh/ode/627.html
• Ideally, assess a broad range of effects
  morphological, electrophysiological,
  neurochemical, behavioral
• For compounds released, cant easily compare
  neurotoxic effects following intracranial vs.
  systemic exposure

34
Genotoxicity and carcinogenicity testing

• Genotoxicity - Damage to somatic and germ cells
• Testing strategy - use battery of tests one
  bacterial assay (e.g., Ames) and 2 mammalian cell
  assays (e.g., chromosomal aberration and mouse
  lymphoma)
• What do positive results mean? Increased
  potential for carcinogenicity or heritable
  mutations? What do you do when you get genotox
  results?

35
Carcinogenicity testing

• Theoretically, compounds released from devices
  could have a carcinogenic effect
• Testing may be recommended for lifetime implants,
  new materials, materials with genotox test
  results
• However, disconnect b/w 10993-1 and reality -
  testing is rarely done
• Solid state carcinogenesis - results not
  applicable to humans

36
Carcinogenicity of medical devices

• Current issue National Toxicology Programs
  listing of nickel as"reasonably anticipated to be
  a human carcinogen" http//www.niehs.nih.gov/oc/ne
  ws/rocrslt.htm
• Implications for stainless steel orthopedic
  implants - labeling? Calls for removal?
• Important to consider form of nickel. Nickel
  alloys used in orthopedic devices are not
  carcinogenic

37
Immunotoxicity

• Sensitization assessed for essentially every
  device
• Other tests/endpoints as recommended by CDRH
  Immunotoxicity Testing Guidance
  http//www.fda.gov/cdrh/ost/ostggp/immunotox.html
• Interpretation of results?

38
Systemic toxicity

• Systemic toxicity testing is conducted according
  to ISO 10993-11, typically with an extract of the
  device.
• From 10993-11 It must be borne in mind that
  subchronic and/or chronic toxicity testing is not
  always necessary for a risk assessment. Such
  assessment might be made on the basis of
  qualitative and quantitative analytical
  measurements to evaluate the exposure of possible
  leachables from the device.

39
Systemic toxicity

• Also from 10993-1 Submission contains
  acceptable tox data and/or justification or risk
  assessment for not conducting appropriate tests.
• Need exists for more explicit guidance on how to
  conduct the chemical characterization/risk
  assessment approach for assessing systemic
  toxicity potential

40
Overview of proposed approach

• What is a risk assessment-based approach?
• What are the advantages of using this approach
  for the to the submitter and FDA?
• What are the limitations of this approach?
• What has FDA done to make use of the approach
  more practical?

41
Risk Assessment-Based Approach for the Evaluation
of Systemic Toxicity

• Risk assessment - simply a framework for
  evaluating data.
• Compare the dose of the compound(s) received by a
  patient to an upper-bound dose not expected to
  result in adverse effects in patients (Tolerable
  Intake or TI).
• If dose to patient is lt TI, then it is unlikely
  that adverse systemic effects will be seen in the
  patient. Consequently, theres no need to
  conduct systemic toxicity testing.

42
How do you do this?

• How do you estimate dose to patient? Chemical
  characterization standard (ISO 10993-18)
• How do you derive a TI value?
• Risk assessment standard
• (ISO 10993-17)

43
Approach is based on ISO standards
44
Comparing dose-to-patient to the TI Value
45
How do you derive a TI value?

• For noncancer effects
• TI (mg/kg/day) NOAEL or LOAEL/Uncertainty
  Factors
• NOAEL and LOAEL - from toxicity studies
• Uncertainty factors - accounts for differences in
  potency b/w animals and humans and for
  interindividual variability in humans

46
Hypothetical Dose-Response Relationship
for Compromised vs. General Populations
Comparison of Mean of General Population to 5th
Percentile of Compromised Population
General Population
Compromised Population
Number of Individuals Responding
5th Percentile
Mean
Dose
47
How to obtain toxicity data for derivation of TI
values

