The ACTTION-American Pain Society Pain Taxonomy (AAPT): An Evidence-Based and Multidimensional Approach to Classifying Chronic Pain Conditions

Page 1

The Journal of Pain, Vol 15, No 3 (March), 2014: pp 241-249Available online at www.jpain.org and www.sciencedirect.com

Focus Article

The ACTTION-American Pain Society Pain Taxonomy (AAPT):

An Evidence-Based and Multidimensional Approach to Classifying

Chronic Pain Conditions

Roger B. Fillingim,* Stephen Bruehl,y Robert H. Dworkin,z Samuel F. Dworkin,x

John D. Loeser,{ Dennis C. Turk,ll Eva Widerstrom-Noga,# Lesley Arnold,**Robert Bennett,yy Robert R. Edwards,zz Roy Freeman,xx Jennifer Gewandter,{{

Sharon Hertz,llll Marc Hochberg,## Elliot Krane,*** Patrick W. Mantyh,yyy

John Markman,zzz Tuhina Neogi,xxx Richard Ohrbach,{{{ Judith A. Paice,llllll

Frank Porreca,### Bob A. Rappaport,**** Shannon M. Smith,yyyy Thomas J. Smith,zzzz

Mark D. Sullivan,xxxx G. Nicholas Verne,{{{{ Ajay D. Wasan,llllllll

and Ursula Wesselmann####

*Pain Research and Intervention Center of Excellence, Gainesville, Florida.yDepartment of Anesthesiology, Vanderbilt University School of Medicine, Nashville, Tennessee.zDepartment of Neurology in the Center for Human Experimental Therapeutics; and Director, Analgesic, Anesthetic,and Addiction Clinical Trial Translations, Innovations, Opportunities, and Networks (ACTTION), University of RochesterSchool of Medicine and Dentistry, Rochester, New York.xDepartment of Psychiatry and Behavioral Sciences, School of Medicine, Department of Oral Medicine, School ofDentistry, University of Washington, Seattle, Washington.{Department of Neurological Surgery, University of Washington, Seattle, Washington.llDepartment of Anesthesiology and Pain Medicine, University of Washington, Seattle, Washington.#Health Scientist VHA, University of Miami, Miller School of Medicine, Miami Project to Cure Paralysis, Miami, Florida.**Department of Psychiatry and Behavioral Neuroscience, and Director,Women’s Health Research Program, Universityof Cincinnati College of Medicine, Cincinnati, Ohio.yyFibromyalgia Research Unit, Oregon Health & Science University, Portland, Oregon.zzBrigham & Women’s Hospital, Chestnut Hill, Massachusetts.xxDepartment of Neurology, Harvard Medical School, Cambridge, Massachusetts.{{Department of Anesthesiology, University of Rochester School of Medicine and Dentistry, Rochester, New York.llllFood and Drug Administration, Center for Drug Evaluation and Research, Silver Spring, Maryland.##Department of Medicine, Division of Rheumatology and Clinical Immunology, University of Maryland School ofMedicine, Baltimore, Maryland.***Departments of Anesthesiology, Perioperative, and Pain Medicine, Pediatrics, Stanford University School ofMedicine, Stanford, California.yyyDepartment of Pharmacology, University of Arizona, Tucson, Arizona.zzzUniversity of Rochester Medical Center, School of Medicine and Dentistry, Rochester, New York.xxxClinical Epidemiology Unit, Boston, Massachusetts.{{{Department of Oral Diagnostic Sciences, University at Buffalo, Buffalo, New York.

The views expressed in this article are those of the authors, none of whomhas a financial conflict of interest related to the specific issues discussed inthis manuscript. All of the authors attended the consensus meeting onwhich this article is based. In addition, the consensus meeting was at-tended by employees of Pfizer, one of the companies that provided unre-stricted grants to ACTTION to support its activities. However, theseindividuals did not contribute to the content of this manuscript.No official endorsement by the U.S. Food and Drug Administration (FDA)or the pharmaceutical companies that have provided unrestricted grantsto support the activities of the Analgesic, Anesthetic, and Addiction Clin-ical Trial Translations, Innovations, Opportunities, and Networks (ACT-TION) public-private partnership with the FDA should be inferred. The

consensus meeting on which this article is based was funded by ACTTION,which has received research grants or other revenue from the FDA,various pharmaceutical companies, and other sources.Address reprint requests to Roger B. Fillingim, PhD, University of FloridaCollege of Dentistry, UF Pain Research and Intervention Center ofExcellence Clinical and Translational Research Building (CTRB), Room3216, 2004 Mowry Road, PO Box 100404, Gainesville, FL 32610-0404.E-mail: [email protected]/$36.00

ª 2014 by the American Pain Society

http://dx.doi.org/10.1016/j.jpain.2014.01.004

241

Page 2

242 The Journal of Pain The ACTTION-APS Pain Taxonomy (AAPT)

llllllDirector, Cancer Pain Program, Division of Hematology-Oncology, Northwestern University, Feinberg School ofMedicine, Chicago, Illinois.###Department of Pharmacology, University of Arizona, Tucson, Arizona.****Food and Drug Administration, Center for Drug Evaluation and Research, Silver Spring, Maryland.yyyyDepartment of Anesthesiology, University of Rochester School of Medicine and Dentistry, Rochester, New York.zzzzDepartment of Oncology, Harry J. Duffey Family Professor of Palliative Medicine, and Director of PalliativeMedicine, Johns Hopkins Medical Institutions, The Johns Hopkins Hospital, Baltimore, Maryland.xxxxDepartment of Psychiatry and Behavioral Sciences, University of Washington, Seattle, Washington.{{{{Department of Medicine, University of Texas Medical Branch, Galveston, Texas.llllllllDepartment of Anesthesiology and Psychiatry, University of Pittsburgh, Pittsburgh, Pennsylvania.####Department of Anesthesiology and Neurology, University of Alabama at Birmingham, Birmingham, Alabama.

