Precision Medicine in Diabetes: AConsensus Report From theAmerican Diabetes Association(ADA) and the EuropeanAssociation for the Study ofDiabetes (EASD)Diabetes Care 2020;43:1617–1635 | https://doi.org/10.2337/dci20-0022

The convergence of advances in medical science, human biology, data science, andtechnologyhasenabled thegenerationofnew insights into thephenotypeknownas“diabetes.” Increased knowledge of this condition has emerged from populationsaround theworld, illuminating thedifferences in howdiabetes presents, its variableprevalence, and how best practice in treatment varies between populations. Inparallel, focus has been placed on the development of tools for the application ofprecision medicine to numerous conditions. This Consensus Report presents theAmerican Diabetes Association (ADA) Precision Medicine in Diabetes Initiative inpartnership with the European Association for the Study of Diabetes (EASD),including its mission, the current state of the field, and prospects for the future.Expert opinions are presented on areas of precision diagnostics and precisiontherapeutics (including prevention and treatment), and key barriers to and oppor-tunities for implementation of precision diabetes medicine, with better care andoutcomes around the globe, are highlighted. Cases where precision diagnosis isalready feasible and effective (i.e., monogenic forms of diabetes) are presented, whilethe major hurdles to the global implementation of precision diagnosis of complexforms of diabetes are discussed. The situation is similar for precision therapeutics, inwhich the appropriate therapy will often change over time owing to the manner inwhich diabetes evolves within individual patients. This Consensus Report describes afoundation for precision diabetes medicine, while highlighting what remains to bedone to realize its potential. This, combinedwith a subsequent, detailed evidence-based review (due 2022), will provide a roadmap for precision medicine in diabetesthat helps improve the quality of life for all those with diabetes.

RATIONALE FOR PRECISION MEDICINE IN DIABETES

The practice of medicine centers on the individual. From the beginning, the physicianhas examined the patient suffering from illness, ascertained his/her signs andsymptoms, related them to themedical knowledge available at the time, recognizedpatterns that fit a certain category and, based on the practical wisdom accumulatedvia empirical trial and error, applied a given remedy that is best suited to thesituation at hand. Thus, the concept of precision medicine, often defined as

1Department of Pediatrics, Columbia UniversityIrving Medical Center, New York, NY2Department of Medicine, Columbia UniversityIrving Medical Center, New York, NY3American Diabetes Association, Arlington, VA4Center for Genomic Medicine, MassachusettsGeneral Hospital, Boston, MA5Diabetes Unit, Massachusetts General Hospital,Boston, MA6Metabolism Program, Broad Institute of MITand Harvard, Cambridge, MA7Program in Medical and Population Genetics,Broad Institute of MIT and Harvard, Cambridge,MA8Department ofMedicine, HarvardMedical School,Boston, MA9Institute of Biomedical and Clinical Science,College of Medicine and Health, University ofExeter, Exeter, U.K.10Department of Population Medicine, HarvardMedical School, Harvard Pilgrim Health CareInstitute, Boston, MA11National Institute of Diabetes and Digestiveand Kidney Diseases, Bethesda, MD12Wellcome Centre for Human Genetics, Univer-sity of Oxford, Oxford, U.K.13Oxford Centre for Diabetes, Endocrinology andMetabolism, University of Oxford, Oxford, U.K.14School of Medicine, Trinity College, Dublin,Ireland15Department of Epidemiology, Colorado Schoolof Public Health, University of Colorado AnschutzMedical Campus, Aurora, CO16Division of Population Health and Genomics,Ninewells Hospital and School of Medicine, Uni-versity of Dundee, Dundee, Scotland, U.K.17Department ofMedicine, Universityof Chicago,Chicago, IL18Department of Pediatrics, University of Chi-cago, Chicago, IL19Duke University School of Medicine, Durham,NC20Center forPublicHealthGenomics,UniversityofVirginia, Charlottesville, VA

Wendy K. Chung,1,2 Karel Erion,3

Jose C. Florez,4,5,6,7,8 Andrew T. Hattersley,9

Marie-France Hivert,5,10 Christine G. Lee,11

Mark I. McCarthy,12,13 John J. Nolan,14

Jill M. Norris,15 Ewan R. Pearson,16

Louis Philipson,17,18 Allison T. McElvaine,19

William T. Cefalu,11 Stephen S. Rich,20,21

and Paul W. Franks22,23

Diabetes Care Volume 43, July 2020 1617

CONSEN

SUSREP

ORT

providing the right therapy for the rightpatient at the right time, is not novel.What has changed radically is our abilityto characterize and understand humanbiological variation through 1) assess-ment of the genetic andmetabolic state,2) leveraging data to inform disease cat-egories, and 3) science-guided preventiveand treatment decisions tailored to spe-cific pathological conditions. Couplingthese with detailed information aboutlifestyleandenvironment,availablethroughdigital devices and technologies thatcollect those measures, as well as dataabstracted from electronic medical re-cords, present unparalleled opportuni-ties to optimize diabetes medicine.Diabetes mellitus is diagnosed by the

presence of hyperglycemia that is higherthan a threshold blood glucose concen-tration which predisposes to microvas-cular end-organcomplications.However,hyperglycemia is the end product of nu-merouspathophysiologicalprocesses thatoften emerge over many years and con-verge on the inability of the pancreaticb-cells to secrete enough insulin to meetthe demands of target tissues. In clinicalpractice, absolute insulin deficiency canbe detected from the autoimmune de-struction of b-cells in type 1 diabetes(T1D), which represents ;10% of alldiabetes cases. Making the diagnosis ofT1D is critical for survival, given thetherapeutic requirement of exogenousadministration of insulin. However, lesscommonly, hyperglycemia might derivefrom an inherited or de novo loss offunction in a single gene (e.g., mono-genic diabetes, comprising 2–3% of alldiabetes diagnosed in children or youngadults). Diabetes can also appear afterpancreatitis or organ transplantation,during pregnancy, or as a result of cysticfibrosis.Most individualswith diabetes,however, are likely to be diagnosedwithtype 2 diabetes (T2D), which includesdefects in one or (more often) multiplephysiological pathways (e.g., b-cell

insufficiency, fataccumulationormiscom-partmentalization, inflammation, incretinresistance, dysfunctional insulin signaling).

Our modern capacity to comprehen-sively interrogate diverse axes of biologyhas facilitated the approach of studyinganindividualtoinfergeneralprinciples,fromwhich a discrete treatment plan is se-lected. These axes include developmental/metabolic context, genomic variation,chromatin signals that mark genes asactive or repressed in tissues, expressedtranscripts, biomarkers of disease, and in-creasedknowledgeof lifestyle/environmentalrisk factors. Parallel advances in compu-tational power and analytical methodsrequired to appropriately interrogate“big data” are driving insights that mayradically transform the practice ofmed-icine. Yet, at this time, the individualphysicianoften lacks the timeand trainingneeded to incorporate these insights intomedical decision making. Thus, the trans-lation of the rapidly accumulating newknowledge into practice requires carefulevaluation and translational strategiesinvolving specialist training, education,and policy considerations.

The failure to adequately understandthediversemolecular andenvironmentalprocesses that underlie diabetes and ourinability to identify the pathophysiolog-ical mechanisms that trigger diabetes inindividual patients limit our ability to pre-vent and treat the disease. Public healthstrategies have struggled to slow the ep-idemic, even in countrieswith the greatestfinancial and scientific resources. Pharma-cological therapies,comprising12differentdrugclasses currently approvedby theU.S.Food andDrug Administration (FDA),may,at best, control blood glucose and modifydisease coursebutdonotprovide a cureorresult in the remission of disease. More-over, these agents are sometimes pre-scribed based on nonmedical considerations(cost, side effects, patient preference, orcomorbidities), which may overlook thebiological mechanism. Thus, more people

are developing diabetes worldwide andhave disease progressing to complica-tions, incurring a significant health careburden and cost.

There are, however, several reasonsfor hope. First, diabetes caused by singlegene defects can be characterized andtargeted therapies are particularly effec-tive (1,2). Second, islet autoantibody bio-markers and genomic risk have clarifiedautoimmune diabetes from other formsof the disease (3,4), thereby facilitatingimmune intervention trials and preonsetmonitoring to reduce risk of severe com-plications and aiding in detection of en-vironmental triggers (5). Third, multiplebiomarkers and genetic variants havebeen shown to alter risk of T2D, re-vealing previously unsuspected biolog-ical pathways and providing new targets.Fourth, T2D has been shown to be acomplex combination of multiple con-ditions andprocesses, definedbyprocess-specific subgroups in which individualswith extremeburdens of risk in particularpathways reside and for whom a specifictherapeutic approach may be optimal(6). Finally, the tools, resources, anddata now exist to determine the bio-logical and lifestyle/environmental pre-dictors of drug response, as measuredby a variety of clinical outcomes (7).

THE PRECISION MEDICINE INDIABETES INITIATIVE

The ideaofprecisiondiabetesmedicine isgainingmomentum,basedupontheprom-ise of reducing the enormous and grow-ing burden of diabetes worldwide. Toaddress this, the Precision Medicine inDiabetes Initiative (PMDI) was launchedin 2018 by the American Diabetes Asso-ciation (ADA), in partnership with theEuropean Association for the Study ofDiabetes (EASD). The PMDI has part-nered subsequently with other organ-izations (the U.S. National Institute ofDiabetes and Digestive and Kidney Dis-eases [NIDDK] and JDRF).

21Department of Public Health Sciences, Universityof Virginia, Charlottesville, VA22Genetic and Molecular Epidemiology Unit,Lund University Diabetes Centre, Lund University,Malmo, Sweden23Department of Nutrition, Harvard T.H. ChanSchool of Public Health, Boston, MA

Corresponding author: Paul W. Franks, paul.franks@med.lu.se

S.S.R. and P.W.F contributed equally to thisConsensus Report and are co-chairs of the Pre-cision Medicine in Diabetes Initiative.

M.I.M is currently affiliated with Genentech,South San Francisco, CA

This article is being simultaneously published inDiabetologia (DOI: 10.1007/s00125-020-05181-w)and Diabetes Care (DOI: 10.2337/dci20-0022)

by the European Association for the Study ofDiabetes and the American Diabetes Association.

© 2020 by the American Diabetes Associationand the European Association for the Study ofDiabetes. Readers may use this article as long asthe work is properly cited, the use is educationaland not for profit, and the work is not altered.More information is available at https://www.diabetesjournals.org/content/license.

1618 Consensus Report Diabetes Care Volume 43, July 2020

The mandate of the PMDI is to establishconsensus on the viability and potentialimplementation of precision medicine forthe diagnosis, prognosis, prevention, andtreatment of diabetes, through expert con-sultation, stakeholder engagement, andsystematic evaluation of available evidence.This mandate is pursued in order to realizea future of longer, healthier lives for peoplewith diabetes.The PMDI is focused on assessing

evidence, promoting research, providingeducation, and developing guidelines forthe application of precision medicine indiabetes. The 2019 ADA Scientific Ses-sions (held in June 2019) sponsored aresearch symposium focused on preci-sionmedicine, followed by a PMDI stake-holder meeting (held in October 2019)that was attended by experts in areasgermane to precision diabetes medicinefrom around the world. Future PMDIsymposia will extend the themes of pre-cision diabetes medicine during the2020 ADA Scientific Sessions and EASDAnnual Meeting. In the coming years,educational approaches to translate thescience into practice will be the target ofa series of postgraduate education sym-posia. A global clinical research networkfocused on precision diabetes medicineis also being planned, along with othereducation and information disseminationactivities (see Fig. 1 for an overview of keyobjectives).The purpose of the work underlying

the ADA/EASD PMDI consensus reports,

of which this is the first, is to definerelevant terminology (Text Box 1) andreview the current status of diagnosticsand therapeutics (prevention and treat-ment) in diabetes, including key areas ofopportunity and where further inquiry isneeded (Text Boxes 2–4). Particular focusis placed on elucidating the etiologicalheterogeneity of diabetes, which involvesa combination of approaches includingcontemporaneous measures of risk fac-tors, biomarkers, and genomics, as well aslifestyle and pharmacological interven-tions. Monogenic diabetes is one of fewareas where precision diabetes medi-cine has been proven feasible and ispracticed (as discussed at a recent Di-abetes Care Editors’ Expert Forum; M.C.Riddle, personal communication). Thisfirst Consensus Report does not seekto address extensively the role of pre-cision medicine in the complications ofdiabetes, which is a topic for futureevaluation. In addition, we donot discussdiabetes digital device technology, asthis is addressed in a joint ADA/EASDconsensus report (8,9). A second PMDIconsensus report will be published docu-menting the findings of a systematicevidence review, focusing on precisiondiagnostics and precision therapeutics(prevention and treatment).

An Executive Oversight Committee,comprising representatives from thefounding organizations, ADA (L.P.) andEASD (J.J.N.), and the two co-chairs ofthe initiative (P.W.F. and S.S.R.), provide

PMDI governance. The Executive Over-sight Committee is responsible for en-suring that the PMDI activities are executed.Leadership and direction of the PMDI areprovided bymembers of the PMDI Steer-ing Committee, currently composed ofacademic leaders in precision diabetesmedicine from the U.S. (W.K.C., J.C.F.,J.M.N.)andEurope(A.T.H.,M.I.M.,E.R.P.),a representative fromNIDDK (C.G.L.), andthe ExecutiveOversight Committeemem-bers (L.P., J.J.N., P.W.F., S.S.R.). The Steer-ingCommittee is responsible forprovidingguidance for PMDI activities and engagesin developing precisiondiabetesmedicineeducation, drafting consensus statements,and building interest/working groups toachieve its mission. The Executive Over-sight Committee and the Steering Com-mittee work closely together under thebanner of the PMDI Task Force.Member-ship of the Steering Committee will ex-pand to include experts from around theworld and across multiple areas of ex-pertise germane to the topic of precisiondiabetes medicine.

