Available online at www.sciencedirect.com

Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 167–171www.nanomedjournal.com

Short Communication: Toxicology

Nanotoxicology and nanomedicine: making hard decisionsIgor Linkov, PhD,a,⁎ F. Kyle Satterstrom, MA,b Lisa M. Corey, MSc

aUS Army Engineer Research and Development Center, Brookline, Massachusetts, USAbHarvard University School of Engineering and Applied Sciences, Cambridge, Massachusetts, USA

cIntertox Inc., Seattle, Washington, USA

Abstract Current nanomaterial research is focused on the medical applications of nanotechnology, whereas

Received 15 AprilThis study was su

and Development Cen⁎Corresponding a

ment Center, 83 Win02446, USA.

E-mail address: ig

Please cite this artic2008;4:167-171, doi

1549-9634/$ – see frodoi:10.1016/j.nano.20

side effects associated with nanotechnology use, especially the environmental impacts, are not takeninto consideration during the engineering process. Nanomedical users and developers are faced withthe challenge of balancing the medical and societal benefits and risks associated withnanotechnology. The adequacy of available tools, such as physiologically-based pharmacokineticmodeling or predictive structure-activity relationships, in assessing the toxicity and risk associatedwith specific nanomaterials is unknown. Successful development of future nanomedical devices andpharmaceuticals thus requires a consolidated information base to select the optimal nanomaterial in agiven situation—understanding the toxicology and potential side effects associated with candidatematerials for medical applications, understanding product life cycle, and communicating effectivelywith personnel, stakeholders, and regulators. This can be achieved through an innovativecombination of toxicology, risk assessment modeling, and tools developed in the field ofmulticriteria decision analysis (MCDA).Published by Elsevier Inc.

Key words: Nanotoxicology; Risk assessment; Multicriteria decision analysis

Nanomaterials have the potential to revolutionize medi-cine because of their ability to affect organs and tissues at themolecular and cellular levels. Current research is focused onthe medical applications of nanotechnology, whereas sideeffects associated with their use, especially the environ-mental impacts of their manufacture and disposal, aregenerally not taken into consideration during the engineeringprocess. Incorporating environmental concerns into nano-material engineering and nanomedicine development isimportant, but it greatly increases decision complexity.

Even though the risk assessment paradigm successfullyused by the scientific community since the early 1980s maybe generally useful, its application to nanomaterials would

2007; accepted 28 January 2008.pported in parts by the US Army Engineer Researchter.uthor. US Army Engineer Research and Develop-chester Street, Suite 1, Brookline, Massachusetts

or.linkov@usace.army.mil (I. Linkov).

le as: I. Linkov, F.K. Satterstrom, L. Corey, Nanotoxico:10.1016/j.nano.2008.01.001.

nt matter. Published by Elsevier Inc.08.01.001

require allowing for an uncertainty in basic knowledgethat is much larger than the uncertainty for other materialsand pharmaceuticals. To combat the uncertainty, decisionmakers need an understanding of product life cycle and theability to communicate effectively with personnel, stake-holders, and regulators. This can be achieved through aninnovative combination of toxicology, risk assessmentmodeling, and tools developed in the field of multicriteriadecision analysis (MCDA).

Biomedical community needs

Nanomaterials have been promoted as a revolutionarytechnology for cell and tissue engineering, medical devicedevelopment, and the encapsulation and delivery of drugs,diagnostics, and genes. Advances in nanotechnology have ledto the introduction of many nanomaterials in these areas, andthe Nanomedicine Initiative of the National Institutes of HealthRoadmap for Medical Research initiative predicts thatnanomaterials will begin yielding significant medical benefitswithin the next 10 years.

logy and nanomedicine: making hard decisions. Nanomedicine: NBM

168 I. Linkov et al / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 167–171

Despite the widespread use of nanomaterials, under-standing of the toxicity and potential health risks associatedwith nanomaterial use is extremely limited. In fact, toxicityissues related to nanomaterials used in nanomedicine areoften ignored.1,2 Thus, along with the development ofnovel nanoparticles, experts in related scientific fields arecalling for a simultaneous assessment of the toxicologicaland environmental effects of nanoparticles.3 Recent in vivoand in vitro studies have suggested that inhalation anddermal absorption of some nanomaterials may have adversehealth effects,3,4 and the use of medical products containingnanomaterials may lead to chronic health risks.5 Spurred bysuch reports, regulatory agencies, as well as the popular andscientific media, are shifting their focus from the initialeuphoria about the potential of the technology to concernabout possible deleterious effects resulting from nano-material manufacture and use.