Abstract: Current approaches to classification of chronic pain conditions suffer from the absence of a

systematically implemented and evidence-based taxonomy. Moreover, existing diagnostic approaches

typically fail to incorporate available knowledge regarding the biopsychosocialmechanisms contributing

to pain conditions. To address these gaps, the Analgesic, Anesthetic, and Addiction Clinical Trial Transla-

tions Innovations Opportunities and Networks (ACTTION) public-private partnership with the U.S. Food

and Drug Administration and the American Pain Society (APS) have joined together to develop an

evidence-based chronic pain classification system called the ACTTION-APS Pain Taxonomy. This paper de-

scribes the outcome of an ACTTION-APS consensus meeting, at which experts agreed on a structure for

this new taxonomy of chronic pain conditions. Several major issues around which discussion revolved

are presented and summarized, and the structure of the taxonomy is presented. ACTTION-APS Pain Tax-

onomywill include the following dimensions: 1) core diagnostic criteria; 2) common features; 3) common

medical comorbidities; 4) neurobiological, psychosocial, and functional consequences; and 5) putative

neurobiological and psychosocial mechanisms, risk factors, and protective factors. In coming months,

expertworkinggroupswill apply this taxonomytoclustersof chronicpainconditions, therebydeveloping

a set of diagnostic criteria that have been consistently and systematically implemented across nearly all

common chronic pain conditions. It is anticipated that the availability of this evidence-based andmecha-

nistic approach topain classificationwill be of substantial benefit to chronic pain research and treatment.

Perspective: The ACTTION-APS Pain Taxonomy is an evidence-based chronic pain classification

system designed to classify chronic pain along the following dimensions: 1) core diagnostic criteria;

2) common features; 3) common medical comorbidities; 4) neurobiological, psychosocial, and func-

tional consequences; and 5) putative neurobiological and psychosocial mechanisms, risk factors,

and protective factors.

ª 2014 by the American Pain Society

Key words: Taxonomy, chronic pain, pain classification, biopsychosocial, evidence-based, mechanism-

based pain classification.

Thepurpose of clinical diagnosis is to provide a validexplanation of symptoms and signs in order toguide treatment and inform prognosis. Several

characteristics of an ideal diagnostic system are pre-sented in Table 1. Although clinical diagnosis in manyfields has progressed considerably in recent decades,incorporating new evidence and improved diagnostictechnologies, classification of pain disorders has wit-nessed limited advances. Indeed, one could argue thatcurrent pain classification systems fail to fulfill the pri-mary purpose of diagnosis (to guide treatment) andmeet few of the characteristics of ideal classification sys-tems. A challenge for pain classification is the need toaccount for individual differences in pain processing,which often result in a weak association between typical‘‘diagnostic’’ measures of tissue damage or disease activ-ity and the severity and clinical symptoms. The mostcomprehensive pain classification system available ispublished by the International Association for the Studyof Pain (IASP). First published in 1979, revised in 1994,and updated in 2011, the IASP system intended to fulfill

a mission described by IASP founder John Bonica, whosaid, ‘‘It is possible to . develop a classification of painsyndromes which are acceptable to many . in the field;even if the adopted definitions and classifications are notperfect they are better than the Tower of Babel condi-tions that currently exist’’ (p. ix).16 However, although itdescribes a large number of chronic pain conditions,this system has never been widely adopted by eitherthe clinical or the research community. Notably, overthe past 15 years, multiple authors have called for amechanism-based approach to pain diagnosis21-24;however, information regarding neurobiologicalmechanisms is limited and has not been routinelyincorporated into any existing pain classificationsystems. Thus, there is an increasingly urgent need todevelop a standardized, systematic, and evidence-basedapproach to pain classification that incorporates infor-mation regarding biopsychosocial mechanisms and thatcan be applied to all common chronic pain conditions.To that end, the Analgesic, Anesthetic, and Addiction

Clinical Trial Translations Innovations Opportunities

Page 3

Table 1. Characteristics of an Ideal Diagnostic System

CHARACTERISTIC DESCRIPTION

Biologically plausible The diagnostic system must be consistent with the biological processes underlying the signs and symptoms

that characterize the disorders of interest.

Exhaustive The diagnostic system must encompass all clinical disorders within the domain of interest.

Mutually exclusive The diagnostic system must encode each disorder once and only once.

Reliable The diagnostic system must be applicable with a high degree of consistency across time and between

diagnosticians.

Clinically useful The diagnostic system must be useful in the clinical setting, guiding prognosis and therapy.

Simple The diagnostic system must be both straightforward and efficient enough for practical use.