Work for this Consensus Report beganat theOctober 2019 stakeholdermeetinginMadrid. The meeting included presen-tations and roundtable discussions. Atthe conclusion of the meeting, a writinggroup meeting attended by the PMDITask Force and stakeholders was held todetermine what should be addressed inthe Consensus Report. Following the meet-ing,consensuswasreachedbythePMDITaskForce throughbimonthly calls andelectronic

Figure 1—PMDI activities. PM, precision medicine; RFA, research funding announcement.

care.diabetesjournals.org Chung and Associates 1619

communication. Relevant experts out-side of the Task Force were asked tocontribute sections as needed. The Con-sensus Report was then peer reviewedby experts in the field and by the clinicalcommittees of the founding organiza-tions. The report was then submitted toDiabetes Care and Diabetologia for si-multaneous publication.

PRECISION DIABETES MEDICINE:WHAT IT IS AND WHAT IT IS NOT

Precision diabetes medicine refers to anapproach to optimize the diagnosis, pre-diction, prevention, or treatment of di-abetes by integrating multidimensionaldata, accounting for individualdifferences(Text Box 1). The major distinction fromstandardmedical approaches is the use ofcomplex data to characterize the individ-ual’s health status, predisposition, prog-nosis, and likely treatment response.Precision medicine also focuses on iden-tifying patients who, despite a diagnosis,do not require treatment (or require lessthanmight conventionally be prescribed).

These data may stem from traditionalsources such as clinical records, as wellas from emergent sources of “big data”such as individual medical records fromvery large cohorts of patients; geomo-bility patterns obtained from devices;behavioralmonitors (e.g., actigraphy forexercise and sleep assessments); ingest-ible, subcutaneous, or wearable sensors(e.g., for blood glucose monitoring); andgenomicandother ’omicsdata. Integrationof patient preferences, patient-centeredoutcomes, cost-effectiveness, and shareddecision making will guide how pre-cision diabetes medicine is formulatedand applied.

There are several terms sometimesused interchangeablywith precisionmed-icine, including “personalized medicine,”“individualized medicine,” and “stratifiedmedicine.” The 2020 ADA Standards ofMedical Care in Diabetes (ADA SOC)places considerable emphasis on thepersonalization of diabetes medicine,highlighting that “clinicians care forpatients and not populations” (10) (p.

S2). This reflects the appreciation ofindividual differences with respect tosymptomatology, presentation, behaviors,preferences, social circumstances, responseto treatment, comorbidities, or clinicalcourse. For precision diabetes medicineto be effective, it must be tailored to theindividual. Thus, the ADA SOC instructsthe clinician to adapt guidelines to eachpatient’s characteristics, circumstances,and preferences, including the patient’sfood security, housing, and financial sta-bility. InthecontextofthePMDI,this isnotconsidered to be precision medicine;rather, this final step in the process oftranslating knowledge into practice ispersonalized (or individualized) medi-cine. In contrast, precision (or stratified)medicineemphasizes tailoringdiagnosticsor therapeutics (prevention or treatment)to subgroups of populations sharing sim-ilar characteristics, thereby minimizingerror and risk while maximizing efficacy.Includedwithin precision diabetesmed-icine is the monitoring of disease pro-gression using advanced technologies or

Text Box 1—Definitions

c Precision diagnosis involves refining the characterization of the diabetes diagnosis for therapeutic optimization or to improve prognostic clarityusing information about a person’s unique biology, environment, and/or context.○ Precision diagnostics may involve subclassifying the diagnosis into subtypes, such as is the case in MODY, or utilizing probabilistic algorithmsthat help refine a diagnosis without categorization.

○Careful diagnosis is oftennecessary for successful precision therapy,whether for preventionor treatment. This is truewhere subgroup(s) of thepopulationmust be defined, within which targeted interventionswill be applied, and alsowhere one seeks to determinewhether progressiontoward disease has been abated.

○Precisiondiagnosis canbe conceptualized as a pathway thatmoves through stages, rather than as a single step. Thediagnostic stages include1)an evaluation of prevalence based on epidemiology, including age, or age at diagnosis of diabetes, sex, and ancestry; 2) probability based onclinical features; and 3) diagnostic tests that are interpreted in the light of 1) and 2). A diagnosis in precision medicine is a probability-baseddecision, typically made at a specific point in the natural history of a disease, and neither an absolute truth nor a permanent state.

c Precision therapeutics involves tailoring medical approaches using information about a person’s unique biology, environment, and/or contextfor the purposes of preventing or treating disease (see Precision prevention and Precision treatment, below).

c Precision prevention includes using information about a person’s unique biology, environment, and/or context to determine their likelyresponses to health interventions and risk factors and/or to monitor progression toward disease.○ Precision prevention should optimize the prescription of health enhancing interventions and/or minimize exposure to specific risk factors forthat individual. Precision prevention may also involvemonitoring of health markers or behaviors in people at high risk of disease, to facilitatetargeted prophylactic interventions.

c Precision treatment involves using information about a person’s unique biology, environment, and/or context to guide the choice of anefficacious therapy to achieve the desired therapeutic goal or outcome, while reducing unnecessary side effects.○ Today, the objective of precision therapy is to maximize the probability that the best treatment of all those available is selected for a givenpatient. It is possible that in the future, precision diabetes medicines will be designed according to the biological features of specific patientsubgroups, rather than for the patient population as a whole.

c Precision prognostics focuses on improving the precision and accuracy with which a patient’s disease-related outcomes are predicted usinginformation about their unique biology, environment, and/or context.○ The focus of precisionprognostics includes predicting the risk and severity of diabetes complications, patient-centered outcomes, and/or earlymortality.

c Precision monitoringmay include the detailed assessment of biological markers (e.g., continuous glucose monitoring), behaviors (e.g., physicalactivity), diet, sleep, and psychophysiological stress.○ Precision monitoring can be achieved using digital apps, cutaneous or subcutaneous sensors, ingestible sensors, blood assays etc.○ The intelligent processing, integration, and interpretation of the data obtained throughprecisionmonitoring are key determinants of success.○ Precision monitoring may be valuable for precision prevention (e.g., in T1D), precision diagnostics (e.g., where diagnoses are based on time-varying characteristics), and precision prognostics (e.g., where disease trajectories are informative of the development of key outcomes).

1620 Consensus Report Diabetes Care Volume 43, July 2020

considering how patient features affectthe reliability of assays. The applicationof precision diabetesmedicinemay sub-stantially reduce errors in diagnostic(Fig. 2), therapeutic (Fig. 3), and prognostic(Fig. 4) processes. For example, the in-terrogation of large sets of longitudinalclinical data could identify disease sub-types and match the patient to otherswith a similar disease profile; throughknowledge of treatment efficacy andoutcomes, more precise prognosis andoptimization of therapies for thispatientbyconcordancetosimilarsubgroupswouldemerge (Text Box 1 and Figs. 3 and 4).

PRECISION DIAGNOSTICS

What are the Requirements forPrecision Diagnosis?Precision diagnostics (Text Box 2) em-ploys methods to subclassify patients toenable the successful application of pre-cision medicine approaches (Fig. 2). Thiswill facilitate matching precise preven-tion strategies and treatments to indi-viduals either at risk for or diagnosedwith diabetes. Ideally, a precision diag-nostic test should be 1) robust (hightest-retest reliability within and betweenlaboratories); 2) able to define a discretesubgroup giving insights into disease

etiology, prognosis, and treatment response;3) widely available; 4) easily performedwithaccepted norms for interpretation; 5) in-expensive (or at least cost-effective); and6) approved by regulatory authorities.

Precision diagnosis can be conceptu-alized as a pathway that moves throughstages, rather than as a single step. Thediagnostic stages include assessing the:

c expected prevalence based on epide-miology, including age, or age at di-agnosis of diabetes, sex, and ancestry,

c probable clinical diagnosis using clinicalfeatures and other data, and

Text Box 2—Precision diagnostics: background, barriers to implementation, and research gaps

c Type1diabetes.Bestdiagnostic resultsdependon integratingall diagnosticmodalities, notby relyingonpriorprevalence, clinical features,or testresults in isolation. Theage atwhich the initial islet autoantibody appears and the typeof autoantibody (e.g.,whichof the four primary antibodiesamong ICA512, insulin, GAD, and ZnT8) may be important in defining etiological subtypes of T1D. The majority of the genetic risk of T1D is nowknown,andthesensitivityandspecificityofaT1Dgenetic risk score (T1D-GRS)bothexceed80%.Despite this, ahighT1D-GRSwill have lowpositivepredictivevalue inpatientpopulationswhere theoverallprevalenceofT1D is low, suchas thoseaged.50yearswhendiabetes isdiagnosed. Itwilllikely provemost useful when the T1D-GRS is combinedwith clinical features and islet autoantibodies. At present, there is no immune-based testsufficiently reproducible and robust that it can be used diagnostically.

c Type2diabetes.Categoriesbasedoncluster analysis atdiagnosis canprovide insights into likelyprogression, riskof complications, and treatmentresponse, which offer an exciting approach to subclassification of T2D. At this time, the available genetic data for T2D do not have sufficientpredictive accuracy to replace existing delineative approaches. Although the subcategorization of T2Dusing genetic data is informative regardingtheetiological processes that underlie thedisease, themethodsdescribed so far (6,101) arenot intended tobeused to subclassify a T2Ddiagnosisnor are the existing genetic data sufficient for this purpose for the majority of individuals with T2D. Treatment response and progression can bepredicted from clinical features (137). An advantage of using clinical features for diagnosis of T2D is that they are widely available and easilyobtained (e.g., sex, BMI, HbA1c); however, a potential limitation is that they may vary over time.

c Barriers to implementation. One of several important translational barriers facing the proposed clustering approach for T1D and T2D is thata fasting C-peptide measurement is required at the time of diagnosis, which is not routinely performed in clinical practice, and the reliability ofC-peptide assays varies considerably between laboratories (41). Another limitation is that the biomarkers used to define these clusters changeover timedependingon thedisease courseor its treatment, such that this approach canonly be applied tonewly diagnosed individuals, but not toindividuals years before disease onset or themanymillions of peoplewith long-standing diabetesworldwide.Moreover, the current approachesfor clustering in T2D require continuously distributed data to be categorized, which typically results in loss of power. Thus, thesemethods do notyield good predictive accuracy, a major expectation in precision medicine, but this may change as the approach is refined.

c Research gaps. Based on limited ideal tests and uncertainty in etiology, more research is needed in T1D and T2D in order to define subtypes anddecide the best interventional and therapeutic approaches.

Text Box 3—Precision prevention: background, barriers to implementation, and research gaps

c Type 1 diabetes. In T1D, precision prevention mainly involves the optimization of monitoring methods, thereby facilitating early detection andtreatment. The reasons most prevention trials in T1D have not been effective may include failure to consider the individual’s unique T1D riskprofile (e.g., genetic susceptibility) and their unique response to the preventive agent (immune therapy or dietary intervention). Withoutconsidering the unique genetic profiles of children, interventions aimed at preventing type 1 diabetes (e.g., dietary intervention orimmunotherapy) may be unlikely to succeed. Thus, precision prevention in T1D is likely to involve stratification of at-risk populations andinnovative monitoring technologies.

c Type 2 diabetes. T2D has many avenues for prevention; thus, the possibilities for precision approaches, possibly through tailoring of diet, arebroad. To date, prevention of T2D has focused on peoplewith prediabetes. To be cost-effective, it will likely be necessary to stratify the population withprediabetes such that only those with other relevant risk factors are the focus of preventative interventions. Relevant risk factors may include lifestyle,socioeconomic status, family history, ethnicity, and/or certain biomarker profiles, including genetics.

c Barriers to implementation. The effective implementation of precision prevention will require that appropriate technologies are available, thegeneral public has the willingness to embrace the approach and that those in greatest need can access precision prevention programs. Acommunication plan used by the interventionalist and the patient’s perception of risk should be a focus of precision prevention strategies.

c Research gaps. There are critical areas of research required for implementation of precision prevention in diabetes, including determining forwhomonline care ismore effective than in-person care, the types of staff delivering the lifestylemodification programs, the impact of group and/or individual interaction, andthe frequencyof suchsessions.There is alsouncertaintyabouthowbest toprovideandsustain lifestylemodification.In addition, emphasis should be placed on identifying profiles that indicate the likely response to specific lifestyle interventions (focusing onspecific diets, exercise programs, and other behavioral factors) and sensitivity to risk factors (such as sleep disturbance, stress, depression, poordiet, sedentary behaviors, smoking, certain drugs, and obesity).

care.diabetesjournals.org Chung and Associates 1621

c modification by diagnostic tests thatare interpreted in the light of preva-lence and diagnosis.

A diagnosis in precision medicine is aprobability-baseddecision, typicallymadeat a point in the natural history of adisease, reflecting neither an absolutetruth nor a permanent state. Presentingthe degree of uncertainty in a mannerthat is intuitive to the patient and prac-titioner is critical if the precision diag-nosis is to be effective.

Precision Diagnosis in Clinical Practice

Interpreting HbA1c in Diagnosis and

Monitoring

Data and outcomes from thewidespreaduseofglycatedhemoglobin (HbA1c), ratherthan blood glucose levels, for diagnosishas led to a precision approach for thediagnosis of diabetes. The level of HbA1c

will depend on factors that impacthemoglobin and red cell stability aswell as average glucose values (10).Genetic testing can reveal unsuspectedvariants that alter HbA1c. Thus, knowl-edge of the patient’s ancestry andspecific genetic information can guideinterpretation of assay results for di-agnosis and the monitoring of bloodglucose.