The US Environmental Protection Agency (EPA) hasraised concerns about the use of nanosilver in severalconsumer products already on the market. Uncertaintyabout the health impacts associated with nanotechnologiesand their potentially uncontrolled market growth hasresulted in calls from environmental and political bodies tolimit the use of nanomaterials, increase the stringencyof governmental regulations, and, in extreme cases, ban theuse of nanomaterials completely. A better understanding ofthese materials is clearly needed, yet experience withinorganic and organic chemicals may not be directly relevantto nanomaterials, in that their physical and biologicalproperties are often determined by novel relationshipsbetween their size, structure, and the presence of addedfunctional groups.

A framework of underlying questions remains tobe addressed:

• What are the specific nanomaterial properties thatshould be characterized for nanomedical applications?

• What data are available on nanomedicine toxicity,exposure, and environmental fate and transport?

• Where are the data gaps?• How do nanomaterial characteristics contribute totoxicity in relation to nanomedical applications?

• How do specific delivery mechanisms influencenanomedicine toxicity?

• What is the role of concurrent exposures to multiplenanomedicines and pharmaceuticals?

Difficulties in applying traditional riskassessment framework

A risk assessment has four general components: hazardidentification, toxicity assessment, exposure assessment, andrisk characterization. Nanomaterials can easily be identifiedas a potential hazard, but they present many complications tothe subsequent three steps.

When a nanomaterial is used for a medical application, itis intentionally given to a patient because of some uniqueproperty that its size (and often chemistry) imparts. Forexample, the nanosized particles may have the abilityto access different tissues than larger particles, such ascrossing the blood-brain barrier, or to be tagged with specificantibodies to home in on and be taken up by specific cells.Nevertheless, nanomaterials can cause side effects, and atoxicity assessment requires knowledge of their metabolismand distribution in the body. A variety of techniques arecurrently available for determining the distribution of ananomedicine in a patient, such as radiolabeling, which canbe used to evaluate distribution and uptake into specificcells and tissues. Distribution depends on several factors,including the mechanism of targeting. Cancer cells can betargeted using antibody conjugation to a medication; directtargeting can be enabled so the nanomedicine can be takenup by specific cells; and nanomedicine can passively diffuseinto tissues or cells, for example taking advantage of theleaky endothelia in the blood vessels around some solidtumors. In each case it is possible for the medicine to reacha different population of unintended cells. This situationis complicated by the possibility of making use of manydifferent delivery routes, including oral, transdermal, intra-venous, and inhalation. Further considerations includewhether the nanomaterial stays localized or re-enters thecirculatory system and how it is used or metabolized.Specific nanomaterials will bring with them their ownspecific factors to consider.

Multiple variables could also influence nanomedicineexposure assessment, including characterization of varia-tions in biological reactivity, size, shape, charge, and routeof administration, as well as factors that complicate thestraightforward estimation of exposure (e.g., metabolism,excretion, adduction to biological molecules, etc.). Forexample, several studies on carbon nanotubes have shownthat the toxicity and distribution of nanoparticles isdependent upon the presence of functional groups, impu-rities, fiber length, and aggregation status.6,7 When ananomaterial is not used for a medical application butexposure is instead environmental, exposure estimation maybe even less straightforward.

Given estimates of exposure and toxicity, the final stepinvolved in estimating the hazard of contaminant exposureis the characterization of the dose-response function, that is,the likelihood of adverse health effects at varying degrees ofexposure. By necessity, a dose-response assessment mustbe developed separately for each nanomaterial. Given therequired effort, detailed dose-response assessments willnot be possible for all nanomaterials. Decision tools anddatabases should be developed to facilitate use of allavailable information as well as proxy data for making thebest judgment on dose-response. Risk to individuals can thenbe quantitatively and qualitatively determined. However,unlike the reference doses used by the EPA in health riskassessments for the general population, safety standards for

169I. Linkov et al / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 167–171

agencies such as the US Food and Drug Administration canbe variable. Drugs designed to treat extreme forms ofdisease, such as cancer and acquired immunodeficiencysyndrome, may not be required to give patients more than alimited margin of safety.

Risk-based decision analysis

As with any new technology or science, developing aframework for selecting appropriate nanomaterials andmaking medical decisions with uncertainty and incompleteinformation is the current challenge for the field of nano-technology. Understanding nanomaterial toxicity requiresmultiple sets of information because of both the complexityof nanomaterials and the often-limited database of relevantexperimental studies. One of the tools widely used in riskassessment applications in similar situations is the weight-of-evidence approach. Weight-of-evidence considerations arerequired in assessing risks to ecological receptors.8 The EPAand other agencies use a weight-of-evidence approach inevaluating the potential carcinogenicity and toxicity ofenvironmental contaminants.9 Traditionally, assessors weighvarious lines of evidence and apply professional judgmentand/or calculations to decide where the weight of evidencelies—that is, whether the various lines of evidence point topotential risk in the case of each receptor or not. Much ofthis effort, however, is not initially transparent. Eventhough weight-of-evidence considerations may includesome quantification, this approach often results in arbitraryweight selection and thus in risk estimates that include anunquantified degree of uncertainty and potential bias.Thus, we have introduced MCDA as a tool for integratingheterogeneous information for regulatory decision makingfor nanomaterials.10