• National Library of Medicine databases
• PubMed www.ncbi.nlm.nih.gov/PubMed
• TOXLINE igm.nlm.nih.gov
• TOXNET toxnet.nlm.nih.gov
• Commercial databases
• SciSearch, EMBASE, BIOSIS, etc.
• Internet Search Engines

48
Check the literature first!
49
Derivation of TI Values

• ISO 10993-17 works best when data are available
  from clinically relevant routes and durations of
  exposure.
• Practical problem Relevant toxicity data are
  unavailable to derive TI values for many
  compounds released from medical device materials.
• FDA and stakeholders are developing new risk
  assessment approaches for deriving TI values in
  the absence of relevant toxicity data.

50
Access to existing health-based exposure limit
values

• EPA Reference Dose (RfD) Values
• IRIS database www.epa.gov/iris
• TERA ITER database
• EPA RfDs, ATSDR MRLs, Health Canada ADIs
• www.tera.org/iter
• CalEPA OEHHA No Significant Risk Level (NSRL)
  Values www.oehha.org/prop65/pdf/tb0194r2.pdf
• Use with caution, not always relevant for medical
  device exposures

51
Concept under developmentThreshold of
toxicological concern

• Threshold - Dose of any compound, released from a
  device, that is thought to be toxicologically
  insignificant
• No need for systemic toxicity testing if compound
  is released from the device at a dose lt threshold
• Thresholds should be science-based but practical
  current proposal 0.1 mg/day
• Concept under development, not currently used by
  CDRH.

52
Concept under development Use of quantitative
cytotoxicity data

• Correlation exists between IC50 from cytotoxicity
  studies and systemic toxicity LOAEL values
• Derive regression equations to predict IC50 from
  LOAEL
• Systemic toxicity testing threshold 0.1 mg/day is
  equivalent to a given IC50 in various studies
• Therefore, if IC50 gt a given value, then no need
  for systemic toxicity testing
• Concept under development, not currently used by
  CDRH.

53
Advantages of the risk assessment-based approach

• Potentially limits the amount of testing
  necessary. Can be done as a paper exercise,
  therefore, potentially reduces product
  development costs/time
• Reduces animal usage
• Flexible, no rigid evaluation criteria

54
What happens if a device fails using the risk
assessment-based approach?

• Conclude that device is unacceptable for intended
  use
• Examine dose/TI ratio, not a bright line
  evaluation criterion
• Reevaluate extraction data
• Obtain relevant toxicity data
• Conduct risk-benefit analysis
• Test extract according to FDA-modified
• ISO 10993-1

55
Can also be used as an adjunct to testing of an
extract

• Focuses preclinical testing
• If you know which constituents are being released
  from the device and the toxicity of those
  compounds, helps to predict which toxicological
  endpoints may be important
• Can be used to design preclinical studies

56
When is this Approach Most Likely to be Used?

• When there are no issues regarding the
  proprietary nature of the material
• When only one or a few chemical constituents are
  changed in a device
• When toxicity data are readily available on the
  compound(s)
• When extraction/analytical chemistry studies are
  easily conducted

57
Biocompatibility testing strategy

• Lets reexamine existing biocompatibility testing
  strategy. Is there a better way?
• Need for evaluation criteria
• Need for hierarchical testing
• How to use data in a biological evaluation
• How to use chemical characterization data

58
Risk assessment issues relevant to medical device
toxicology

• Development of new, simple approaches
• How to apportion TI over multiple devices
• How to take benefit and feasibility into account
• Radiation/microbial risk assessment (see Gaylor
  et al. paper)
• Growing acceptance of the Precautionary Principle

59
Precautionary principle

• Definition When an activity raises threats of
  harm to human health or the environment,
  precautionary measures should be taken even if
  some cause and effect relationships are not fully
  established scientifically".
• Implications for medical device risk assessment?
• - DEHP in PVC medical devices
• - Soy oil- or guar gum-filled breast implants