Diagnosing T1D Versus T2D

Currently, themost commonstep towardprecisiondiagnosis that ismade in clinicaldiabetes medicine is the classification ofT1D versus T2D, the two most prevalentsubcategories with different etiologiesand different treatment requirements.Part of the diagnostic dilemma is thatneither T1D nor T2D are monolithicentities and robust “gold standards”are not universally agreed. Diagnosticissues arise when expected clinical

features are discordant from establishednorms (e.g., people diagnosed with di-abetes who are young and have obesity,or old and slim, orwho are a rare subtypein that clinical setting) (11). Islet auto-antibody positivity varies by clinical set-ting (e.g., in people without diabetes,individuals diagnosed with probable T1Das children, individuals with clinical fea-tures of T2D), resulting in an altered priorprobability of T1D that reflects the dif-ferent prevalence in these diverse set-tings. The best diagnosis depends onintegrating all diagnostic modalities, asdemonstrated in predicting long termC-peptide negativity in individuals diag-nosed with diabetes between 20 and40 years of age, where an integratedmodel outperformed diagnosis based onclinical features, circulating antibodies,or genetics used in isolation (3). Thefrequency of misdiagnosis of T1D and

Text Box 4—Precision medicine approaches to treat diabetes: background, barriers to implementation, and research gaps

c Type 1 diabetes. The only existing therapy is insulin for T1D. Developments in long-acting and glucose-sensitive insulins are improving the healthandwell-being of people with T1D, as are technological advances in continuous glucosemonitoring devices, insulin pumps, closed-loop systems,and the artificial pancreas.

c Type 2 diabetes. It has long been recognized that T2D is heterogeneous in its etiology, clinical presentation, and pathogenesis. Yet, traditionally,trials of therapeutic intervention do not recognize this variation.

c Monogenic formsof diabetes are already amenable to precision treatment, if correctly diagnosed. For example,HNF1A-MODY (MODY3),HNF4A-MODY (MODY1), and ABCC8-MODY (MODY12) are acutely sensitive to the glucose-lowering effects of sulfonylureas. Alternatively, individualswith GCK-MODY (MODY2) can have unnecessary treatments stopped.

c With increasing efforts to map patients with T2D in etiological space using clinical andmolecular phenotype, physiology, and genetics, it is likelythat this increasingly granular view of T2D will lead to increasing precision therapeutic paradigms requiring evaluation and potentialimplementation.Genetic variationnot only can capture etiological variation (i.e., genetic variants associatedwithdiabetes risk) but also variationin drug pharmacokinetics (absorption, distribution, metabolism, and excretion [ADME]) and in drug action (pharmacodynamics).

c In contrast, “true” T2D is a common complex disease characterized by thousands of etiological variants, each contributing to a small extent todiabetes risk. Thus, it remains uncertain that genetic variantswill be identified that are highly predictive of drug outcomes in T2D, even if process-specific polygenic risk scores are derived (where all variants on an etiological pathway are combined to increase power).

c Barriers to implementation. The current and growing burden of diabetes is not fromwesternwhite populations but fromother ethnic groups, inparticular South and East Asians. Yet, these populations are underrepresented in clinical trials and, in particular, in attempts to understandvariation in drug outcomes.○ Because the diabetes phenotype can vary markedly by ethnic group, it is likely that complications and drug outcomes will differ betweenpopulations.

○Many of the approaches gaining traction in precisionmedicine generatemassive data sets that are burdensome to store and require powerfulcomputational servers for analysis.

○Undertaking appropriately designed clinical trials for precision treatments thatmeet the current expectations of regulatory authoritiesmaybechallenging, given the many subgroups within which treatments will need to be evaluated. Innovative clinical trials will likely be needed andreal-world evidence will likely need to be part of the evaluation process.

○Translating complex information topatients aboutgenetic (andother ’omics) tests in a clear, concise, andclinically relevantmannerwill requirehealth care providers to be appropriately trained.

c Research gaps. For drug outcomes, there is a pressing need tomove beyond early glycemic response and examine variation in response in termsof cardiovascular outcomes and mortality rates, especially of the newer agents such as SGLT2i and GLP-1RA, with focus on specific patientsubgroups. Identifying predictive markers (especially genetic markers) of serious adverse events in patients treated with these drugs presentsan additional area urgently in need of greater attention.○ Need for functional studies to determine the mechanism(s) of action underlying specific gene variants○ Need for better understanding of the pathophysiology of diabetes to inform on new therapeutic targets○ Need to study broader populations/ethnic groups○ Need for understanding outcomes of highest relevance to patients○ Need for decision-support tools to implement precision diabetes medicine in clinical practice○ Need to demonstrate that approaches are cost-effective

1622 Consensus Report Diabetes Care Volume 43, July 2020

T2D in middle-aged and elderly adults(11,12) suggests that precise diagnosticapproachesareneeded,especiallyas failureto recognize insulin-deficient states canbe fatal.

Monogenic Diabetes

A Diabetes Care Editors’ Expert Forum(M.C. Riddle, personal communication)has concluded recently that amonogenicdiabetes diagnosis is closest to meetingall criteria for a perfect diagnostic test asit defines a discrete subgroup givinginsights into etiology, prognosis, andtreatment response (1,2). Most casesof monogenic diabetes remain misdiag-nosed. Perhaps the best example of pre-cision diabetes medicine is the excellentand long-lasting glycemic response tooral sulfonylureas in insulin-dependentinfantsdiagnosedwithneonatal diabetescaused by abnormalities in the b-cellpotassium channel (13–17). In GCK-MODY (MODY2), it is established thatpatients do not require (18), or respondto, oral medication (19). Other MODYdiagnoses (HNF1A [MODY3], HNF4A[MODY1] and ABCC8 [MODY12]) areacutely sensitive to the glucose-loweringeffects of sulfonylureas (20–22); how-ever, unless thediagnosis is precise, thesetherapeutic benefits are lost. With theclear benefits of precision diagnosis of

monogenic diabetes, it is important toreduce barriers to its implementation.For example, the cost of performingmolecular genetic testing is high anduniversal testing is not cost-effective. Itis thus necessary to limit testing to thosemost likely to have a monogenic diag-nosis. Moreover, identification proto-cols require prescreening based onclinical features (e.g., family history,age at onset, phenotype including syn-dromic features) and nongenetic testing(islet autoantibodies and C-peptide).

One approach for implementing pre-cisionmedicine in the case of monogenicdiabetes would be to:

c test all infants diagnosed with diabe-tes in the first 6 months of age, be-cause .80% have a monogenic causeof neonatal diabetes;

c use aMODY calculator to identify thosewhose clinical features suggest a highlikelihoodofMODY (www.diabetesgenes.org/mody-probability-calculator/) (23);

c test individuals with pediatric diabeteswhen at least three islet autoantibod-ies are antibody negative (24).

The effective use of these pregeneticselection criteria should greatly improvethe likelihood of correctly diagnosing

monogenic diabetes without the burdenof costly genetic screens. Although di-agnostic molecular genetic testing uti-lizes robust analysis of germline DNA,which is virtually unchanged throughoutlife, there are still issues with its imple-mentation. One issue is the incorrectinterpretation of the genetic informa-tion, leading to inaccurate identificationof causal mutations in both clinical prac-tice and in the published research lit-erature (25). Curation of pathogenicvariants formonogenicdiabetes is criticaland is currently being addressed by in-ternational consortia. As a result of tech-nological advances, multiple causes ofmonogenic diabetes canbe tested for in asingle next-generation sequencing test.This approach is generally advantageousas it does mean that syndromic mono-genic diabetes is diagnosed geneticallywhen the patient presents with isolateddiabetes. Thiswill allow other features tobe examined and treated appropriatelybefore clinical presentation. Examples ofthis are neonatal diabetes (2), HNF1B-MODY (MODY5) (26), WFS1 (Wolframsyndrome) (27), and mitochondrial di-abetes (28). For these patients, the ge-netic diagnosis of diabetes will haveimplications far beyond the prognosisand care of diabetes, as the patient

Figure 2—Precision diagnostics

care.diabetesjournals.org Chung and Associates 1623

with certain types ofmonogenic diabeteswill also be at high risk of developmentaldelay, neurological disease, developmen-tal kidney disease, liver failure, deafness,and cardiomyopathy.

Diagnosing Latent Autoimmune Diabetes

in Adults

Latent autoimmune diabetes in adults(LADA) is not currently recognized by theADA as a formal subtype of diabetes.Nevertheless, LADA reveals some of thedifficulties in diabetes subtyping. It wasshown that the presence of GAD auto-antibodies in patients with T2D wasassociated with progression to early in-sulin therapy (29); yet, controversy re-mains as to whether LADA is a discretesubtype, a milder form of T1D, or amixture of some patients with T1Dand others with T2D. The uncertaintyis increased by variation in the diagnosticcriteria, with initial treatment basedupon physician preference as well asthe patient’s presentation (30). In addi-tion, among those with GAD autoanti-bodies, the phenotype varies withdifferent autoantibody levels (31).

Subcategories of Common Forms of

Diabetes

The subcategorization of T1D or T2Dmaynot always be the optimal approach forprecision diabetes diagnosis or therapy.

Nevertheless, the ability to delineate T1Dor T2D using nontraditional data andapproaches may lead to improvementsin prevention or treatment of the dis-ease, including diabetes subclassifica-tions beyond T1D or T2D.Subcategories in T1D.The age at which theinitial islet autoantibody appears and thetype of autoantibody (e.g., which of thefour primary antibodies among islet cellautoantigen 512/islet antigen 2 [ICA512/IA-2], insulin, GAD, zinc transporter8 [ZnT8]) may be important in definingetiological subtypes of T1D (32). Datasupporting this potential subcategoryare based upon those diagnosed inthe first 10 years of life and in pre-dominantly white European popula-tions. The relevance to other ethnicgroups and those diagnosed later inlife is uncertain.

Thegenetic variants accounting for themajority of risk of T1D are now known,and the sensitivity and specificity of T1Dgenetic risk scores (T1D-GRS) both ex-ceed 80% (5,33–35); however, a highT1D-GRSwill have lowpositive predictivevalue in populations with a typically lowprevalence. A T1D-GRS may prove mostuseful when integrated with clinical fea-tures and islet autoantibodies (3,4).There is variation in the genetic suscep-tibility with age at diagnosis but, at

present, genetics is not suggested asan approach for defining subtypes ofT1D.

There is strong evidence for enrich-ment of immune cell types that areassociated with genetic risk of T1D, par-ticularly T cells (CD41 and CD81) and Bcells (CD191). However, at present, thereis no immune-based test sufficiently re-producible and robust that it can be useddiagnostically for T1D.

Persistent endogenous b-cell functionin T1D is associated with greater poten-tial for improved glycemic control andreduced complications (36). A stimulatedC-peptide measurement represents acandidate for defining subcategories ofT1D with different treatment aims.C-peptide levels exponentially fall inthe “honeymoon period” after T1D di-agnosis (37) but have been shown to bestable 7 years after diagnosis (38). Per-sistent C-peptide is associated with alater age of diagnosis, although thereare few data to predict those likely tomaintain high levels of C-peptide.Subcategories in T2D. Family history ofT2D, as a surrogate for precise geneticevaluation, fails to meet many of thecriteria of a robust test as any assessmentchanges over time and depends on therelatives selected for reporting the “fam-ily.” The value of a family history may be

Figure 3—Precision therapeutics

1624 Consensus Report Diabetes Care Volume 43, July 2020

greatest in monogenic diabetes, in whicha pedigree will often demonstrate a pat-ternof inheritanceconsistentwithasinglegene disorder and a consistent phenotype.T2D treatment response and disease

progression can be predicted from con-tinuous clinical features with specificmodels. Thesemodels appear to performbetter than dividing into cluster-basedsubgroups (7). An advantage of usingclinical features is that they are widelyavailable and easily obtained (e.g., sex,BMI, HbA1c). However, they are limitedby the fact that clinical features may varyover time and with the natural history ofthedisease. Incorporationof longitudinalchange with treatment response couldbe a strength as the model’s predictionwould change in concert with changes inthe phenotype of the patient.Recent research has attempted to de-

fine subcategories of T2D (and T1D)based on cluster analysis at diagnosisto provide insights into likely progres-sion, riskof complications, and treatmentresponse (39,40). Barriers facing this andother approaches include collection ofdata that are not routinely obtained(e.g., a fasting C-peptide at the timeof diagnosis, with considerable variationin results between laboratories [41]) andthe change in biomarkers over time thatare dependent on disease course or its

treatment. Genetic data have been usedtodefineT2D subcategories by clusteringgenetic variants that associatewith phys-iological traits and which are correlatedwith clinical outcomes (6). At this time,the available genetic data for T2D andthe clustering does not have sufficientpredictive accuracy to replace existingdelineative approaches. None of themethods described above are estab-lished for subclassification of T2D inclinical practice; nevertheless, it is truethat in a minority of patients, theirspecific type of diabetes may be ade-quately characterized using genetic clus-tering (42,43).

PRECISION THERAPEUTICS

Accurate diagnosis is necessary for suc-cessful precision therapy, whether forprevention or treatment (Fig. 3). This istruewhere subgroup(s) of thepopulationmust be defined to determine whichtargeted interventions will be applied,aswell as for determination of treatmentoutcome. In monogenic diabetes, thereare no currently known options for pre-vention. In T1D, precision preventioncurrently involves mainly the optimiza-tion of monitoring methods (Text Box 3),thereby facilitating timely early detec-tion, preventing early complications

and allowing appropriate treatment. Incontrast, T2D has many avenues forprevention; thus, the possibilities forprecision approaches, possibly throughtailoring of lifestyle (e.g., diet), are broadin T2D.