We believe that MCDA could be applied widely tosupport decisions on the border of nanomedicine andnanotoxicology. The advantages of using MCDA techniquesover other less structured decision-making methods arenumerous: MCDA provides a clear and transparent metho-dology for making decisions and also provides a formal wayfor combining information from disparate sources. Thesequalities make decisions made through MCDA morethorough and defensible than decisions made through lessstructured methods. For example, MCDA could be used tosupport weight-of-evidence evaluation of nanomaterials.11

Moreover, MCDA could be easily linked with adaptivemanagement for nanomedicine development. In an adaptivemanagement paradigm, the uncertainty in nanomaterial riskswould be acknowledged, and strategies would be formulatedto manage or reduce the uncertainty. The basic adaptivemanagement process is straightforward: one chooses amanagement action, monitors the effects of the action, andadjusts the action based on the monitoring results andupdated social and economic factors.12 During the adaptivemanagement process, in contrast to traditional management,changes are expected and discussed, learning is emphasized,

and objectives can be revised based on the performance of amanagement alternative, changing societal values, orinstitutional learning. An ideal governance frameworkwould be adaptive,13 and a combination of adaptivemanagement and MCDA would provide a powerful frame-work for a wide range of environmental managementproblems, including nanotechnologies. It would allowstructured, clear decisions to be made and also theadjustment of those decisions based on their performance.14

Proposed approach

Integrating this heterogeneous and uncertain informationdemands a systematic and understandable framework toorganize scarce technical information and expert judgment.Current work for the EPA and US Department of Defense12

shows that MCDA methods provide a sound approach tomanagement of heterogeneous information and risks.

Our approach for making efficient decisions on appro-priate nanomaterials for medical applications will allow jointconsideration of the medical factors and side effects alongwith associated uncertainties relevant to selection ofalternative nanomaterials and treatments. It will follow asystematic decision framework developed by Linkov et al10

A generalized MCDA process will follow two basic themes:(1) generating alternative nanomaterials and treatmentoptions, success criteria, and value judgments; and (2)ranking the alternatives by applying value weights. The firstpart of the process generates and defines choices, perfor-mance levels, and preferences. The latter section methodi-cally prunes nonfeasible alternatives by first applyingscreening mechanisms (e.g., significant toxicity, excessivecost) and then ranking in detail the remaining alternativenanomaterials by MCDA techniques that use the variouscriteria levels generated by toxicity models, experimentaldata, or expert judgment. Although it is reasonable to expectthat the process may vary in specific details among nano-medical applications and project types, emphasis should begiven to designing an adaptive management structure thatuses adaptive learning as a means for incorporating changingdecision priorities or new knowledge from toxicity testing orother data into nanotechnology strategy selection or change.

The tools used within group decision making andscientific research are essential elements of the overalldecision process. The applicability of the tools is symbolizedin Figure 1 by solid lines (direct involvement) and dottedlines (indirect involvement). Decision analysis tools help togenerate and map technical data as well as individualjudgments into organized structures that can be linked withother technical tools from risk analysis, modeling, monitor-ing, and cost estimations. Decision analysis software canalso provide useful graphical techniques and visualizationmethods to express the gathered information in under-standable formats. When changes occur in the requirementsor the decision process, decision analysis tools can respondefficiently to reprocess and iterate with the new inputs. This

Figure 1. Example decision process. Dark lines indicate direct involvement / applicability, and dotted lines indicate less direct involvement / applicability.

170 I. Linkov et al / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 167–171

integration of decision tools and scientific and engineeringtools allows users to have a unique and valuable role in thedecision process without attempting to apply either type oftool beyond its intended scope.

Three basic groups of stakeholders (nanotechnologymanagers and decision makers, scientists and engineers,and patients or nanomedicine users) are symbolized inFigure 1 by dark lines for direct involvement and dottedlines for less direct involvement. Although the actualmembership and function of these three groups mayoverlap or vary, the roles of each are essential inmaximizing the utility of human input into the decisionprocess. Each group has its own way of viewing theproblem, its own method of envisioning solutions, and itsown responsibility. Nanotechnology managers spend mostof their effort defining the problem’s context and theoverall constraints on the decision. In addition, they mayhave responsibility for final nanomaterial selection.Patients and technology recipients may provide input indefining nanomedical and nanomaterial alternatives, butthey contribute the most input in helping formulateperformance criteria and making value judgments toweight success criteria. Depending on the problem andcontext, patients and users may have some responsibilityin ranking and selecting the final nanomaterial usealternative. Scientists and engineers have the mostfocused role in that they provide the measurements orestimations for the desired criteria that determine thesuccess of various nanomaterials and alternatives.

