American Geriatrics Society 2019 Updated AGS Beers ...American Geriatrics Society 2019 Updated AGS Beers Criteria ... Case reports, case series, letters to the editor, and editorials

CLINICAL INVESTIGATION

American Geriatrics Society 2019 Updated AGS Beers Criteria®

for Potentially Inappropriate Medication Use in Older AdultsBy the 2019 American Geriatrics Society Beers Criteria® Update Expert Panel*

The American Geriatrics Society (AGS) Beers Criteria®

(AGS Beers Criteria®) for Potentially Inappropriate Medica-tion (PIM) Use in Older Adults are widely used by clini-cians, educators, researchers, healthcare administrators, andregulators. Since 2011, the AGS has been the steward of thecriteria and has produced updates on a 3-year cycle. TheAGS Beers Criteria® is an explicit list of PIMs that are typi-cally best avoided by older adults in most circumstances orunder specific situations, such as in certain diseases or con-ditions. For the 2019 update, an interdisciplinary expertpanel reviewed the evidence published since the last update(2015) to determine if new criteria should be added or ifexisting criteria should be removed or undergo changes totheir recommendation, rationale, level of evidence, orstrength of recommendation. J Am Geriatr Soc 00:1–21, 2019.

Key words: medications; drugs; older adults; Beers list;Beers Criteria

The American Geriatrics Society (AGS) Beers Criteria®

(AGS Beers Criteria®) for Potentially InappropriateMedication (PIM) Use in Older Adults are widely used byclinicians, educators, researchers, healthcare administrators,and regulators. Since 2011, the AGS has been the stewardof the criteria and has produced updates on a 3-year cyclethat began in 2012.1,2 The AGS Beers Criteria® are anexplicit list of PIMs that are typically best avoided by olderadults in most circumstances or under specific situations,such as in certain diseases or conditions.

For the 2019 update, an interdisciplinary expert panelreviewed the evidence published since the last update(2015) to determine if new criteria should be added or ifexisting criteria should be removed or undergo changes totheir recommendation, rationale, level of evidence, orstrength of recommendation. Each of the five types of cri-teria in the 2015 update were retained in this 2019 update:medications that are potentially inappropriate in most olderadults, those that should typically be avoided in olderadults with certain conditions, drugs to use with caution,drug-drug interactions, and drug dose adjustment based onkidney function.

OBJECTIVES

The specific aim was to update the 2015 AGS Beers Criteria®

using a comprehensive, systematic review and grading of theevidence on drug-related problems and adverse events inolder adults. The strategies to achieve this aim were to:

• Incorporate new evidence on PIMs included in the 2015 AGSBeers Criteria® and evidence regarding new criteria or modifi-cations of existing criteria being considered for the 2019update.

• Grade the strength and quality of each PIM statement based onthe level of evidence and strength of recommendation.

• Convene an interdisciplinary panel of 13 experts in geriatriccare and pharmacotherapy who would apply a modified Delphimethod, informed by the systematic review and grading, toreach consensus on the 2019 update.

• Incorporate exceptions in the AGS Beers Criteria® that thepanel deemed clinically appropriate. These exceptions wouldbe designed to make the criteria more individualized to clinicalpractice and be more relevant across settings of care.

INTENT OF CRITERIA

The primary target audience for the AGS Beers Criteria®

is practicing clinicians. The criteria are intended for use inadults 65 years and older in all ambulatory, acute, andinstitutionalized settings of care, except for the hospiceand palliative care settings. Consumers, researchers, phar-macy benefits managers, regulators, and policymakers alsowidely use the AGS Beers Criteria®. The intention of theAGS Beers Criteria® is to improve medication selection;

From the *American Geriatrics Society, New York, New York.

Address correspondence to Mary Jordan Samuel, American GeriatricsSociety, 40 Fulton St, 18th Floor, New York, NY 10038.E-mail: [email protected]

See related editorial by Michael Steinman et al.