Precision Prevention in Diabetes(Text Box 3)

Type 1 Diabetes

T1D is characterized by damage, impair-ment, and eventual destruction of theinsulin-producing pancreatic b-cells,thought to be the result of an autoim-mune process. T1D progression has beengrouped into discrete “stages” (44).Stage 1 is defined by the presenceof $2 islet autoantibodies, with normalblood glucose; stage 2 is defined by thepresence of$2 islet autoantibodies withelevation of blood glucose, signaling thefunctional impairment of theb-cells; andstage 3 is characterized by symptoms ofdysglycemia, such as polyuria or diabeticketoacidosis, although not all symptomsneed be present. A clinical diagnosis ofT1D typically is not given until stage 3.T1D is nearly inevitable once $2 isletautoantibodies appear, particularly inthose of younger age, with a lifetimediabetes risk approaching 100% (45,46).Approximately half of the risk of T1D isdue to genetic factors, with over 30% of

Figure 4—Precision prognostics

care.diabetesjournals.org Chung and Associates 1625

the genetic risk attributable to genes ofthe human leukocyte antigen (HLA) com-plex but also includingmore than 50 non-HLA loci (35). Unknown environmentalfactors are thought to trigger the auto-immune process that results in initialb-cell damage and progression towardsymptomatic T1D (47).Primary prevention trials in genetically

susceptible individuals who have notyet developed autoantibodies (i.e., pre-stage1)and secondaryprevention trials inchildren with stages 1 and 2 have beenconducted (48) using dietary interven-tions and immune-targeting approaches.Dietary manipulation studies have beenlargely unsuccessful in reducing islet au-toimmunity (49–51) or T1D (52). Previousintervention studies among individualsat stage 1 or stage 2 have been unableto slow, halt, or reverse the destructionof insulin-producing b-cells. Of ninecompleted secondary prevention trials(53–60), only one (using an anti-CD3antibody) has shown a slight delay inprogression to T1D (61).Most prevention trials in T1D have not

been effective, partially because theunique T1D genetic risk profile of theindividual and their unique response tothepreventiveagent (immune therapyordietary intervention) have not been con-sidered. For example, the inflammatoryresponse to infection with enterovirusesimplicated in the onset of T1D has beenshown to be genetically mediated (62)and diet has had different effects ondevelopment of autoimmunity and pro-gression to T1D (63) dependent on

genetic risk. Several studies have sug-gested that susceptibility to islet auto-immunity and progression to T1Dmay berelated to the ability to adequately usevitamin D, as higher cord blood 25-hydroxyvitamin D was associated witha decreased risk of T1D, but only inchildren who were homozygous for avitamin D receptor gene (VDR) variant(64). Risk of islet autoimmunity wasobserved with reduced dietary intakeof the n-3 fatty acid a-linolenic acid, butonly in those with a specific genotype inthe fatty acid desaturase gene (FADS)cluster (65). Thus, without consideringthe unique genetic profiles of children,dietary supplementation may not besuccessful, arguing for an appropriatelyvalidated precision approach.

Type 2 Diabetes

The emergence of T2D as a global publichealth crisis during recent decades hasmotivated numerous large randomizedcontrolled trials assessing the efficacy ofpharmacological or lifestyle interven-tions for prevention. An emphasis hasbeen placed on intervening in peoplewith “prediabetes,” defined as a personwith levels of fasting blood glucose, 2-hblood glucose, or HbA1c that are chron-ically elevated but below the diagnosticthresholds for diabetes. Although pre-diabetes is a major risk factor for T2Dand other diseases (66), intervening ineveryone with prediabetes may not becost-effective (67). Aggressive precisionprevention in those with relevant riskfactors is discussed in the current ADA

SOC (68). Youth with prediabetes should bethe focus of preventive interventions, es-pecially those with overweight or obesityandwho have one ormore additional riskfactors (e.g., maternal history or exposureto gestational diabetes mellitus [GDM], apositive family history of diabetes in first-or second-degree relatives, signs of insulinresistance, or specific high-risk ancestry).

Multiple interventions in adults withT2D have been evaluated for risk reduc-tion and prevention, both in the shortand the long term. A recent systematicreview (69) reported that after activeinterventions lasting from 6 monthsto.6years, relativerisk reductionachievedfrom lifestyle interventions (39%) was simi-lar to that attained fromuseof drugs (36%);however, only lifestyle interventions had asustained reduction in risk once the in-tervention period had ended. Analysis ofthe postintervention follow-up period (;7years) revealed a risk reduction of 28%with lifestyle modification compared witha nonsignificant risk reduction of 5%from drug interventions.

Most lifestyle intervention programsusestandardizedapproachesdesignedtochange diet and exercise habits for re-ducing body weight. The Diabetes Pre-vention Program (DPP) evaluated theefficacy of lifestyle intervention andmet-formin therapy, compared with standardof care and placebo (control), for delay orprevention of diabetes in those withimpaired glucose regulation at baseline.Although the reductions in diabetesrisk from lifestyle (58% reduction) andmetformin (31% reduction) compared

Text Box 5—Precision medicine approaches to lessen treatment burden and improve quality of life

c Diagnosis and disease management. A more specific diagnosis has the potential to reduce uncertainty and manage future expectations aboutdisease course. This is clearly the case for somemonogenic formsof diabetes,wherediagnosis is nearly certain, given its stronggenetic indication,and the specific treatment is coupled to the subcategory (genetic subtype) of disease. Emerging knowledge regarding subtypes of T2D indicatesthat there is potential to classify individuals with diabetes at risk for progression to complications.

c Misdiagnosis. Inaccurate classificationof the type of diabetes, either from lack of precisionor inadequate clinical attention to detail at the timeofpresentation, can have long-lasting, adverse effects onmental health and quality of life. In the pediatric and younger adult population, the risk ofmisclassification is increasing as both “true” T1Dand “true” T2D classifications are confused through the growing obesity epidemic in youth (T2D)and older ages at onset (T1D). In addition, monogenic variants of diabetes can be misdiagnosed as either T1D or T2D. A precision approach todiagnosis with appropriate standardized laboratory support and increased research to obtain novel biomarkers of disease has the potential tosolve this problem.

c Complications.Worry about complications is an issue for all people with diabetes. Currently, people with diabetes (either T1D or T2D) are givena label of being unequivocally at risk for reduced life span, amputation, kidney failure, and blindness. A more precise diagnosis, prognosis, andstrategy to predict and prevent complications has the potential to greatly reduce disease burden and distress and improve quality of life.Nevertheless, there is also a risk that more precise prognostification may cause distress if the options for successful intervention are limited orincompatible with the patient’s needs or desires.

c Stigmatization.Amajor burden for peoplewith diabetes is that the disease is often considered the fault of the patient. This is particularly true forT2D, as it is often labeled as “just” a lifestyle disease. Clinical care of those with diabetes often results in a singular approach to treatment,regardless of their specific needs, life situation, and other conditions. A clinical process thatmakes diagnosismore precise and includes a patient-oriented evaluation and response to needs has the potential to lessen stigma and reduce associated distress.

1626 Consensus Report Diabetes Care Volume 43, July 2020

with the control intervention were im-pressive (70), there was considerablevariation across the study population(71), with many participants developingT2D during the active intervention period(the first 2.8 years of the trial). Thus, theDPP lifestyle intervention did not truly“prevent” diabetes. Indeed, in the de-cade after randomization, during whichparticipants were offered lifestyle rein-forcement semiannually, the averageduration before disease onset was ;3years (72). Those participants in the DPPwho progressed most rapidly were thosewho lost the least weight in the earlystages of the intervention (73), withgenetic variants representing significantpredictors of peakweight loss andweightloss maintenance (74). Results from theDPP and other large prevention trialssuggest that a “one-size-fits-all” lifestyleintervention strategy will not be effica-cious for everyone, particularly if it can-not be sustained, strengthening the casefor precision lifestyle interventions inT2D prevention.Although precision diabetes medicine

is muchmore than genetics, themajorityof relevant research has focused on

evaluating the role of genetic variantsin precision prevention. Large epidemi-ological studies (75) and interventiontrials (76,77) strongly suggest that stan-dard approaches for lifestyle modifica-tion are equally efficacious in preventingdiabetes regardless of the underlyinggenetic risk. This contrasts with theextensive epidemiological evidence sug-gesting that the relationship of lifestylewith obesity is dependent on genetic risk(78–81); however, with few exceptions(e.g., [74]), analyses in large randomizedcontrolled trials have failed to show thatthese same genetic variants modifyweight loss in response to lifestyle in-tervention (82). It is also important torecognize that knowledge of increasedgenetic risk for diabetes may not moti-vate improvements in lifestyle behaviors.Indeed, knowledge of increased geneticrisk for diabetes may decrease motiva-tion to modify behavior in genetic fatal-ists (83).

Diet recommendations optimized tothe individual have been shown to re-duce postprandial glycemic excursionsto a greater extent than standardapproaches in healthy individuals (84).

Meal compositions that induce the mostfavorable glycemic profiles have beenguided by models derived from an indi-vidual’s biological data (e.g., microbiome,genome, and metabolome), informationon lifestyle factors (e.g., sleep and exer-cise), and postprandial glycemia followingthe consumption of a series of standard-ized meals. Although these studies indi-cate that personalized diet plansmay helpminimize postprandial glycemic excur-sions, no studies have reported thelong-term impact of adhering to person-alized diets on glycemic control.

Of the 12 approved classes of diabetesdrugs, many having been assessed forefficacy in prevention. Overall, drugs thatenhance insulin action haveprovenmoreeffective in diabetes prevention thanthose that increase insulin secretion.Some of the variability in the diabetes-reducing effect of metformin in the DPPhas been associated with variation in theSLC47A1 gene that encodes the multi-drug and toxin extrusion 1 (MATE1)transporter protein (85). In the DPP Out-comes Study, the effects of lifestyle,metformin, and placebo interventionson weight reduction during the 6–15

Figure 5—The path to precision diabetes medicine. HEA, health economic assessment. Adapted from Fitipaldi et al. (136).

care.diabetesjournals.org Chung and Associates 1627

years that followed the end of therandomized intervention phase wereassessed (86). As a percentage of base-lineweight, those assigned tometforminmaintained an average weight loss of6.2% compared with the lifestyle inter-vention group, which maintained aweight loss of 3.7%, and the placebogroup, which maintained a weight lossof 2.8%. In the subgroup of DPP partic-ipants who lost ,5% baseline weight at1 year post-randomization (poor res-ponders), bodyweight during the follow-ing 14 years remained essentially unchanged,whether receiving metformin or placebointerventions. In contrast, those partic-ipants in the lifestyle intervention groupwho lost ,5% baseline weight gainedand sustained;2 kg excess body weightin the years that followed. These findingsreveal a subgroup of DPP participantsin whom lifestyle intervention led toweight gain, which presents a potentialavenue for stratified intervention, whereindividuals who are unlikely to respondwell to lifestyle modification might bebetter served by other therapeuticapproaches.

Precision Treatment (Text Box 4)Once diabetes develops, a variety oftherapeutic steps may be clinically in-dicated to improve disease manage-ment. These steps include:

c glucose monitoringc patient education and lifestyle inter-vention (87)

c surgeryc drug treatments to lower HbA1cc drug treatments to lower cardiovascu-lar risk (e.g., statins, antihypertensives)

c drug treatments targeting specificcomplications (e.g., ACE inhibitors/angio-tensin II receptor blockers [ARBs] andsodium–glucose cotransporter 2 [SGLT2]inhibitors for proteinuric kidney disease,fibrates for retinopathy, atypical analge-sics for painful neuropathy, and statinsand antihypertensives for cardiovasculardisease)

For each of these treatments, therewill be patients who respond well andthose who respond less well, in additionto those who have adverse outcomesfromthe therapy. Thus, precision treatmentcan be considered as using patient char-acteristics to guide the choice of anefficacious therapy to achieve the

desired therapeutic goal or outcomewhile reducing unnecessary side ef-fects (Fig. 3). Given the broad scopeof precision treatment, pharmacolog-ical therapy in T2D has the best evi-dence base for precision therapeuticsat present.

Subcategories and Drug Outcomes

Traditionally, trials of therapeutic inter-ventions do not recognize variation inetiologic processes that lead to develop-ment of T2D. The MASTERMIND consor-tium recently reanalyzed data from theA Diabetes Outcome Progression Trial(ADOPT) and Rosiglitazone Evaluatedfor Cardiac Outcomes and Regulationof Glycaemia in Diabetes (RECORD)studies in order to highlight how clin-ical phenotype can be used to helpguide treatment intervention. In ADOPT,on average, men without obesity showeda greater HbA1c reduction over 5 yearswith sulfonylureas than they did withthiazolidinediones; however, womenwithobesity treated with thiazolidinedioneshad sustained HbA1c lowering over the5 years compared with sulfonylureas(88). When considering the clinical andphysiological variables used to subgroupindividuals with diabetes (39), the in-sulin-resistant cluster defined in ADOPTand RECORD responded better to thia-zolidinediones while the older patientcluster responded better to sulfonyl-ureas (7).

Similar studies have been undertakento investigate how simple clinical varia-bles can be used to predict glycemicresponse to dipeptidyl peptidase 4 inhib-itors (DPP4i). In studies undertaken usingprospective (Predicting Response to In-cretin Based Agents in Type 2 Diabetesstudy [PRIBA]) and primary care data inthe U.K. (Clinical Practice Research Data-link [CPRD]), an insulin-resistant pheno-type of obesity and high triacylglycerolswas associated with reduced initial re-sponse toDPP4i andmore rapid failure oftherapy (89).

As outlined under PRECISION DIAGNOSTICS

and elsewhere (the upcoming ExpertForum), the most current examples ofhow genetics impacts precision treat-ment can be seen in monogenic diabe-tes, for which single gene mutations arecausal for the development of diabetesand for which targeted treatments can,in effect, bypass the etiological defect(e.g., sulfonylurea sensitivity in HNF1A-

MODY [MODY3] [20] and insulin inde-pendence with high-dose sulfonylureasin neonatal diabetes due to KATP channeldefects [14]). In some instances, pre-cision treatment may result in cessa-tion of unnecessary medication, as isthe case in people with GCK-MODY(MODY2), where blood glucose remainssomewhat elevated, but stable, overtime.