Fillingim et al The Journal of Pain 243

andNetworks (ACTTION) public-private partnershipwiththeU.S. Food andDrugAdministration and theAmericanPain Society (APS) have joined together to develop a clas-sification system that incorporates current knowledge ofbiopsychosocial mechanisms, called the ACTTION-APSPain Taxonomy (AAPT). The overriding objective ofAAPT is to develop an evidence-based chronic pain tax-onomy based on a consistently appliedmultidimensionalframework and then to encourage experts to apply thismultidimensional framework to specific chronic painconditions. A major impetus for the AAPT initiativederived from observing the transformative impact ofevidence-based diagnostic classifications in relatedfields. For example, the third edition of the Diagnosticand Statistical Manual of Mental Disorders (DSM-III) ofthe American Psychiatric Association arguably revolu-tionized research and clinical practice in psychiatry bysystematically implementing an evidence-based anddescriptive taxonomy to replace prior theoretically basedapproaches that did not have adequate reliability andwere not widely used.1,15 The DSM-III unveiled a multi-axial system wherein specific diagnostic criteria werepresented on Axes 1 and 2, and the remaining axes al-lowed the diagnostician to provide additional clinicallyrelevant information, including comorbidmedical condi-tions, psychosocial stressors, and overall functioning. TheDSM has seen further developments in subsequent edi-tions, including the recently released DSM-5,2 and it re-mains an evidence-based descriptive taxonomy that isuniversally applied in psychiatry. Thus, DSM-III and subse-quent updates served as an exemplary model for AAPT.Another successful diagnostic system from which AAPTdrew guidance is the International Classification ofHeadache Disorders (ICHD), an evidence-based systemfor classifying headaches.17,18 The ICHD, now in itsthird edition,19 was first released in 1988 and has beena tremendous boon to headache research and clinicalpractice. The ICHD is the gold standard for headacheresearch, including clinical trials, which has led to thedevelopment of evidence-based treatments for severalheadache disorders.There are multiple benefits to developing a unified

and standardized evidence-based taxonomy for classi-fying chronic pain disorders. First, to the extentpossible, it can help align current knowledge of bio-psychosocial mechanisms with an evidence-based paintaxonomy, and subsequently with tailored pain treat-ment. As previously described,24 current pain classifica-tion is based primarily on symptoms, signs, and body

location, sometimes combined with informationregarding structural pathology (eg, magnetic resonanceimaging evidence of spinal stenosis), and rarely in an in-tegrated biopsychosocial framework. However, paintreatments produce their therapeutic effects by impact-ing specific neurobiological and/or psychosocial mecha-nisms. Hence, an integrated biopsychosocial taxonomythat includes current and emerging evidence regardingpain mechanisms can assist clinicians with treatment se-lection, thereby enhancing patient outcomes. Second, ataxonomy in which diagnostic criteria for all chronicpain conditions are consistently and systematically im-plemented will facilitate communication about chronicpain and related research by ensuring comparabilityacross studies of the same condition. That is, at present,case definitions of the same chronic pain condition varywidely across professional organizations and studies,which limits the ability to compare findings acrossstudies. A widely accepted, consistently applied, andevidence-based taxonomy would help overcome thisimpediment to research and ultimately treatment. Inparticular, a systematically implemented taxonomycould enhance clinical trial methodology by promotingconsistently applied inclusion criteria, thereby facili-tating treatment development. Third, AAPT would pro-vide educational benefits because this systematicallydeveloped taxonomy would represent an excellentevidence-based and mechanistic foundation fortraining pain researchers and clinicians. Finally, by iden-tifying gaps in the evidence regarding diagnostic symp-toms of particular disorders and their neurobiologicaland psychosocial underpinnings, such a taxonomy couldhighlight important avenues for future pain research.

MethodsIn order to develop the multidimensional framework

and chronic pain taxonomy, AAPT held a consensusmeeting in May 2013 that brought together clinicaland basic scientists with expertise in pain mechanismsand in the major chronic pain disorders in both adultsand children. In order to facilitate the process, a set offoundational papers was distributed to all participantsprior to the meeting and the agenda included back-ground presentations addressing the following topics:1) the importance of pain mechanisms in chronic painclassification (F.P., P.W.M.); 2) the development and evo-lution of DSM to illustrate a successful classificationdeveloped outside the area of pain; 3) 3 presentations

Page 4

244 The Journal of Pain The ACTTION-APS Pain Taxonomy (AAPT)

describing successful efforts in developing evidence-based diagnostic criteria for specific pain disorders,namely, temporomandibular disorders (TMD; S.F.D.),complex regional pain syndrome (S.B.), and chronicpain associated with spinal cord injury (E.W.-N.); andfinally, 4) guiding principles for developing a newchronic pain taxonomy (R.B.F.). The remainder of themeeting was devoted to building consensus regardingthe multidimensional framework and the organizationof the taxonomy and the chronic pain disorders to whichthe framework would be applied. During the course ofdiscussion, several important issues arose, which will bedescribed individually below.