DOI: 10.1111/jgs.15767

JAGS 00:1–21, 2019© 2019 The American Geriatrics Society 0002-8614/18/$15.00

mailto:[email protected]://doi.org/10.1111/jgs.15766http://crossmark.crossref.org/dialog/?doi=10.1111%2Fjgs.15767&domain=pdf&date_stamp=2019-01-29

educate clinicians and patients; reduce adverse drug events;and serve as a tool for evaluating quality of care, cost, andpatterns of drug use of older adults.

As with previously published AGS Beers Criteria®, thegoal of the 2019 update continues to be improving thecare of older adults by reducing their exposure to PIMs thathave an unfavorable balance of benefits and harms com-pared with alternative treatment options. This is accom-plished by using the AGS Beers Criteria® as both aneducational tool and a quality measure—two uses that arenot always in agreement—and the panel considered andvigorously deliberated both. The AGS Beers Criteria® arenot meant to be applied in a punitive manner. Prescribingdecisions are not always clear-cut, and clinicians must con-sider multiple factors, including discontinuation of medica-tions no longer indicated. Quality measures must be clearlydefined, easily applied, and measured with limited informa-tion and, thus, although useful, cannot perfectly distinguishappropriate from inappropriate care. The panel’s review ofevidence at times identified subgroups of individuals whoshould be exempt from a given criterion or to whom a spe-cific criterion should apply. Such a criterion may not be eas-ily applied as a quality measure, particularly when suchsubgroups cannot be easily identified through structuredand readily accessible electronic health data. As an exam-ple, the panel thought that a criterion should not beexpanded to include all adults 65 years and older whenonly certain subgroups have an adverse balance of benefitsvs harms for the medication, or conversely when a sizablesubgroup of older adults may be appropriate candidates fora medication that is otherwise problematic.

Despite past and current efforts to translate the cri-teria into practice, some controversy and myths abouttheir use in practice and policy continue to prevail. Thepanel addressed these concerns and myths by writing acompanion article to the 2015 update of the AGS BeersCriteria® and an updated 2019 short piece, which remainsthe best way to advise patients, providers, and health sys-tems on how to use (and not use) the 2019 AGS BeersCriteria®.3

METHODS

Methods used for the 2019 update of the AGS Beers Criteria®

were similar to those used in the 2015 update, with additionalemphasis on extending the rigor of the evidence review andsynthesis process.2 These methods were based on the Gradingof Recommendations Assessment, Development and Evalua-tion (GRADE) guidelines for clinical practice guideline devel-opment and are consistent with recommendations from theNational Academy of Medicine.4,5

Panel Composition

The AGS Beers Criteria® expert update panel comprised13 clinicians and included physicians, pharmacists, andnurses, each of whom had participated in the 2015 update.Panelists had experience in different practice settings,including ambulatory care, home care, acute hospital care,skilled-nursing facility, and long-term care. In addition,the panel included ex-officio representatives from the Cen-ters for Medicare and Medicaid Services, the National

Committee for Quality Assurance, and the PharmacyQuality Alliance. Potential conflicts of interest were dis-closed at the beginning of the process and before each fullpanel call and are listed in the disclosures section of thisarticle. Panelists were recused from discussion in areas inwhich they had a potential conflict of interest.

Literature Review

Literature searches were conducted in PubMed and theCochrane Library from January 1, 2015, to September30, 2017. Search terms for each criterion included individ-ual drugs, drug classes, specific conditions, and combina-tions thereof, each with a focus on “adverse drug events”and “adverse drug reactions.” Medications believed to havelow utilizations (eg, meprobamate and central α-agonistantihypertensives other than clonidine) or no longer avail-able in the United States were excluded from the literaturesearch. Searches targeted controlled clinical trials, observa-tional studies, and systematic reviews and meta-analyses,with filters for human participants, 65 years and older, andEnglish language. Clinical reviews and guidelines were alsoincluded to provide context. Case reports, case series, lettersto the editor, and editorials were excluded.