Unlike monogenic forms of diabetes,T2D is a common complex disease char-acterized by thousands of etiologicalgene variants. It is uncertain whetherindividual genetic variants will be highlypredictive of drug outcomes. Similar tothe underlying genetic architecture ofT2D, it is possible that drug response inT2D will be influenced by many geneticvariants of small to modest effect. Ge-netic studies of drug response in T2Dhave largely been based on candidategenes of known etiological processes ordrug pathways. These studies have beenlimited in their success. For example,some studies have shown that theKCNJ11/ABCC8 E23K/S119A risk variantincreases glycemic response to sulfony-lureas (90–92); in contrast, the TCF7L2diabetes risk variant reduces glycemicresponse to sulfonylureas (93–95). ThePPARG Pro12Ala diabetes risk varianthas been associated with reduced gly-cemic response to thiazolidinediones(96–98).

Genome-wide association studies(GWAS) have the potential to providenovel insights as they make no assump-tions about drug mechanism or diseaseprocess, in contrast to candidate gene/pathway studies. Only GWAS of metfor-min have been reported to date (99,100),identifying that variants at the ATM/NPAT and SLC2A2 loci are associatedwith an altered glycemic response. InSLC2A2, the noncoding rs8192675 vari-ant C allele is associated with greaterresponse to metformin and is associatedwith reduced expression of the SLC2A2transporter in liver, intestines, and kid-neys. In individuals with obesity, thosewith two copies of the C allele had anabsolute HbA1c reduction of ;1.55%(compared with a reduction of ;1.1%in those without the C allele). While thismay appear to be a small difference, theSLC2A2 genotype effect is the equivalentof a difference in metformin dose of550 mg, or about half the average effectof starting a DPP4i.

1628 Consensus Report Diabetes Care Volume 43, July 2020

When considering etiological varia-tion, recent work partitioning diabe-tes-associated genetic variants by theirpresumed etiological process (parti-tioned polygenic scores) (6,42,101)may define genetically driven dominantprocesses. These processes, such asb-cell dysfunction, lipodystrophy, orobe-sity, could respond differently to drugsthat act on these pathways, such assulfonylureas, glucagon-like peptide1 re-ceptor agonist (GLP-1RA), DPP4i, andthiazolidinediones.Genetic variation can not only capture

etiological variation but also variation indrug pharmacokinetics (absorption, dis-tribution, metabolism, excretion [ADME])and in drug action (pharmacodynamics).Studies of ADME genes have revealedsome variants with a moderate to largeeffect. For example, the 8% of the whitepopulation who carry two loss-of-functionvariants in CYP2C9 are 3.4 times morelikely to achieve HbA1c target than thosewith normal function cytochrome P450family 2 subfamily C member 9 (CYP2C9)due to reduced metabolism of sulfonyl-ureas and increased serum concentrations(102). SLCO1B1 and CYP2C8 genotypesthat alter liver uptake andmetabolism ofrosiglitazone can alter glycemic response(HbA1c) by as much as 0.7% (103). Whilethese studies have promoted pharmaco-genetic approaches in precision diabetestherapeutics, some studies have beensurprisingly negative. For example, loss-of-function variants in the SLC22A1 gene,encoding the organic cation transporter1 (OCT1), which transports metformininto the liver (104,105), donot reduce theglucose-lowering efficacy of metforminin patients with T2D (106,107). Thus,there is genetic evidence that metformindoes not work to lower glucose solely viahepatic mechanisms.The diabetes phenotype is markedly

different across ethnic groups; thus, it islikely that drug outcomes will differbetween populations. The current andgrowing burden of diabetes is growingrapidly in all populations, particularly inSouth and East Asians, yet these pop-ulations are underrepresented in clinicaland drug outcomes trials. A lack ofsystematic reviews and meta-analysesfrom these high-prevalence regions stillpoints to differences in drug response.For example, the DPP4i response isgreater in Asian than white people(108), a result supported by a subgroup

analysis of the Trial Evaluating Cardio-vascular Outcomes with Sitagliptin (TE-COS) showing a greater HbA1c reductionto sitagliptin in East Asians comparedwith white individuals (109). Glycemicresponse to metformin has also beenreported to differ by ethnic group, withAfrican American individuals having agreater response than European Amer-icans (110).

At this time, it is evident that we havethe potential to use simple clinical (e.g.,BMI, sex, ethnicity), physiological, andgenetic variables to predict who is moreor less likely to benefit from a treatment.The reducing costs of genotyping panelsmean that genotype information couldpotentially be available at the point ofprescribing, when the modest effect sizesdescribedmay start to have clinical utility.There is a need to develop implementa-tion and evaluation strategies to assessthe effectiveness and cost-effectivenessof such approaches compared with con-ventional treatment approaches.

PRECISION APPROACHES TODIABETES IN PREGNANCY

In women, being affected by GDM is amajor risk factor for T2D. The risk ofdeveloping T2D in women with priorGDM approaches 70% after the indexpregnancy (111), climbing to an 84% riskof developing T2D in women of EastIndian ancestry (112). Currently, geneticstudies of GDM have identified thosevariants known to increase risk of T2D(113); however, other variants have beenshown to influence glycemic traits spe-cifically inpregnancy (114). Furthermore,like T2D, GDM is a heterogeneous con-dition linked to primary defects in eitherinsulin secretion or sensitivity (115,116).GDM can also result from monogenicforms of diabetes, as numerous studieshave shown. Models that attempt topredict pregnancy complications (117)or subsequent T2D (118) in GDM usingclinical characteristics, biomarkers, and/or genetic variants have yet to be adopted,even though both lifestyle interventionsand metformin use have demonstratedbenefits in reducing the risk of T2D inwomen with prior GDM (119).

The target for all patients with T1D orT2D in pregnancy is to achieve as nearnormal glucose as possible, particularlyaround the timeof conception (to reducedevelopmental anomalies) and in the

third trimester (to reduce the risk ofmacrosomia) (120). In pregnancy, theonly clear exception so far is for motherswithGCK-MODY (MODY2) as fetal growthis determined predominantly by fetalgenotype (121). In mothers whose fetusinherits the mother’s GCK-MODY muta-tion, fetal growth is normal despite thematernal hyperglycemia; thus, treatmentof the maternal hyperglycemia is notrecommended (121,122). Establishingwhether the fetus is likely to be affectedis usually determinedbyultrasound scan.In the future, the use of noninvasive cell-free DNA methods in maternal bloodwill likely establish fetal risk (123). InGDM, whether maternal hyperglycemiais closely monitored and treated in thethird trimester is based on the degree ofhyperglycemia determined by an oralglucose tolerance test at 24–28 weeks’gestation (10). In the future, this decisioncould be modified by nonglycemic factorsthat impact fetal growth.

PATIENT-CENTERED MENTALHEALTH AND QUALITY-OF-LIFEOUTCOMES

Precision diabetes medicine holds thepromise of reducing uncertainty by pro-viding therapies that are more effective,less burdensome, andwith fewer adverseoutcomes, which ultimately improve qual-ity of life and reduce premature death (seeText Box 5). Highly relevant in this contextis mental health (e.g., risk of distress anddepression), yet little has been done toinvestigate how precision medicine mightplay a useful role in improving mentalhealth outcomes.

Depression and anxiety are twice ascommon in people with diabetes than inthe general population, occurring in upto 20% of adult patients (124). Distressoccurs in ;30% of people with diabetes(125) reflecting the emotional and psy-chological burden that comes with di-abetes and its complications, the lifeadjustments it requires, and anxietyabout hypoglycemia or the impact onthe fetus for GDM. Distress has beenreported as being more common inpatients in secondary rather than pri-mary care and in populations with non-European ancestry. Depression is morecommon in lower- and middle-incomecountries, where ;75% of people withT2D reside (125). Both depression anddistress in diabetes are more common in

care.diabetesjournals.org Chung and Associates 1629

those who progress from oral agents toinsulin therapy (126). The onset of com-plications with the initiation of a morecomplexpatternof treatment is associatedwith increased rates of depression (126).There are key points in the life course

of a person with diabetes when bothrational and irrational fears are oftenelevated, typically coinciding with “events,”including:

c increased medication dosec transition to insulin or other injectablesor devices

c emergence of complications or wors-ening of complications

c following a severe hypoglycemic eventc change in diabetes care provider.

Inmany cases, patient self-evaluationsmay be distorted at these times becausethe patient attributes blame for thedisease to themself, the future feelsuncertain and distress peaks. In thesetting of precision diabetes medicine,providers should assess symptoms ofdiabetes distress, depression, anxiety,disordered eating, and cognitive capaci-ties using appropriate standardized andvalidated tools at the initial visit, atperiodic intervals, and when there is achange in disease, treatment, or lifecircumstance (127), information that,when combined with other data, arelikely to improve the precision of clinicaldecision making.Psychological counseling can help pa-

tients understand and manage theiremotional reactions to major eventsby developing a more optimistic outlookandmore realistic,modulated, and adap-tive emotional reactions (128). Precisionmedicine may be used in the future tohelp predict the frequency and extent ofemotional crises. As a result, precisiondiabetes medicine may lessen the patientburden, help patients to objectivize theirdisease, and provide targets for behav-ioral and point-of-care interventions atcriticalmoments in the clinical care cycle.Effective and tailored education and pro-fessional counseling will be necessary tomitigate the risk that a clearer prognosismay raise anxiety about the future forsome patients.

EQUITY IN PRECISION DIABETESMEDICINE

Theexperiencewithmonogenic diabeteshas shown that there is a large degree of

regional, national, and international var-iation in how, and howoften, these casesare diagnosed (1,129,130). This variationis, in part, due to differences in access togeneral medical care and treatments,access to relevant health care professio-nals with the necessary education, tra-ining, and experience, and access tolaboratories with the necessary experi-ence, assays, and standards (131). Aprecision approach to diabetes carewill require that the relevant laboratorymethods and assays are carefully stan-dardized and comparable. Assessmentsthat need to be standardized include:

c T1D-associated autoantibodiesc C-peptidec clinical genetic/genomic risk scoresc decision-support interpretation.

A challenge is that the frequency ofvarious diabetes phenotypes and riskgenotypes may vary by regions of theworld and between ethnicities within aregion. For example, T2Doftenmanifestsvery differently in Native Americans thanin people of European ancestry, withNative Americans tending to developdiabetes at a much younger age andexperience loss of b-cell function earlierin the life course of the disease (132).Recent insights following the ADA Pre-cision Diabetes Medicine meeting inMadrid (held in October 2019) confirmthat case-based interactive learning is anexcellent way to support this type ofpostgraduate education for clinicians atall levels of training.

THE ROAD TO IMPLEMENTATION

Advances in science allow for generationof large-scale biological and physiologicaldata that can be harnessed for precisiondiagnostic (Fig. 2), therapeutic (Fig. 3),and prognostic (Fig. 4) purposes. Pro-grams are needed to train, foster, andretain individuals with biological anddata science expertise who will contrib-ute to precision diabetes medicine ef-forts. Furthermore, clinicians, scientists,and regulators must collaborate to de-velop standards and safeguards for pro-tecting the accumulated “precise” data,which in some instances may lead tounintended and sensitive revelations, onindividuals in a secure manner acrosspopulations and across countries.World-wide differences in prevalence of the

forms of diabetes necessitates inclusionof currently understudied populationsfor the development of precision diag-nostics and therapeutics. As a result, theprecise subtype of diabetes a particularindividual is diagnosed with may vary indifferent populations based on subtypefrequency or genetic or dietary or life-style differences.

The communication strategy used bythe interventionalist and the patient’sperception of risk may be importantfactors contributing to the successfulimplementation of precision diabetesmedicine. Both personal and societalbarriersmay exist to the implementationof precision prevention across geo-graphic regions and countries. Discus-sions with global and regional regulatoryagencieswill be needed todetermine thelevel of evidence needed for approvaland adoption of precision diagnosticsand therapeutics. The development oftools and strategies to synthesize patientdata and facilitate shared decision mak-ing will be needed to translate evidencefor precision diabetes medicine into in-dividualized diabetes care, accountingfor patient preferences and behaviors,health literacy, and socioeconomic con-siderations. Pragmatic studiesofdecision-support systems utilizing rich informationin these health care systems, particularlythose with biobank-linked electronichealth care records, are needed to guideimplementation of precision diabetesmedicine into clinical practice and togenerate the much needed cost-efficacydata for broader adoption.