Review of Recently Developed Evidence-Based Diagnostic Criteria for SpecificPain DisordersTo provide examples of implementing an evidence-

based approach to pain classification, 3 investigatorswho were involved in different initiatives to developevidence-based diagnostic criteria for 3 different paindisorders presented their approach and findings. S.F.D.discussed the development of the Research DiagnosticCriteria (RDC) for Temporomandibular Disorders (RDC/TMD).7 The RDC/TMD aimed to 1) develop standardizeddiagnostic criteria for major TMD subtypes to be widelyused in research, 2) provide reliable specifications forclinical examination and history, 3) use 2 axes to reflectphysical disease (Axis I) and subjective illness experience(Axis 2), and 4) invoke an iterative research processmodeled after the DSM, which requires periodicallyreestablishing the evidence-based reliability and validityof the newly emerging iterations of the RDC/TMD. Eachof these objectives has beenmet, and the newest versionto evolve is the DC-TMD-1, an evidence-based diagnosticand classification system for common forms of TMDscientifically suitable for clinical practice as well asresearch. Thus, the RDC/TMD has evolved to becomethe gold standard for classification of TMDs in bothresearch and clinical settings. S.B. presented his experi-ence regarding the development of revised diagnosticcriteria for complex regional pain syndrome.10,11 Basedon an external validation study conducted acrossmultiple sites in Europe and North America, theprevailing IASP consensus criteria were found to havepoor specificity, leading to potential overdiagnosis ofcomplex regional pain syndrome. An expert paneldeveloped new diagnostic criteria, which onsubsequent validation were found to have muchgreater specificity than the IASP criteria, and these newcriteria were adopted by IASP in 2011. E.W.-N. presentedthe International Spinal Cord Injury Pain Classification,an effort to reconcile the multiple spinal cord injurypain taxonomies present at the time in order to developvalid diagnostic criteria for pain associated with spinalcord injury.3,4 The newly developed classificationincorporated feedback from multiple professionalorganizations and was subsequently validated usingclinical vignettes. These 3 efforts in pain classificationdemonstrate the feasibility and utility of adopting an

evidence-based approach in developing a pain taxon-omy. However, each classification system differs substan-tially from the others, reflecting the absence of anoverarching framework for pain classification. AAPThas established such a framework in order to produce aconsistent pain taxonomy that includes all commonchronic pain disorders.

Important Characteristics of the AAPTTaxonomyThe single most important characteristic of the taxon-

omy is that it be based on the best available evidencerather than based solely on consensus or expert opinion.That is, to the greatest extent possible, diagnostic criteriafor specific chronic pain disorders should be determinedusing existing mechanistic and diagnostic evidence,rather than historical precedent or theoretical biases.When necessary, additional data will be collected to pro-vide the required evidence to guide the working groupin developing diagnostic criteria. It is acknowledgedthat the classification will evolve and be revised on thebasis of accumulating evidence and knowledge (as inthe evolution of DSM-III). It is also important to notethat AAPT proposes a coordinated effort to implementthe taxonomy systematically across all common chronicpain conditions. Another critical characteristic of thetaxonomy is that it reflects the multidimensional andbiopsychosocial nature of chronic pain in which relevantpsychological and social variables are integrated withneurophysiological knowledge. Thus, the template forthe AAPT taxonomy includes not only pain-related diag-nostic criteria and features but also psychosocial featuresand functional impact of pain conditions. Additionally,AAPT emphasizes the inclusion of existing informationregarding mechanistic features and risk factors for painconditions, including not only neurobiological processesbut also psychosocial contributions, which are consid-ered mechanisms in their own right. Another essentialcharacteristic is that the taxonomy should be applicablefor both research and clinical purposes, such that thediagnostic criteria could be as easily used by primarycare providers as by pain scientists; however, it is recog-nized that widespread clinical use is likely to developgradually as the evidence base expands and the taxon-omy evolves further. Although the taxonomy is not de-signed with consideration of factors related to billingor third-party reimbursement for clinical services, au-thors of diagnostic criteria for each condition willattempt to provide information regarding internationalclassification of diseases (ICD) codes that are related tothe AAPT condition being described. Finally, the taxon-omy is meant to be a starting point based on currentlyavailable evidence, and the goal is to systematically up-date the criteria as new evidence, especially regardingneurobiological mechanisms, becomes available. Indeed,it seems likely that refinements and enhancements to thecriteria will occur according to the experience of theworking groups in applying the taxonomy to specificpain conditions.

Page 5

Table 2. Organization of Chronic Pain Disordersto Be Included in the AAPT*

Peripheral and central nervous systems

Peripheral neuropathic pain

Central neuropathic pain

Musculoskeletal pain system

Osteoarthritis

Other arthritides (eg, rheumatoid arthritis, gout, connective tissuediseases)

Musculoskeletal low back pain

Myofascial pain, chronic widespread pain, and fibromyalgia

Other predominantly musculoskeletal pain

Orofacial and head pain system

Headache disorders*

Temporomandibular disorders

Other orofacial pain

Visceral, pelvic, and urogenital pain

Visceral pain: abdominal, pelvic, and urogenital pain

Disease-associated pains not classified elsewhere (eg, pain associated

with active cancer, with sickle cell disease, or with Lyme disease)

*AAPTwill not develop diagnostic criteria for headache condition, because the

ICHD-2 already exists and provides an evidence-based classification that is highly

consistent with the AAPT template.