Searches identified 17,627 references; 5403 abstractswere sent to panelists for review, of which 1422 referenceswere selected for full-text review. Among these, 377 articleswere abstracted into evidence tables, including 67 systematicreviews and/or meta-analyses, 29 controlled clinical trials,and 281 observational studies.

Development Process

Between February 2016 and May 2018, the full panel con-vened for a series of conference calls and 1 full-day, in-person meeting. In addition, the panel divided into fourwork groups, each assigned a subset of the criteria. Eachwork group led the review and synthesis of evidence for itssubset of the criteria, convening via conference calls andelectronically via e-mail.

The development process began by soliciting ideas fromthe panelists about criteria that should be explored for addi-tion, modification, or removal. Suggestions from otherswere also welcomed. To guide the evidence selection,review, and synthesis process, each work group then under-took an exercise to identify a priori which clinical out-comes, indications, and comparison groups were mostrelevant when considering evidence for each criterion(ie, the “desired evidence” for reviewing each criterion).These discussions were not considered binding but providedguidance for keeping the evidence review and synthesisfocused on what was most clinically relevant.

Each work group reviewed abstracts from the literaturesearches for the criteria in its purview and collectivelyselected a subset for full-text review. This selection processconsidered the methodologic quality of each study, its rele-vance to older adults, and its concordance with the desiredevidence noted above. After reviewing the full text of eachselected article, the work group then decided by consensuswhich articles represented the best available evidence, basedon a balance of these same three key criteria (methodologicquality, relevance to older adults, and concordance with

2 2019 AGS BEERS CRITERIA® UPDATE EXPERT PANEL MONTH 2019–VOL. 00, NO. 00 JAGS

desired evidence). Special emphasis was placed on selectingsystematic reviews and meta-analyses when available,because resource constraints precluded the panel from con-ducting these types of comprehensive analyses. In general, astudy was considered relevant to older adults if the mean ormedian age of participants was older than 65 years, andespecially relevant if most or all participants were olderthan this age threshold.

Articles comprising the best available evidence wereabstracted by AGS staff into evidence tables. These tablessummarized the design, population, and findings of eachstudy, and identified markers of methodologic qualityhighlighted by the GRADE criteria for clinical trials andobservational studies and by A MeaSurement Tool toAssess Systematic Reviews (AMSTAR).6–8 Each work groupthen synthesized evidence for each criterion from the 2015to 2017 literature reviews based on GRADE guidelines andthe American College of Physicians’ evidence gradingframework (Table 1).6,9

Using evidence from the 2015 to 2017 literaturereview, evidence findings from previous updates in 2012and 2015, and clinical judgment, each work group pre-sented to the full panel its findings and suggestions forchanges (or no change) to the criteria, with ensuing discus-sion. For most criteria, a consensus emerged, to leave anexisting criterion from the 2015 update unchanged, to mod-ify it, to remove it entirely, or to add a new criterion. Poten-tial modifications included the drug(s) included in thecriterion, the recommendation, the rationale, the quality ofevidence, and the strength of recommendation. As noted inthe GRADE guidelines, strength of recommendation ratingsincorporate a variety of considerations, including expertopinion and clinical judgment and context, and thus do notalways align with quality of evidence ratings.

After discussion of proposed changes, an anonymous Del-phi process was used to ascertain panel consensus, using afive-point Likert scale with anchors of “strongly disagree” and“strongly agree.” As a general rule, criteria receiving “agree”or strongly agree ratings from more than 90% of panelistswere included. The remainder were brought back for groupdiscussion, with final decisions resolved through consensus.

In addition to changes made on the basis of evidence,the panel decided on several modifications to improve clar-ity and usability of the AGS Beers Criteria®. These includedremoving a number of medications that are used onlyrarely. These removals should not be interpreted as condon-ing use of these medications but rather are intended to“declutter” the AGS Beers Criteria® and not distract frominformation on more commonly used medications. Inselected cases, the panel changed the wording of certain cri-teria, recommendations, and rationale statements toimprove clarity and avoid potential misinterpretations.