The result of the entire process is a comprehensive,structured process for selecting the optimal alternative inany given situation, drawing from stakeholder preferencesand value judgments as well as scientific modeling and

risk analysis. This structured process would be of greatbenefit to decision making in nanotechnology manage-ment, where there is currently no structured approach formaking justifiable and transparent decisions with explicittrade-offs between social and technical factors (e.g.,using a cancer medication with the potential for adverseside effects).

Discussion

Because the nanomedical field is growing, it is importantto be proactive in response to stakeholder concerns andvalues. Although nanotechnology holds great promises forthe future, there is apprehension that there may be unseenadverse effects. Although the potential for adverse effects isunknown at this time, researchers and developers must makedecisions on how to continue to grow the field whilebalancing the safety of the public. Likewise, consumersand patients need to understand the level of risk, if any,associated with the use of nanomedicines and makedecisions based on the best information available to them.The MCDA framework links heterogeneous information oncauses, effects, and risks for different nanomaterials withdecision criteria and weightings elicited from decisionmakers, allowing visualization and quantification of thetrade-offs involved in the decision-making process. Theproposed framework can also be used to prioritize researchand information-gathering activities and thus can be usefulfor value-of-information analysis. New data are constantlybeing researched and presented. With the growth of thispowerful scientific database, MCDA offers an innovativeand effective way to integrate and evaluate the wealth ofknowledge relating to nanomedicines.

171I. Linkov et al / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 167–171

References

1. Kagan VE, Bayir H, Shvedova AA. Nanomedicine and nanotoxicology:two sides of the same coin. Nanomedicine 2005;1:313-6.

2. Moghimi SM, Hunter AC, Murray JC. Nanomedicine: current statusand future prospects. FASEB J 2005;19:311-30.

3. Seaton A, Donaldson K. Nanoscience, nanotoxicology, and the need tothink small. Lancet 2005;365:923-4.

4. Shvedova AA, Kisin ER, Mercer R, Murray AR, Johnson VJ,Potapovich AI, et al. Unusual inflammatory and fibrogenic pulmonaryresponses to single-walled carbon nanotubes in mice. Am J PhysiolLung Cell Mol Physiol 2005;289:L698-708.

5. Peters K, Unger RE, Kirkpatrick CJ, Gatti AM, Monari E. Effects ofnano-scaled particles on endothelial cell function in vitro: studies onviability, proliferation and inflammation. J Mater Sci Mater Med 2004;15:321-5.

6. Donaldson K, Aitken R, Tran L, Stone V, Duffin R, Forrest G, et al.Carbon nanotubes: a review of their properties in relation to pulmonarytoxicology and workplace safety. Toxicol Sci 2006;92:5-22.

7. Lacerda L, Bianco A, Prato M, Kostarelos K. Carbon nanotubes asnanomedicines: from toxicology to pharmacology. Adv Drug Deliv Rev2006;58:1460-70.

8. US Environmental Protection Agency. Ecological risk assessmentguidance for Superfund: process for designing and conductingecological risk assessments—interim final. Washington, DC: US EPA;1997.

9. US Environmental Protection Agency. Nanotechnology white paper,external review draft. Washington, DC: US EPA; 2005.

10. Linkov I, Satterstrom FK, Steevens J, Pleus R. Multi-criteria decisionanalysis and nanotechnology. J Nanopart Res 2007;9:543-54.

11. Linkov I, Satterstrom FK, Zemba S, Steevens J. Use of multi-criteriadecision analysis tools to facilitate weight-of-evidence evaluation innanotechnology risk assessment. NSTI Nanotechnology Conferenceand Trade Show, Boston, Massachusetts; 2006.

12. Linkov I, Satterstrom FK, Kiker G, Batchelor C, Bridges T, Ferguson E.From comparative risk assessment to multi-criteria decision analysisand adaptive management: recent developments and applications.Environ Int 2006;32:1072-93.

13. Roco MC. Possibilities for global governance of converging technol-ogies. J Nanopart Res 2008;10:11-29.

14. Linkov I, Satterstrom FK, Kiker GA, Bridges TS, Benjamin SL, BelluckDA. From optimization to adaptation: shifting paradigms in environ-mental management and their application to remedial decisions. IntegrEnviron Assess Manag 2006;2:92-8.