Fillingim et al The Journal of Pain 245

Should the New Taxonomy Be anEvolution or a Revolution?An important topic of discussion related towhether the

new taxonomy should significantly depart from currentand historical practice versus retaining features of currentapproaches to classification. Proponents of the formerargued that current approaches reflectdescriptive systemsorganizedon the basis of an inconsistently applied combi-nation of body location, affected tissues, and associateddisease states,which provides little information regardingthe pathophysiological mechanisms underlying thepain itself that should in principle be the targets of treat-ment. Moreover, this approach treats chronic pain disor-ders that often share biopsychosocial mechanisms (eg,fibromyalgia, irritable bowel syndrome, and TMDs)as completely independent.5,20 Thus, a revolutionaryapproach to chronic pain taxonomy might completelyabandon current diagnostic labels and approaches basedon anatomical structures and organ systems in favorof an approach that prioritizes the neurobiologicalmechanisms underlying chronic pain disorders. Althoughthere was unanimity regarding the importance ofincorporating pathophysiologic mechanisms into thenew taxonomy, 2 concerns prevented endorsement of arevolutionary approach based primarily if not exclusivelyon mechanisms. First, there was agreement that existingknowledge regarding the mechanisms underlying manychronic pain disorders was insufficient to support suchan approach. Second, there was concern that such aradical departure from prevailing practice would faceresistance from clinicians and scientists who are familiarwith classical systems and who would be reluctant toaccept a significant change from current approaches.Therefore, the consensus dictated that the AAPT wouldretain similarities to existing systems, but, as exemplifiedby the RDC/TMD to DC-TMD-1 evolution, the AAPTapproach would incorporate existing and emergingevidence regarding neurobiological and psychosocialmechanisms into all diagnostic criteria.

Should AAPT Adopt a Medical or aSyndromal Approach to PainClassification?Medical diagnostic approaches (eg, ICD-10) prioritize

identification of pathophysiological mechanisms,whereas syndromalapproaches (eg,DSM-V) classify condi-tions primarily based on clusters of symptoms. Arguably,AAPTrepresents a hybrid of these 2 approaches. Althoughthe AAPT Core Diagnostic Criteria (Dimension 1) dictatethat signs and symptoms represent the primary basis onwhichdiagnoseswill bebased, the taxonomyalso includesDimension 4, on which biopsychosocial mechanismscontributing to the condition canbedelineated.However,it is important to recognize that the AAPT mechanisticdimension differs from the historical biomedical view ofthe pathophysiology of pain conditions, which empha-sized peripheral markers of structural pathology and/ordisease severity. These pathophysiologic measures havegenerally corresponded poorly to chronic pain severityand have failed to account for interindividual variability

in clinical symptoms. In contrast, AAPT intends themecha-nistic dimension to specify the neurobiological and psy-chosocial factors that contribute to the development ofchronic pain and account for the robust individual differ-ences inclinicalpresentationthatareahallmarkof chronicpain. Indeed, given the subjective and personal nature ofthe pain experience, a solely medical/pathophysiologicalapproach to pain classification seems neither realistic noradvisable. Thus, ultimately, AAPT represents a syndromaltaxonomy that incorporates existing informationregardingmechanisms, while recognizing the importanceof individual differences in clinical presentation. Thisapproach is designed to produce a practically useful andevidence-based taxonomy that allows a person-centeredapproach to classification and clinical care.

How Should Chronic Pain Disorders BeCategorized in the Taxonomy?Considerable discussion addressed the basis upon

which chronic pain disorders would be grouped in thistaxonomy. For example, should conditions be groupedby anatomical locations (eg, upper extremity, lowerextremity, spine) or by organ system (eg, nervous system,musculoskeletal, visceral)? Relying on anatomical sitealone was rejected, because this would cluster togetherchronic pain conditions that have very distinct patho-physiological mechanisms. For instance, lower extremitypain would include peripheral diabetic neuropathy ofthe foot and leg as well as knee osteoarthritis.Conversely, the same disorder affecting differentanatomical sites (eg, peripheral diabetic neuropathy ofthe foot and hand) would be separately categorized.Therefore, the consensus was that the dimension alongwhich pain disorders will be categorized is organ sys-tem/anatomical structure, which will include peripheral

Page 6

Table 3. The Dimensions Comprising the AAPT

DIMENSION DESCRIPTION

Dimension 1: Core diagnostic criteria Includes symptoms and signs required for diagnosis of the disorder (eg, periauricular pain,

palpation sensitivity, joint sounds in the case of TMD). Also includes diagnostic tests and

differential diagnosis considerations.

Dimension 2: Common features Provides additional information regarding the disorder, including common pain

characteristics (eg, location, temporal qualities, descriptors), nonpain features

(numbness, fatigue), and the epidemiology of the disorder. These features are helpful in

describing the disorder but are not used as part of the diagnosis.

Dimension 3: Common medical comorbidities Includes medical diagnoses that co-occur with high frequency with the pain disorder. For

example, diabetes mellitus is often comorbid with osteoarthritis, andmajor depression is

comorbid with many chronic pain disorders.

Dimension 4: Neurobiological, psychosocial,

and functional consequences

Includes information regarding neurobiological and psychosocial consequences of chronic

pain, as well as the functional impact of the pain disorder. Examples include allostatic

load, sleep quality, mood/affect, coping resources, physical function, and pain-related

interference with daily activities

Dimension 5: Putative neurobiological

and psychosocial mechanisms, risk factors,

and protective factors

Includes putative neurobiological and psychosocial mechanisms contributing to the pain

disorder, including potential risk factors and protective factors.