The final set of criteria was reviewed by the AGS Exec-utive Committee and Clinical Practice and Models of CareCommittee and subsequently released for public comment.Comments were solicited from the general public and sentto 39 organizations. Comments were accepted over a3-week period from August 13, 2018, until September4, 2018. A total of 244 comments were received from47 individuals (79 comments), 6 pharmaceutical companies(10 comments), and 22 peer organizations (155 comments).All comments were reviewed and discussed by the panel

cochairs. All comments along with proposed changes to thecriteria were shared with the entire panel for final approval.

RESULTS

Noteworthy Changes to PIMs for Older Adults

Tables 2 through 6 show the 2019 criteria. Table 7 liststhose drugs with strong anticholinergic properties that aresometimes referenced in Tables 2 through 6. Comparedwith the 2015 criteria, several drugs were removed fromTable 2 (medications that are potentially inappropriate inmost older adults), Table 3 (medications that are potentiallyinappropriate in older adults with certain conditions), andTable 4 (medications that should be used with caution).These removals are summarized in Table 8 and includeremoval of drugs no longer available in the United States(ticlopidine, oral pentazocine). In other cases, the recom-mendation was removed entirely because the panel decidedthe drug-related problem was not sufficiently unique toolder adults (eg, using stimulating medications in patientswith insomnia or avoiding medications that can lower theseizure threshold in patients with a seizure disorder). Theseremovals do not imply that these medications are now con-sidered safe for older adults; rather, they were made to helpkeep the AGS Beers Criteria® streamlined and focused onmedications particularly problematic for older adults.

The H2-receptor antagonists were removed from the“avoid” list in patients with dementia or cognitive impair-ment. This is because evidence for adverse cognitive effectsin these conditions is weak, and because the panelexpressed concern that the intersection of this criterion withanother criterion that discourages chronic use of proton-pump inhibitors in the absence of strong indications wouldoverly restrict therapeutic options for older adults withdementia who have gastroesophageal reflux or similarissues. However, H2-receptor antagonists remain on the cri-teria as “avoid” in patients with delirium. In addition,wording of this criterion was modified to affirm that non-benzodiazepine, benzodiazepine receptor agonist hypnotics(ie, the “Z drugs”: zolpidem, eszopiclone, and zaleplon)should be avoided in older adults with delirium.

Two drugs with strong anticholinergic properties, pyri-lamine and methscopolamine, were added to the list of anti-cholinergic drugs to avoid. Changes to criteria oncardiovascular drugs include minor updates to the rationaleand a minor change to clarify the recommendation foravoiding digoxin as first-line therapy for atrial fibrillationand heart failure (Table 2). The rationale to avoid sliding-scale insulin has been revised to clarify its meaning andintent (Table 2). Glimepiride has been added to the list ofsulfonylureas with a greater risk of severe prolonged hypo-glycemia (Table 2). The duration of use of metoclopramidehas been added to be consistent with US Food and DrugAdministration labeling (Table 2).

The serotonin-norepinephrine reuptake inhibitors(SNRIs) have been added to the list of drugs to avoid inpatients with a history of falls or fractures (Table 3). Fol-lowing a principle that applies to all criteria, the panel rec-ognizes there may be situations when SNRIs, otherantidepressants, and other medications listed in this crite-rion may be appropriate for people with a history of falls

JAGS MONTH 2019–VOL. 00, NO. 00 2019 AGS BEERS CRITERIA® UPDATE EXPERT PANEL 3

or fractures, based on potential benefits and the lack ofavailability of safer alternatives. After reviewing and dis-cussing the evidence on antipsychotics to treat psychosis inpatients with Parkinson disease, the panel decided toremove aripiprazole as preferred and add pimavanserin.Thus, the 2019 AGS Beers Criteria® recognize quetiapine,clozapine, and pimavanserin as exceptions to the generalrecommendation to avoid all antipsychotics in older adultswith Parkinson disease (Table 3). However, none of thesethree excepted drugs is close to ideal in either efficacy orsafety, each having its own limitations and concerns.