BUILDING PARTNERSHIPS

Partnerships must be established be-tween the scientific community, pa-tients, health care systems, providers,payors, industry, and regulatory bodiesinvolved in the development, evaluation,approval, adoption, and implementationof precision diagnostics, monitoring, andtherapeutics that aredeemedacceptablefor safe, efficacious, and cost-effectiveuse in precision diabetes care. Makingthe most of the opportunities offered byprecision diabetes medicine will requiremany different stakeholders to formhighly effective partnerships. Withoutnetworks of partnerships that span aca-demic institutions, corporations, payors,regulators, andmedical and public interestgroups with shared understanding and

1630 Consensus Report Diabetes Care Volume 43, July 2020

vision (Fig. 5), precision diabetes medi-cine is destined to fail. Partners inmakingprecision diabetes medicine a realityinclude:People with diabetes. People with di-

abetes are the most important sta-keholders. In Western countries,between 1 in 10 and 1 in 20 peoplehave diabetes, while in other parts ofthe world, diabetes is more prevalent(1 in 3 in some Middle Eastern popula-tions (133), and 1 in 2 in some NativeAmerican tribes [132]). The precisionapproach to diabetes will require effec-tive patient-facing, bidirectional commu-nication strategies that explain whatprecision medicine is and how it works.People with diabetes should be invitedto contribute to research through advisoryand advocacy positions, to contribute topostgraduate educational programs forclinicians and to play a central role indiscussions with politicians, regulators,and payors.Regulatory agencies. The transition

from current diabetes clinical practiceto a precision medicine approach willhave important implications for the de-velopment, prescription, and regulationof diagnostics and therapeutics. Involve-ment of regulators at the earliest stagesof the precision diabetes medicine work-flow will be critical to the successfulimplementation of the precision ap-proach. Recognizing these challenges,the FDA and the European MedicinesAgency have initiated discussions relat-ing to standards for evidence and thedesign of future clinical trials for pre-cision diabetes medicine (134).Payors. Payment for medical care re-

lated to diabetes varies greatly, includ-ing between regions within countries,with costs for diabetes often hidden inother areas of medical care. Fragmenta-tion of sites of delivery for diabetes careand its costs directly impact paymentpolicies. There is evidence in the case ofmonogenic diabetes that a precisionmedicine approach is cost-effective(135). The delay, or prevention, of com-plications (the major contributor to di-abetes costs) through precision diabetesmedicine may be the strongest driver foradoption.Product manufacturers. Diabetes

technology, including the developmentof wearable devices for glucose mo-nitoring and for regulating insulin infu-sions (i.e., the artificial pancreas), has

developed rapidly and is an exampleof widespread personalized diabetesmedicine. Technology and pharmaceu-tical implementation is currently at apre-precision level, and treatmentguidelines are quite generic. The Euro-pean Federation of PharmaceuticalIndustries and Associations (EFPIA) Di-abetes Platform, in which six leadingpharmaceutical companies are develop-ing shared policy goals focused on im-proving diabetes clinical outcomes, hasinitiated multiple projects with strongprecision diabetes medicine agendas,with other public-private partnershipsfocused on precision diabetes medicineunderway (136).

Private and public supporters of re-search. Support for diabetes researchfunding has struggled as its priorityhas fallen among the general publicand some political decision makers,where cancer and cardiovascular diseaserank consistently higher thandiabetes onthe public agenda. For precision diabetesmedicine to meaningfully improve thelives of patients, it will be necessary tobuild highly effective networks of keystakeholders, such that common agen-das are agreed to and funding for re-search and implementation is madeavailable. This, in turn, requires thatthe evidence justifying a precision di-abetes medicine approach is clearly ar-ticulated to all major decision makers,including funders.

Clinicians and professional organiza-tions. Medical care for the person withdiabetes involves a wide spectrum ofhealth care providers, including tertiaryand secondary specialists, general intern-ists, primary care doctors, nurses, die-titians, podiatrists, pharmacists, andother paramedical professionals. Severalorganizations are engaged in the PMDI(ADA, EASD,NIDDK) and representativesof professional bodies in Asia, Africa, andelsewhere are being engaged by thePMDI to ensure global impact. Tailoringeducational modules and content todifferent professional and cultural set-tings is ideally suited to these partnerorganizations.

General public. The enormous burdenthat diabetes places on many healthcare systems is usually shouldered bythe general public, owing to the highcosts of treating the disease and loss ofpublic revenue through decreased pro-ductivity. The effective implementation

of precision prevention will requirethat the general public embraces theapproach and that those in greatest needcan access precision prevention programs.Diabetes messaging for the general publiccan be modeled on precision oncology,for which public advocacy and engage-ment have been successful, effectivelyutilizing social media as well as traditionalmedia to communicate not only itsstrengths and weaknesses but also itsbenefits and risks.

SUMMARY AND FUTUREPERSPECTIVES

Precision diabetes medicine has found afirm foothold in the diagnosis and treat-ment of monogenic diabetes, while theapplication of precision medicine toother types of diabetes is at this timeaspirational, rather than standard ofcare. The ability to integrate the diag-nosis ofmonogenic diabetes into routineclinical care is one example where diag-nostics are essential and meet many ofthe characteristics of the ideal test. De-spite an excellent diagnostic paradigm,there are no known avenues for pre-vention inmonogenic diabetes, althoughcareful monitoring in presymptomaticvariant carriers may lead to early de-tection of diabetes and rapid treatment.

Future precision diabetes medicineapproaches are likely to include diagnos-tic algorithms for defining diabetessubtypes in order to decide the bestinterventional and therapeutic ap-proaches. The scope and potential forprecision treatment in diabetes is vast,yet deep understanding is lacking. It willbe imperative to determine when andhow the application of therapeutics inprecision diabetes medicine improvesoutcomes in a cost-effective fashion.

There are many important stakehold-ers whose engagement will be necessaryfor the implementation of precision di-abetes medicine to succeed (Fig. 5).Progress in translating advances in bio-logy and technology will be governed bythe identification, accurate measure-ment, and scalable deployment of agentsfor diagnosis and therapy, so broadstakeholder engagement is essential. Itis crucial that precision approaches areavailable to the full diversity of humanpopulations and societal contexts, suchthat precision diabetes medicine doesnot widen health disparity but achieves

care.diabetesjournals.org Chung and Associates 1631

the greatest benefits to all individualsand society as a whole. Highly functionalpartnerships with patient representa-tives and public organizations will berequired to reap the benefits of precisiondiabetes medicine.

Acknowledgments. The authors thank P. Sim-ing (Lund University) for editorial assistance, H.Fitipalidi (LundUniversity) for assistancewith thedesign of the figures and Prof. H. Mulder (LundUniversity) for technical critique. The authorsacknowledge the invited peer reviewers whoprovided comments on an earlier draft of thisreport: Helen Colhoun (University of Edinburgh),Boris Draznin (University of Colorado School ofMedicine), Torben Hansen (University of Copen-hagen), Pal Njølstad (University of Bergen), andMatthew C. Riddle (Oregon Health and ScienceUniversity).Funding. Funding for the PMDI is from theAmerican Diabetes Association. In-kind supporthas been provided by the academic institutionsof each Task Force member. The ideas andopinions expressed in this report were derivedinpart fromworkundertakenbythecoauthors, forwhich they report the following support: W.K.C.(NIH: R01DK52431, P30DK26687, U54 TR001873,and U54DK118612); A.T.H. (Wellcome Trust Se-nior Investigator: 098395; National Institute forHealth Research [NIHR] Senior Investigator andsupport of Exeter NIHR Clinical Research Facility;Medical ResearchCouncil [MRC]:MR-K005707-1);M.-F.H. (ADA 1-15-ACE-26, NIH 5R01HD94150-02);M.I.M. (Wellcome Trust Senior Investigator andNIHR Senior Investigator: 203141, 212259, 098381;NIDDK: U01-DK105535); J.M.N. (NIH: R01 DK104351,R21 AI142483); E.R.P. (Wellcome Trust New In-vestigator award: 102820/Z/13/Z); L.P. (NIH:R01DK104942, P30 DK02059, U54DK118612);S.S.R. (NIH: DP3 DK111906, R01 DK122586; Uni-versity of Virginia Strategic Investment Fund SIF88);P.W.F. (European Research Council: CoG-2015_681742_NASCENT; Swedish Research Council; NovoNordisk Foundation; European Diabetes ResearchFoundation; Swedish Heart Lung Foundation; In-novative Medicines Initiative of the European Union:no. 115317 –DIRECT and no. 115881 – RHAPSODY;no. 875534 – SOPHIA).Duality of Interest. W.K.C. is on the scientificadvisory board of the Regeneron Genetics Cen-ter. J.C.F. has received a speaking honorariumfrom Novo Nordisk and consulting fees fromJanssen Pharmaceuticals. M.I.M. has in the past3 years served on advisory panels for Pfizer, NovoNordisk A/S, and Zoe Global Ltd.; has receivedhonoraria fromMerck, Pfizer, Novo Nordisk, andEli Lilly; and has received research funding fromAbbvie, AstraZeneca, Boehringer Ingelheim, EliLilly, Janssen, Merck, Novo Nordisk A/S, Pfizer,Roche, Sanofi, Servier, and Takeda. As of June2019, M.I.M. is an employee of Genentech and aholder of Roche stock. E.R.P. has received re-search funding from Boehringer Ingelheim, EliLilly, Janssen, Novo Nordisk A/S, Sanofi, andServier and honoraria from Eli Lilly. L.P. hasreceived research funding from Janssen andProvention Bio. P.W.F. has received researchfunding from Boehringer Ingelheim, Eli Lilly,Janssen, Novo Nordisk A/S, Sanofi, and Servier;

received consulting fees from Eli Lilly, NovoNordisk, and Zoe Global Ltd.; and has stockoptions in Zoe Global Ltd. No other potentialconflicts of interest relevant to this article werereported.

References1. Hattersley AT, Patel KA. Precision diabetes:learning frommonogenic diabetes. Diabetologia2017;60:769–7772. De Franco E, Flanagan SE, Houghton JA, et al.The effect of early, comprehensive genomictesting on clinical care in neonatal diabetes:an international cohort study. Lancet 2015;386:957–9633. OramRA,Patel K,Hill A, et al.A type1diabetesgenetic risk score canaid discriminationbetweentype 1 and type 2 diabetes in young adults.Diabetes Care 2016;39:337–3444. Grubb AL, McDonald TJ, Rutters F, et al. Atype 1 diabetes genetic risk score can identifypatients with GAD65 autoantibody-positivetype 2 diabetes who rapidly progress to insulintherapy. Diabetes Care 2019;42:208–2145. Sharp SA, Rich SS, Wood AR, et al. Develop-ment and standardization of an improved type 1diabetes genetic risk score for use in newbornscreening and incident diagnosis. Diabetes Care2019;42:200–2076. Udler MS, McCarthy MI, Florez JC, Mahajan A.Genetic risk scores for diabetes diagnosis andprecisionmedicine.EndocrRev2019;40:1500–15207. Dennis JM, Shields BM, HenleyWE, Jones AG,Hattersley AT. Disease progression and treat-ment response in data-driven subgroups oftype 2 diabetes compared with models basedon simple clinical features: an analysis usingclinical trial data. Lancet Diabetes Endocrinol2019;7:442–4518. Fleming GA, Petrie JR, Bergenstal RM, HollRW,PetersAL, HeinemannL.Diabetesdigital apptechnology: benefits, challenges, and recommenda-tions.AconsensusreportbytheEuropeanAssociationfor the Study of Diabetes (EASD) and the AmericanDiabetes Association (ADA) Diabetes TechnologyWorking Group. Diabetes Care 2020;43:250–2609. Fleming GA, Petrie JR, Bergenstal RM, HollRW,PetersAL, HeinemannL.Diabetesdigital apptechnology: benefits, challenges, and recom-mendations.A consensus reportby theEuropeanAssociation for the Study of Diabetes (EASD) andthe American Diabetes Association (ADA) Dia-betes Technology Working Group. Diabetologia2020;63:229–24110. American Diabetes Association. 2. Classifi-cation and Diagnosis of Diabetes: Standards ofMedical Care in Diabetesd2020. Diabetes Care2020;43(Suppl. 1):S14–S3111. Thomas NJ, Jones SE, Weedon MN, ShieldsBM, Oram RA, Hattersley AT. Frequency andphenotype of type 1 diabetes in the first sixdecades of life: a cross-sectional, geneticallystratified survival analysis from UK Biobank.Lancet Diabetes Endocrinol 2018;6:122–12912. Thomas NJ, Lynam AL, Hill AV, et al. Type1 diabetes defined by severe insulin deficiencyoccurs after 30 years of age and is commonlytreatedas type2diabetes. Diabetologia 2019;62:1167–117213. Gloyn AL, Pearson ER, Antcliff JF, et al.Activating mutations in the gene encoding theATP-sensitive potassium-channel subunit Kir6.2

and permanent neonatal diabetes. N Engl J Med2004;350:1838–184914. Pearson ER, Flechtner I, Njølstad PR,et al.; Neonatal Diabetes International Col-laborative Group. Switching from insulin tooral sulfonylureas in patients with diabetesdue to Kir6.2 mutations. N Engl J Med 2006;355:467–47715. Babenko AP, Polak M, Cave H, et al. Acti-vating mutations in the ABCC8 gene in neonataldiabetes mellitus. N Engl J Med 2006;355:456–46616. Sagen JV, Raeder H, Hathout E, et al. Per-manent neonatal diabetes due to mutations inKCNJ11 encoding Kir6.2: patient characteristicsand initial response to sulfonylurea therapy.Diabetes 2004;53:2713–271817. BowmanP,SulenA,Barbetti F,etal.;NeonatalDiabetes International Collaborative Group. Ef-fectiveness and safety of long-term treatmentwith sulfonylureas in patients with neonataldiabetes due to KCNJ11 mutations: an interna-tional cohort study. Lancet Diabetes Endocrinol2018;6:637–64618. Steele AM, Shields BM, Wensley KJ,Colclough K, Ellard S, Hattersley AT. Prevalenceof vascular complications among patients withglucokinase mutations and prolonged, mild hy-perglycemia. JAMA 2014;311:279–28619. Stride A, Shields B, Gill-Carey O, et al. Cross-sectional and longitudinal studies suggest phar-macological treatment used in patients withglucokinase mutations does not alter glycaemia.Diabetologia 2014;57:54–5620. Pearson ER, Starkey BJ, Powell RJ, GribbleFM, Clark PM, Hattersley AT. Genetic cause ofhyperglycaemia and response to treatment indiabetes. Lancet 2003;362:1275–128121. Pearson ER, Pruhova S, Tack CJ, et al. Mo-lecular genetics and phenotypic characteristicsof MODY caused by hepatocyte nuclear factor4alpha mutations in a large European collection.Diabetologia 2005;48:878–88522. Bowman P, Flanagan SE, Edghill EL, et al.Heterozygous ABCC8 mutations are a cause ofMODY. Diabetologia 2012;55:123–12723. Shields BM,McDonald TJ, Ellard S, CampbellMJ, Hyde C, Hattersley AT. The development andvalidation of a clinical predictionmodel to determinetheprobabilityofMODY inpatientswithyoung-onsetdiabetes. Diabetologia 2012;55:1265–127224. Carlsson A, Shepherd M, Ellard S, et al.Absence of islet autoantibodies and modestlyraised glucose values at diabetes diagnosisshould lead to testing for MODY: lessonsfrom a 5-year pediatric Swedish national cohortstudy. Diabetes Care 2020;43:82–8925. Ellard S, ColcloughK, Patel KA, HattersleyAT.Prediction algorithms: pitfalls in interpretinggenetic variants of autosomal dominant mono-genic diabetes. J Clin Invest 2020;130:14–1626. ClissoldRL,HamiltonAJ,HattersleyAT, EllardS, Bingham C. HNF1B-associated renal and extra-renal disease-an expanding clinical spectrum.Nat Rev Nephrol 2015;11:102–11227. Tranebjaerg L, Barrett T, RendtorffND.WFS1Wolfram syndrome spectrum disorder. InGeneReviews�. Adam MP, Ardinger HH, PagonRA, et al., Eds. Seattle, WA, University of Wash-ington, Seattle, 1993-2020 [Internet]. Availablefrom https://www.ncbi.nlm.nih.gov/pubmed/20301750. Accessed 13 May 2020