246 The Journal of Pain The ACTTION-APS Pain Taxonomy (AAPT)

and central neuropathic pain, musculoskeletal pain, andpelvic/urogenital and visceral pain (see Table 2). Finally, itwas recognized that certain types of disease-related painmay not be included in one of the other categories;therefore, a category was created for disease-relatedpains not classified elsewhere (eg, pain associated withactive cancer, sickle cell disease, and Parkinson’s disease).The preference is to classify disease-related pain in one ofthe primary organ system/anatomical categories (eg,painful diabetic peripheral neuropathy would be cate-gorized as a peripheral neuropathic pain); however,when this is not possible, the pain disorder will be classi-fied in the disease-related pain category. It is also impor-tant to note that headache disorders were intentionallyexcluded from the present taxonomy, because the ICHDhas been carefully and systematically developed andthere was agreement that the existing criteria fullymeet the standards of the AAPT.

ResultsAfter considerable discussion, a multidimensional

framework for the new chronic pain taxonomy wasdeveloped. The dimensions comprising the AAPT frame-work, whichwill be applied to each chronic pain disorder(see Table 2), are presented in Table 3. Each dimensionwill be discussed in greater detail below. It is importantto recognize that the order of the dimensions is notintended to reflect their priority or significance. Indeed,the consensus meeting unanimously endorsed theimportance of Dimension 5, reflecting underlying mech-anisms. However, this Dimension was not included aspart of the essential diagnostic criteria, because currentevidence provides definitive mechanistic informationfor very few chronic pain disorders.

Dimension 1: Core Diagnostic CriteriaThe core diagnostic criteria reflect those signs, symp-

toms, and test results that form the basis of the diagnosis.Ideally, the core diagnostic criteria should be applied in

an algorithmic manner, such that people meeting thesespecific criteria would be classified as having the disor-der. Signs and symptoms would include diagnosticnonpain features (eg, diminished range of motion,edema, altered sensation) and pain-related characteris-tics (eg, pain descriptors, location, and temporal quali-ties) that are considered pathognomonic of thedisorders. For example, based on RDC/TMD, diagnosticsymptoms of TMD might include periauricular pain,whereas diagnostic clinical signs could include palpationsensitivity, reduced pain-free range of motion, and jointsounds (eg, popping or clicking).7 In addition, for somedisorders, the results of clinical or laboratory tests willbe incorporated into the diagnostic criteria. Finally, dif-ferential diagnoses are also considered in Dimension 1.

Dimension 2: Common FeaturesThis dimension is intended to provide additional

descriptive information regarding the disorder byincluding common features that often characterize thedisorder but are not required for the diagnosis. Forexample, a given chronic pain disorder might commonlybe associated with pain occurring in a specific body loca-tion (eg, diabetic neuropathy pain in the feet) ordescribed in a particular way (eg, burning or shooting);however, these features would not be a requirementfor making the diagnosis. In addition to pain-relatedqualities, this dimension will also include informationregarding nonpainful features as well as informationregarding the epidemiology of the disorder.

Dimension 3: Common MedicalComorbiditiesThis dimension provides information regarding medi-

cal and psychiatric conditions that are often comorbidwith, but not required for diagnosis of, the chronicpain disorder. For example, if major depression andgeneralized anxiety disorder are considerably more com-mon in patients with a specific pain disorder than in thegeneral population, then these would be noted as

Page 7

Fillingim et al The Journal of Pain 247

psychiatric comorbidities. Similarly, Sj€ogren’s syndromecould be included as a medical condition that iscommonly comorbid with burning mouth syndrome.

Dimension 4: Neurobiological,Psychosocial, and FunctionalConsequencesThe neurobiological, psychosocial, and functional

consequences of chronic pain disorders have been welldocumented, and these will be included in Dimension4. Indeed, some individuals may meet diagnostic criteriafor a given pain disorder though their neurobiological,psychosocial, and physical function remain excellent,whereas other patients with the disorder may presentwith considerable dysfunction across one or more ofthese domains. Thus, psychosocial and functional fea-tures could be used for subgrouping individuals withina given pain disorder, which could have importanttreatment implications. Importantly, premorbid psycho-social functioning represents a consistent predictor ofthe development of chronic pain; however, psychologi-cal antecedents of the pain disorder, which could reflectcausal risk factors, will be specified on Dimension 5(see below). We recognize that the distinctionbetween causes and consequences will be challengingwhen applying the taxonomy. However, we also believeit is important to acknowledge that neurobiologicaland psychosocial differences between chronic paincases and controls can reflect both causes (or riskfactors) and consequences of chronic pain. For example,altered neurosensory processing, as measured byquantitative sensory testing, has predicted futuredevelopment of chronic pain in some studies.6,25

However, other research has demonstrated that successfultreatment of pain normalizes quantitative sensory testingresponses.9,12,13 Hence, the neurobiological underpinningsof altered pain processing may be a risk factor for pain insome cases and a consequence of pain in other cases. Asimilar scenario exists for psychological factors as well,because data support that depression, for example, is arisk factor for pain development and that chronic pain is arisk factor for development of depression.8,14 Thus, webelieve it is important that the taxonomy incorporates thebidirectional nature of these associations. These issueswill be addressed on a case-by-case basis for each pain con-dition. Information reflected on this dimension may bederived from the clinical examination or via administrationof psychometric instruments designed to assess these do-mains.

Dimension 5: Putative Mechanisms, RiskFactors, and Protective FactorsThis dimension is intended to provide information

regarding the potential neurobiological and psychoso-cial mechanisms and risk factors contributing to chronicpain disorders. This includes neurobiological mecha-nisms, such as specific molecular or neurochemical path-ways that have been demonstrated to contribute to thedisorder. Risk or protective factors might include specificgenetic polymorphisms. Based on current evidence,

information regarding specific mechanisms and riskfactors may not be available for some chronic pain disor-ders; however, more general information may be avail-able. For example, it is widely accepted thatfibromyalgia is characterized by generalized hypersensi-tivity to painful stimuli, though the precise mechanismsunderlying this phenotype remain unknown. In thisinstance, widespread hypersensitivity could be includedas a potential mechanism or risk factor.In addition to traditional neurobiological mechanisms,

psychosocial factors also represent important painmech-anisms and risk factors to be considered in the taxonomy,including potential protective factors (eg, social support,optimism, coping). It is recognized that psychosocialinfluences on pain must be transduced through moreproximal neurobiological processes, though these spe-cific mechanisms may not be well understood. Forexample, fear-avoidance processes can contribute topain and pain-related disability not only at a behav-ioral/functional level but also at a neurobiological level.For each pain disorder, available and recommendedmethods for assessing psychosocial and neurobiologicalmechanisms will be described. Moreover, it is anticipatedthat new information regarding specific neurobiologicaland psychosocial risk factors and mechanisms for manypain disorders will rapidly emerge, based on genetic as-sociation studies, brain imaging research, quantitativesensory testing, and additional psychosocial mechanisticresearch. Thus, the intent is that the diagnostic criteriawill be updated with this information on an ongoingbasis.

Organization of Chronic Pain DisordersThe consensus meeting next turned its attention to

organizing the chronic pain disorders that would beincluded in the taxonomy. As noted above, there wasconsiderable discussion regarding the best approachfor categorizing the pain disorders, and the finalconsensus was that chronic pain disorders would beprimarily categorized by anatomical/organ system (seeTable 2). There are limitations to this approach to cate-gorization. For example, TMDs could be consideredmusculoskeletal pains. Therefore, in developing thediagnostic criteria for each chronic pain disorder, theexpert working groups will attempt to address featuresof the condition that may overlap with other conditionswithin and across categories. For instance, generalizedhypersensitivity to painful stimuli can characterizechronic pain disorders across virtually every category,and this feature is likely relevant to the underlyingpathophysiology of many chronic pain disorders. Byhighlighting these potentially overlapping features, itmay be possible to build clusters of conditions basedon underlying pathophysiological and psychosocialmechanisms.

Next StepsThis article presents the structure and organization of

the proposed taxonomy, based on the outcomes of our

Page 8

248 The Journal of Pain The ACTTION-APS Pain Taxonomy (AAPT)

consensus meeting. The AAPT Steering Committee hasnow identified leaders for each of the individual work-ing groups who will apply the taxonomy to each groupof conditions specified in Table 2. The working groupleaders will proceed as follows: 1) identify and inviteother experts to be included in the working group, andconvene a working group meeting; 2) identify the mostcommon pain conditions within the working group’spurview for which diagnostic criteria will be specified;3) review the existing literature regarding current andpreviously proposed diagnostic criteria for each of themajor disorders (including strengths/weaknesses ofeach); 4) propose a comprehensive list of potential signsand symptoms for each of the pain conditions identifiedin point 2 (It is anticipated that at this point all workinggroups will reconvene in order to facilitate developmentof draft diagnostic criteria for each chronic pain condi-tion.); 5) complete studies to demonstrate the reliabilityand validity of the proposed diagnostic methods andcriteria; and 6) finalize the diagnostic criteria for eachchronic pain condition based on the outcomes of thevalidation studies and submit a manuscript to a peer-reviewed journal. Working group activities will be over-seen and supported by both an Executive Committee anda Research Committee.

ConclusionsClassification of chronic pain disorders has historically

been a clinically driven and piecemeal exercise, and amore systematic and evidence-based approach to chronicpain diagnosis would confer considerable scientific,clinical, and educational benefit. We have proposed amultidimensional chronic pain taxonomy that will beevidence-based and systematically applied to all commonchronic pain disorders. Notably, AAPT explicitly includes adimension on which information regarding neurobiolog-ical and psychosocial mechanisms will be provided. AAPTworking groups will use this taxonomy to developevidence-based diagnostic criteria for most chronic paindisorders,whichare intended forboth researchand clinicaluse. The intent is for AAPT to be a dynamic and evolvingtaxonomy that will be updated and revised as new evi-dence emerges. As has been the case with other newlydeveloped taxonomies, AAPT may initially be most widelyused in research settings, but as the taxonomy evolves itis expected that more widespread clinical use will follow.It is hoped thatAAPTwill produce robust scientific and clin-ical impacton the chronic painfield, similar to the transfor-mative influence of DSM-III and ICHD on research andtreatment in mental health and headache, respectively.