1632 Consensus Report Diabetes Care Volume 43, July 2020

28. Murphy R, Turnbull DM, Walker M,Hattersley AT. Clinical features, diagnosis andmanagement of maternally inherited diabetesand deafness (MIDD) associated with the3243A.G mitochondrial point mutation. DiabetMed 2008;25:383–39929. Tuomi T, Groop LC, Zimmet PZ, Rowley MJ,Knowles W, Mackay IR. Antibodies to glutamicacid decarboxylase reveal latent autoimmunediabetes mellitus in adults with a non-insulin-dependent onset of disease. Diabetes 1993;42:359–36230. Brophy S, Yderstraede K, Mauricio D, et al.;Action LADA Group. Time to insulin initiation cannotbe used in defining latent autoimmune diabetes inadults. Diabetes Care 2008;31:439–44131. Hawa MI, Kolb H, Schloot N, et al.; ActionLADA consortium. Adult-onset autoimmune di-abetes in Europe is prevalentwith abroad clinicalphenotype: Action LADA 7. Diabetes Care 2013;36:908–91332. Battaglia M, Ahmed S, Anderson MS, et al.Introducing the endotype concept to address thechallenge of disease heterogeneity in type 1diabetes. Diabetes Care 2020;43:5–1233. Onengut-Gumuscu S, Chen WM, RobertsonCC, et al.; SEARCH for Diabetes in Youth; Type 1Diabetes Genetics Consortium. Type 1 diabetesrisk in African-Ancestry participants and utility ofan ancestry-specific genetic risk score. DiabetesCare 2019;42:406–41534. RichSS.Geneticsand itspotential to improvetype 1 diabetes care. Curr Opin Endocrinol Di-abetes Obes 2017;24:279–28435. Onengut-Gumuscu S, Chen WM, Burren O,et al.; Type 1 Diabetes Genetics Consortium. Finemapping of type1diabetes susceptibility loci andevidencefor colocalizationof causalvariantswithlymphoid gene enhancers. Nat Genet 2015;47:381–38636. Rickels MR, Evans-Molina C, Bahnson HT,etal.; T1DExchangeb-Cell FunctionStudyGroup.High residual C-peptide likely contributes toglycemic control in type 1 diabetes. J Clin Invest2020;130:1850–186237. HaoW,GitelmanS,DiMeglioLA,BoulwareD,Greenbaum CJ; Type 1 Diabetes TrialNet StudyGroup. Fall in C-peptide during first 4 years fromdiagnosis of type 1 diabetes: variable relation toage,HbA1c, and insulin dose. Diabetes Care 2016;39:1664–167038. Shields BM,McDonald TJ, OramR, et al.; TIGIConsortium. C-peptide decline in type 1 diabeteshas two phases: an initial exponential fall and asubsequent stablephase.DiabetesCare2018;41:1486–149239. Ahlqvist E, StormP,KarajamakiA, et al.Novelsubgroups of adult-onset diabetes and theirassociation with outcomes: a data-driven clusteranalysis of six variables. Lancet Diabetes Endo-crinol 2018;6:361–36940. Zaharia OP, Strassburger K, Strom A, et al.;German Diabetes Study Group. Risk of diabetes-associated diseases in subgroupsof patientswithrecent-onset diabetes: a 5-year follow-up study.Lancet Diabetes Endocrinol 2019;7:684–69441. Little RR, Rohlfing CL, Tennill AL, et al.Standardization of C-peptide measurements.Clin Chem 2008;54:1023–102642. Udler MS, Kim J, von Grotthuss M, et al.;Christopher D. Anderson on behalf of META-STROKEandthe ISGC.Type2diabetesgenetic loci

informed by multi-trait associations point todisease mechanisms and subtypes: A soft clus-tering analysis. PLoS Med 2018;15:e100265443. Mahajan A, Wessel J, Willems SM, et al.;ExomeBP Consortium; MAGIC Consortium; GI-ANT Consortium. Refining the accuracy of val-idated target identification through codingvariant fine-mapping in type 2 diabetes. NatGenet 2018;50:559–57144. Insel RA, Dunne JL, Atkinson MA, et al.Staging presymptomatic type 1 diabetes: a sci-entific statement of JDRF, the Endocrine Society,and theAmericanDiabetesAssociation.DiabetesCare 2015;38:1964–197445. Krischer JP; Type 1 Diabetes TrialNet StudyGroup. The use of intermediate endpoints in thedesign of type 1 diabetes prevention trials.Diabetologia 2013;56:1919–192446. Ziegler AG, Rewers M, Simell O, et al. Se-roconversion to multiple islet autoantibodiesand risk of progression to diabetes in children.JAMA 2013;309:2473–247947. Rewers M, Stene LC, Norris JM. Chapter 11:Risk factors for type 1 diabetes. In Diabetes inAmerica, 3rd ed. Cowie CC, Casagrande SS,Menke A, et al., Eds. Bethesda, MD, NationalInstitutes of Health, 201848. Skyler JS, Kirchner JP, Becker D, Rewers M.Chapter 37: Prevention of type 1 diabetes InDiabetes in America, 3rd ed. Cowie CC,Casagrande SS, et al., Eds. Bethesda, MD, Na-tional Institutes of Health, 201849. KnipM, AkerblomHK, Becker D, et al.; TRIGRStudy Group. Hydrolyzed infant formula andearly b-cell autoimmunity: a randomized clinicaltrial. JAMA 2014;311:2279–228750. Vaarala O, Ilonen J, Ruohtula T, et al. Re-moval of bovine insulin from cow’s milk formulaand early initiation of beta-cell autoimmunity inthe FINDIA pilot study. Arch Pediatr AdolescMed2012;166:608–61451. Hummel S, Pfluger M, Hummel M, BonifacioE, Ziegler AG. Primary dietary intervention studyto reduce the risk of islet autoimmunity in childrenat increased risk for type 1 diabetes: the BABYDIETstudy. Diabetes Care 2011;34:1301–130552. KnipM, AkerblomHK, Al Taji E, et al.;WritingGroup for the TRIGR Study Group. Effect ofhydrolyzed infant formula vs conventional for-mula on risk of type 1 diabetes: the TRIGRrandomized clinical trial. JAMA 2018;319:38–4853. Nanto-Salonen K, Kupila A, Simell S, et al.Nasal insulin to prevent type 1 diabetes inchildrenwith HLA genotypes and autoantibodiesconferring increased risk of disease: a double-blind, randomised controlled trial. Lancet 2008;372:1746–175554. Krischer JP, Schatz DA, Bundy B, Skyler JS,Greenbaum CJ; Writing Committee for the Type 1DiabetesTrialNetOral Insulin StudyGroup. Effectof oral insulin on prevention of diabetes inrelatives of patients with type 1 diabetes: a ran-domized clinical trial. JAMA 2017;318:1891–190255. Elding Larsson H, Lundgren M, Jonsdottir B,Cuthbertson D, Krischer J; DiAPREV-IT StudyGroup. Safety and efficacy of autoantigen-specifictherapy with 2 doses of alum-formulated gluta-mate decarboxylase in children with multipleislet autoantibodies and risk for type 1 diabetes:a randomized clinical trial. Pediatr Diabetes2018;19:410–419

56. Lampeter EF, KlinghammerA, ScherbaumWA,et al.; DENIS Group. The Deutsche NicotinamideIntervention Study: an attempt to prevent type 1diabetes. Diabetes 1998;47:980–98457. Gale EA, Bingley PJ, Emmett CL, Collier T;European Nicotinamide Diabetes InterventionTrial (ENDIT) Group. European Nicotinamide Di-abetes Intervention Trial (ENDIT): a randomisedcontrolled trial of intervention before the onsetof type 1 diabetes. Lancet 2004;363:925–93158. Diabetes Prevention Trial–Type 1 DiabetesStudy Group. Effects of insulin in relatives ofpatients with type 1 diabetes mellitus. N Engl JMed 2002;346:1685–169159. Vehik K, Cuthbertson D, Ruhlig H, Schatz DA,Peakman M, Krischer JP; DPT-1 and TrialNetStudy Groups. Long-term outcome of individualstreated with oral insulin: Diabetes PreventionTrial–Type 1 (DPT-1) oral insulin trial. DiabetesCare 2011;34:1585–159060. Vandemeulebroucke E, Gorus FK, DecochezK, et al.; Belgian Diabetes Registry. Insulin treat-ment in IA-2A-positive relativesof type1diabeticpatients. Diabetes Metab 2009;35:319–32761. Herold KC, Bundy BN, Long SA, et al.; Type1 Diabetes TrialNet Study Group. An anti-CD3antibody, teplizumab, in relatives at risk fortype 1 diabetes. N Engl J Med 2019;381:603–61362. Hober D, Alidjinou EK. Enteroviral pathogen-esis of type1 diabetes: queries and answers. CurrOpin Infect Dis 2013;26:263–26963. Hakola L, Takkinen HM, Niinisto S, et al.Infant feeding in relation to the risk of advancedislet autoimmunity and type 1 diabetes in chil-dren with increased genetic susceptibility: a co-hort study. Am J Epidemiol 2018;187:34–4464. Tapia G, Marild K, Dahl SR, et al. Maternaland newborn vitamin D-binding protein, vitaminD levels, vitamin D receptor genotype, and child-hood type 1 diabetes. Diabetes Care 2019;42:553–55965. Norris JM, Kroehl M, Fingerlin TE, et al.Erythrocyte membrane docosapentaenoic acidlevelsareassociatedwith isletautoimmunity: theDiabetes Autoimmunity Study in the Young.Diabetologia 2014;57:295–30466. DECODE Study Group; the European Diabe-tes Epidemiology Group. Glucose tolerance andcardiovascular mortality: comparison of fastingand 2-hour diagnostic criteria. Arch Intern Med2001;161:397–40567. Diabetes Prevention ProgramResearchGroup.Within-trial cost-effectiveness of lifestyle interven-tion or metformin for the primary prevention oftype 2 diabetes. Diabetes Care 2003;26:2518–252368. American Diabetes Association. 12. Olderadults: Standards of Medical Care in Diabetesd2020. Diabetes Care 2020;43(Suppl. 1):S152–S16269. Haw JS, Galaviz KI, Straus AN, et al. Long-term sustainability of diabetes prevention ap-proaches: a systematic review andmeta-analysisof randomized clinical trials. JAMA Intern Med2017;177:1808–181770. Knowler WC, Barrett-Connor E, Fowler SE,et al.; Diabetes Prevention Program ResearchGroup. Reduction in the incidence of type 2diabetes with lifestyle intervention or metfor-min. N Engl J Med 2002;346:393–40371. Crandall J, Schade D, Ma Y, et al.; DiabetesPrevention Program Research Group. The in-fluence of age on the effects of lifestyle mod-ification and metformin in prevention of

care.diabetesjournals.org Chung and Associates 1633

diabetes. J Gerontol A Biol Sci Med Sci 2006;61:1075–108172. Knowler WC, Fowler SE, Hamman RF, et al.;Diabetes Prevention Program Research Group.10-year follow-up of diabetes incidence andweight loss in the Diabetes Prevention ProgramOutcomes Study. Lancet 2009;374:1677–168673. Delahanty LM, Pan Q, Jablonski KA, et al.;Diabetes Prevention Program Research Group.Effects of weight loss, weight cycling, andweightloss maintenance on diabetes incidence andchange in cardiometabolic traits in the DiabetesPrevention Program. Diabetes Care 2014;37:2738–274574. Papandonatos GD, Pan Q, Pajewski NM,et al.; GIANT Consortium; Diabetes PreventionProgram and the Look AHEAD Research Groups.Genetic predisposition to weight loss and regainwith lifestyle intervention: analyses from theDiabetes Prevention Program and the LookAHEAD randomized controlled trials. Diabetes2015;64:4312–432175. Langenberg C, Sharp SJ, Franks PW, et al.Gene-lifestyle interaction and type 2 diabetes:the EPIC interact case-cohort study. PLoS Med2014;11:e100164776. Hivert MF, Christophi CA, Franks PW, et al.;Diabetes Prevention Program Research Group.Lifestyle and metformin ameliorate insulin sen-sitivity independently of the genetic burden ofestablished insulin resistance variants in Diabe-tes Prevention Program participants. Diabetes2016;65:520–52677. Hivert MF, Jablonski KA, Perreault L, et al.;DIAGRAMConsortium; Diabetes Prevention Pro-gram Research Group. Updated genetic scorebased on 34 confirmed type 2 diabetes loci isassociated with diabetes incidence and regres-sion to normoglycemia in the diabetes preven-tion program. Diabetes 2011;60:1340–134878. Kilpelainen TO, Qi L, Brage S, et al. Physicalactivity attenuates the influence of FTO variantson obesity risk: a meta-analysis of 218,166adults and 19,268 children. PLoS Med 2011;8:e100111679. Shungin D, DengWQ, Varga TV, et al.; GIANTConsortium. Ranking and characterization ofestablished BMI and lipid associated loci ascandidates for gene-environment interactions.PLoS Genet 2017;13:e100681280. Graff M, Scott RA, Justice AE, et al.; CHARGEConsortium; EPIC-InterAct Consortium; PAGEConsortium. Genome-wide physical activity in-teractions in adiposityda meta-analysis of200,452 adults. PLoS Genet 2017;13:e100652881. Tyrrell J, Wood AR, Ames RM, et al. Gene-obesogenic environment interactions in the UKBiobank study. Int J Epidemiol 2017;46:559–57582. Livingstone KM, Celis-Morales C, PapandonatosGD, et al. FTO genotype and weight loss: systematicreview and meta-analysis of 9563 individual partic-ipant data from eight randomised controlled trials.BMJ 2016;354:i470783. Godino JG, van Sluijs EM,MarteauTM, SuttonS, Sharp SJ, Griffin SJ. Lifestyle advice combinedwith personalized estimates of genetic or pheno-typic risk of type 2 diabetes, and objectivelymeasured physical activity: a randomized con-trolled trial. PLoS Med 2016;13:e100218584. Zeevi D, Korem T, Zmora N, et al. person-alized nutrition by prediction of glycemic re-sponses. Cell 2015;163:1079–1094