References

1. American Psychiatric Association: Diagnostic and Statisti-cal Manual of Mental Disorders, 3rd ed. Washington, DC,American Psychiatric Association, 1980

2. American Psychiatric Association: Diagnostic and Statisti-cal Manual of Mental Disorders, 5th ed. Washington, DC,American Psychiatric Association, 2013

3. Bryce TN, Biering-Sorensen F, Finnerup NB, Cardenas DD,Defrin R, Ivan E, Lundeberg T, Norrbrink C, Richards JS,Siddall P, Stripling T, Treede RD, Waxman SG, Widerstrom-Noga E, Yezierski RP, Dijkers M: International Spinal CordInjury Pain (ISCIP) Classification: Part 2. Initial validationusing vignettes. Spinal Cord 50:404-412, 2012

4. Bryce TN, Biering-Sorensen F, Finnerup NB, Cardenas DD,Defrin R, Lundeberg T, Norrbrink C, Richards JS, Siddall P,Stripling T, Treede RD, Waxman SG, Widerstrom-Noga E,Yezierski RP, Dijkers M: International Spinal Cord InjuryPain Classification: Part I. Background and description.March 6-7, 2009. Spinal Cord 50:413-417, 2012

5. Diatchenko L, Nackley AG, Slade GD, Fillingim RB,MaixnerW: Idiopathic pain disorders—Pathways of vulnera-bility. Pain 123:226-230, 2006

6. Diatchenko L, Slade GD, Nackley AG, Bhalang K,Sigurdsson A, Belfer I, Goldman D, Xu K, Shabalina SA,Shagin D, Max MB, Makarov SS, Maixner W: Genetic basis forindividual variations in pain perception and the developmentof a chronic pain condition. HumMol Genet 14:135-143, 2005

7. Dworkin SF, LeResche L: Research diagnostic criteria fortemporomandibular disorders. J Craniomandib Disord 6:302-355, 1992

8. Fishbain DA, Cutler R, Rosomoff HL, Rosomoff RS: Chronicpain-associated depression: Antecedent or consequence ofchronic pain? A review. Clin J Pain 13:116-137, 1997

9. Graven-Nielsen T, Wodehouse T, Langford RM, Arendt-Nielsen L, Kidd BL: Normalization of widespread hyperes-thesia and facilitated spatial summation of deep-tissuepain in knee osteoarthritis patients after knee replacement.Arthritis Rheum 64:2907-2916, 2012

10. Harden RN, Bruehl S, Perez RS, Birklein F, Marinus J,Maihofner C, Lubenow T, Buvanendran A, Mackey S,Graciosa J, Mogilevski M, Ramsden C, Chont M, Vatine JJ:Validation of proposed diagnostic criteria (the ‘‘BudapestCriteria’’) for complex regional pain syndrome. Pain 150:268-274, 2010

11. Harden RN, Bruehl S, Stanton-Hicks M, Wilson PR:Proposed new diagnostic criteria for complex regionalpain syndrome. Pain Med 8:326-331, 2007

12. Kosek E, Ordeberg G: Abnormalities of somatosensoryperception in patients with painful osteoarthritisnormalize following successful treatment. Eur J Pain 4:229-238, 2000

13. Kosek E, Ordeberg G: Lack of pressure pain modulationby heterotopic noxious conditioning stimulation in patientswith painful osteoarthritis before, but not following, surgi-cal pain relief. Pain 88:69-78, 2000

14. Lepine JP, Briley M: The epidemiology of pain in depres-sion. Hum Psychopharmacol 19(Suppl 1):S3-S7, 2004

15. Mayes R, Horwitz AV: DSM-III and the revolution in theclassificationofmental illness. JHistBehavSci41:249-267,2005

16. Merskey H, BogdukN: Classification of Chronic Pain, 2nded. Seattle, WA, IASP Press, 1994

17. Olesen J: International Classification of Headache Disor-ders, Second Edition (ICHD-2): Current status and futurerevisions. Cephalalgia 26:1409-1410, 2006

18. Olesen J: The International Classification of HeadacheDisorders. Headache 48:691-693, 2008

Page 9

Fillingim et al The Journal of Pain 249

19. Olesen J: ICHD-3 beta is published. Use it immediately.Cephalalgia 33:627-628, 2013

20. Phillips K, Clauw DJ: Central pain mechanisms in chronicpain states—Maybe it is all in their head. Best Pract Res ClinRheumatol 25:141-154, 2011

21. von Hehn CA, Baron R, Woolf CJ: Deconstructing theneuropathic pain phenotype to reveal neural mechanisms.Neuron 73:638-652, 2012

22. Woodcock J, Witter J, Dionne RA: Stimulating the devel-opment of mechanism-based, individualized pain therapies.Nat Rev Drug Discov 6:703-710, 2007

23. Woolf CJ, Bennett GJ, Doherty M, Dubner R, Kidd B,Koltzenburg M, Lipton R, Loeser JD, Payne R, Torebjork E:Towards a mechanism-based classification of pain? Pain 77:227-229, 1998

24. Woolf CJ, Max MB: Mechanism-based pain diagnosis:Issues for analgesic drug development. Anesthesiology 95:241-249, 2001

25. Yarnitsky D, Crispel Y, Eisenberg E, Granovsky Y,Ben-Nun A, Sprecher E, Best LA, Granot M: Predictionof chronic post-operative pain: Pre-operativeDNIC testing identifies patients at risk. Pain 138:22-28,2008