85. Jablonski KA,McAteer JB, de Bakker PI, et al.;Diabetes Prevention Program Research Group.Common variants in 40 genes assessed for di-abetes incidence and response tometformin andlifestyle intervention in the diabetes preventionprogram. Diabetes 2010;59:2672–268186. Apolzan JW, Venditti EM, Edelstein SL, et al.;Diabetes Prevention Program Research Group.Long-term weight loss with metformin or life-style intervention in the Diabetes PreventionProgramOutcomes Study. Ann InternMed 2019;170:682–69087. LeanME, LeslieWS, BarnesAC, et al. Primarycare-led weight management for remission oftype 2 diabetes (DiRECT): an open-label, cluster-randomised trial. Lancet 2018;391:541–55188. Dennis JM, Shields BM, Jones AG, PearsonER, Hattersley AT, Henley WE; MASTERMINDconsortium. Evaluating associations betweenthe benefits and risks of drug therapy intype 2 diabetes: a joint modeling approach.Clin Epidemiol 2018;10:1869–187789. Dennis JM, Shields BM, Hill AV, et al.; MAS-TERMIND Consortium. Precision medicine intype 2 diabetes: clinical markers of insulin re-sistance are associated with altered short- andlong-term glycemic response to DPP-4 inhibitortherapy. Diabetes Care 2018;41:705–71290. Feng Y, Mao G, Ren X, et al. Ser1369Alavariant in sulfonylurea receptor gene ABCC8 isassociated with antidiabetic efficacy of gliclazideinChinese type2diabetic patients.DiabetesCare2008;31:1939–194491. Zhang H, Liu X, Kuang H, Yi R, Xing H.Association of sulfonylurea receptor 1 genotypewith therapeutic response to gliclazide in type 2diabetes. Diabetes Res Clin Pract 2007;77:58–6192. Javorsky M, Klimcakova L, Schroner Z, et al.KCNJ11 gene E23K variant and therapeutic re-sponse to sulfonylureas. Eur J Intern Med 2012;23:245–24993. Pearson ER, Donnelly LA, Kimber C, et al.Variation in TCF7L2 influences therapeutic re-sponse to sulfonylureas: a GoDARTs study. Di-abetes 2007;56:2178–218294. Schroner Z, Javorsky M, Tkacova R, et al.Effect of sulphonylurea treatment on glycaemiccontrol is related to TCF7L2 genotype in patientswith type2diabetes. DiabetesObesMetab2011;13:89–9195. Javorsky M, Babjakova E, Klimcakova L, et al.Association between TCF7L2 genotype and gly-cemic control in diabetic patients treated withgliclazide. Int J Endocrinol 2013;2013:37485896. Kang ES, Park SY, Kim HJ, et al. Effects ofPro12Ala polymorphism of peroxisome prolifer-ator-activated receptor gamma2 gene on rosi-glitazone response in type 2 diabetes. ClinPharmacol Ther 2005;78:202–20897. Hsieh MC, Lin KD, Tien KJ, et al. Commonpolymorphisms of the peroxisome proliferator-activated receptor-gamma (Pro12Ala) and per-oxisome proliferator-activated receptor-gammacoactivator-1 (Gly482Ser) and the response topioglitazone in Chinese patients with type 2diabetes mellitus. Metabolism 2010;59:1139–114498. Pei Q, Huang Q, Yang GP, et al. PPAR-g2 andPTPRD gene polymorphisms influence type 2diabetes patients’ response to pioglitazone inChina. Acta Pharmacol Sin 2013;34:255–261

99. Zhou K, Bellenguez C, Spencer CC, et al.;GoDARTS and UKPDS Diabetes Pharmacoge-netics StudyGroup;WellcomeTrust CaseControlConsortium 2; MAGIC investigators. Commonvariants near ATM are associated with glycemicresponse to metformin in type 2 diabetes. NatGenet 2011;43:117–120100. Zhou K, Yee SW, Seiser EL, et al.; MetGenInvestigators; DPP Investigators; ACCORD Inves-tigators. Variation in the glucose transporter geneSLC2A2 is associated with glycemic response tometformin. Nat Genet 2016;48:1055–1059101. Mahajan A, TaliunD, ThurnerM, et al. Fine-mapping type 2 diabetes loci to single-variantresolution using high-density imputation andislet-specific epigenome maps. Nat Genet 2018;50:1505–1513102. Zhou K, Donnelly L, Burch L, et al. Loss-of-function CYP2C9 variants improve therapeuticresponse to sulfonylureas in type 2 diabetes:a Go-DARTS study. Clin Pharmacol Ther 2010;87:52–56103. Dawed AY, Donnelly L, Tavendale R, et al.CYP2C8 and SLCO1B1 variants and therapeuticresponse to thiazolidinediones in patients withtype 2 diabetes. Diabetes Care 2016;39:1902–1908104. Shu Y, Sheardown SA, Brown C, et al. Effectof genetic variation in the organic cation trans-porter 1 (OCT1)onmetforminaction. J Clin Invest2007;117:1422–1431105. Sundelin E, Gormsen LC, Jensen JB, et al.Genetic polymorphisms in organic cation trans-porter 1attenuateshepaticmetforminexposure inhumans. Clin Pharmacol Ther 2017;102:841–848106. Zhou K, Donnelly LA, Kimber CH, et al.Reduced-function SLC22A1 polymorphisms en-coding organic cation transporter 1 and glycemicresponse to metformin: a GoDARTS study. Di-abetes 2009;58:1434–1439107. Dujic T, Zhou K, Yee SW, et al. Variants inpharmacokinetic transporters and glycemic re-sponse to metformin: a MetGen meta-Analysis.Clin Pharmacol Ther 2017;101:763–772108. Kim YG, Hahn S, Oh TJ, Kwak SH, Park KS,Cho YM. Differences in the glucose-loweringefficacy of dipeptidyl peptidase-4 inhibitors be-tween Asians and non-Asians: a systematic re-view and meta-analysis. Diabetologia 2013;56:696–708109. Davis TME, Mulder H, Lokhnygina Y, et al.;TECOS Study Group. Effect of race on the gly-caemic response to sitagliptin: insights from theTrial Evaluating Cardiovascular Outcomes withSitagliptin (TECOS). Diabetes Obes Metab 2018;20:1427–1434110. Williams LK, PadhukasahasramB, AhmedaniBK, et al. Differing effects of metformin onglycemic control by race-ethnicity. J Clin En-docrinol Metab 2014;99:3160–3168111. Kim C, Newton KM, Knopp RH. Gestationaldiabetes and the incidence of type 2 diabetes:a systematic review. Diabetes Care 2002;25:1862–1868112. Das Gupta R, Gupta S, Das A, Biswas T,Haider MR, Sarker M. Ethnic predisposition ofdiabetes mellitus in the patients with previoushistory of gestational diabetesmellitus: a review.Expert Rev Endocrinol Metab 2018;13:149–158113. LoweWL Jr, Scholtens DM, Sandler V, HayesMG. Genetics of gestational diabetes mellitus andmaternal metabolism. Curr Diab Rep 2016;16:15

1634 Consensus Report Diabetes Care Volume 43, July 2020

114. Hayes MG, Urbanek M, Hivert MF, et al.;HAPO Study Cooperative Research Group. Iden-tification of HKDC1 and BACE2 as genes influ-encing glycemic traits during pregnancy throughgenome-wide association studies. Diabetes2013;62:3282–3291115. Powe CE, Allard C, Battista MC, et al.Heterogeneous contribution of insulin sensitivityand secretion defects to gestational diabetesmellitus. Diabetes Care 2016;39:1052–1055116. Benhalima K, Van Crombrugge P,Moyson C,et al. Characteristics and pregnancy outcomesacross gestational diabetes mellitus subtypesbased on insulin resistance. Diabetologia 2019;62:2118–2128117. CooraySD,Boyle JA, SoldatosG,WijeyaratneLA, Teede HJ. Prognostic prediction models forpregnancy complications in women with gesta-tional diabetes: a protocol for systematic review,critical appraisal and meta-analysis. Syst Rev2019;8:270118. Tobias DK. Prediction and prevention oftype 2 diabetes in womenwith a history of GDM.Curr Diab Rep 2018;18:78119. Aroda VR, Christophi CA, Edelstein SL,et al.; Diabetes Prevention Program ResearchGroup. The effect of lifestyle intervention andmetformin on preventing or delaying diabetesamong women with and without gestationaldiabetes: the Diabetes Prevention Program out-comes study 10-year follow-up. J Clin EndocrinolMetab 2015;100:1646–1653120. American Diabetes Association. 14. Man-agement of diabetes in pregnancy: Standards ofMedical Care in Diabetesd2020. Diabetes Care2020;43(Suppl. 1):S183–S192

121. SpyerG,MacleodKM,ShepherdM,EllardS,Hattersley AT. Pregnancy outcome in patientswith raised blood glucose due to a heterozygousglucokinasegenemutation.DiabetMed2009;26:14–18122. Sanyoura M, Letourneau L, KnightJohnson AE, et al. GCK-MODY in the US mono-genic diabetes registry: description of 27 un-published variants. Diabetes Res Clin Pract2019;151:231–236123. DeFrancoE, Caswell R, Houghton JA, IotovaV, Hattersley AT, Ellard S. Analysis of cell-freefetal DNA for non-invasive prenatal diagnosis ina family with neonatal diabetes. Diabet Med2017;34:582–585124. Petrak F, Baumeister H, Skinner TC, BrownA, Holt RIG. Depression and diabetes: treatmentand health-care delivery. Lancet Diabetes Endo-crinol 2015;3:472–485125. Snoek FJ, Bremmer MA, Hermanns N.Constructsofdepressionanddistress indiabetes:time for an appraisal. Lancet Diabetes Endocrinol2015;3:450–460126. Moulton CD, Pickup JC, Ismail K. The linkbetween depression and diabetes: the search forshared mechanisms. Lancet Diabetes Endocrinol2015;3:461–471127. American Diabetes Association. 5. Facili-tating behavior change and well-being to im-prove health outcomes: Standards of MedicalCare in Diabetesd2020. Diabetes Care 2020;43(Suppl. 1):S48–S65128. Fisher L, Polonsky WH, Hessler D. Ad-dressing diabetes distress in clinical care:a practical guide. Diabet Med 2019;36:803–812

129. Shields BM, Hicks S, Shepherd MH,Colclough K, Hattersley AT, Ellard S. Maturity-onset diabetes of the young (MODY): how manycases are we missing? Diabetologia 2010;53:2504–2508130. ShepherdM,ColcloughK,EllardS,HattersleyAT. Ten years of the national genetic diabetesnurse network: a model for the translation ofgenetic information into clinical care. Clin Med(Lond) 2014;14:117–121131. Owen KR. Monogenic diabetes in adults:what are the new developments? Curr OpinGenet Dev 2018;50:103–110132. Poudel A, Zhou JY, Story D, Li L. Diabetesand associated cardiovascular complications inAmerican Indians/Alaskan Natives: a review ofrisks and prevention strategies. J Diabetes Res2018;2018:2742565133. Al Busaidi N, Shanmugam P, Manoharan D.Diabetes in the Middle East: government healthcare policies and strategies that address thegrowing diabetes prevalence in the MiddleEast. Curr Diab Rep 2019;19:8134. Meyer RJ. Precision medicine, diabetes,and the U.S. Food and Drug Administration.Diabetes Care 2016;39:1874–1878135. Naylor R. Economics of genetic testing fordiabetes. Curr Diab Rep 2019;19:23136. Fitipaldi H, McCarthy MI, Florez JC,Franks PW. A global overview of precisionmedicine in type 2 diabetes. Diabetes 2018;67:1911–1922137. O’Brien RL, Brinster RL, Storb U. Somatichypermutation of an immunoglobulin transgenein kappa transgenic mice. Nature 1987;326:405–409

care.diabetesjournals.org Chung and Associates 1635