The criteria on drugs to avoid in older adults with heart fail-ure were reorganized to add clinical nuance based on evidence,other guideline recommendations, and clinical considerations.The updated recommendations are that nondihydropyridine cal-cium channel blockers should be avoided in older adults whohave heart failure with reduced ejection fraction; that nonsteroi-dal anti-inflammatory drug (NSAIDs), cyclooxygenase-2 inhibi-tors, thiazolidinediones (“glitazones”), and dronedarone should

be used with caution in older adults with heart failure who areasymptomatic (ie, excellent control of heart failure signs andsymptoms, with or without use of medications) and avoided inolder adults who are symptomatic; and that cilostazol shouldcontinue to be avoided in older adults with heart failure ofany type.

Drugs To Be Used With Caution

Table 4 contains drugs to be used with caution in olderadults. The purpose of this table is to identify drugs forwhich there is some cause for concern, but for which theevidence and/or clinical context is as of yet insufficient tomerit inclusion in the main tables. Compared with the pre-vious update, the following changes and additionswere made:

• The age threshold beyond which extra caution is advised forusing aspirin for primary prevention of cardiovascular disease

Table 1. Designations of Quality of Evidence and Strength of Recommendationsa

Quality of EvidenceQuality of evidence ratings for each criterion are based on synthetic assessment of two complementary approaches to evaluating thequality of evidence.

ACP-based approach9 GRADE-based approach4

High-quality evidence “Evidence…obtained from 1 or more well-designed and well-executed randomized,controlled trials (RCTs) that yield consistent anddirectly applicable results. This also means thatfurther research is very unlikely to change ourconfidence in the estimate of effect.”

Consider the following five factors for the studiesthat comprise the best-available evidence for agiven criterion:1. Risk of bias: Severity of threats to studies’

internal validity (eg, randomized vsobservational design, potential forconfounding, bias in measurement)

2. Inconsistency: Do different studies providesimilar or different estimates of effect size

3. Indirectness: How relevant are the studies tothe clinical question at hand (eg, nature ofstudy of population, comparison group, typeof outcomes measured)

4. Imprecision: Precision of estimates of effect5. Publication bias: Risk of bias due to selective

publication of results

Moderate-quality evidence “Evidence…obtained from RCTs with importantlimitations…. In addition, evidence from well-designed controlled trials without randomization,well-designed cohort or case-control analyticstudies, and multiple time series with or withoutintervention are in this category. Moderate-quality evidence also means that furtherresearch will probably have an important effecton our confidence in the estimate of effect andmay change the estimate.”

Low-quality evidence “Evidence obtained from observational studieswould typically be rated as low quality becauseof the risk for bias. Low-quality evidence meansthat further research is very likely to have animportant effect on our confidence in theestimate of effect and will probably change theestimate. However, the quality of evidence maybe rated as moderate or even high, dependingon circumstances under which evidence isobtained from observational studies.”

# # # # #Overall quality of evidence that supports a given criterion: high, moderate, low

Strength of EvidenceStrength of evidence ratings for each criterion are based on synthetic integration of the quality of evidence, the frequency and severityof potential adverse events and relationship to potential benefits, and clinical judgment.Strong Harms, adverse events, and risks clearly outweigh benefits.Weak Harms, adverse events, and risks may not outweigh benefits.

Abbreviations: ACP, American College of Physicians; GRADE, Grading of Recommendations Assessment, Development and Evaluation.aAdapted from: Qaseem A, Snow V, Owens DK, et al. The development of clinical practice guidelines and guidance statements of the American College ofPhysicians: summary of methods. Ann Intern Med. 2010;153:194-–199. Guyatt G, Oxman AD, Sultan S, et al. GRADE guidelines,: 11.: making an overallrating of confidence in effect estimates for a single outcome and for all outcomes. J Clin Epidemiol. 2013;66(2):151-–157. Andrews JC, Schünemann HJ,Oxman AD, et al. GRADE guidelines,: 15.: going from evidence to recommendation-determinants of a recommendation’s direction and strength. J Clin Epi-demiol. 2013;66(7):726–735.


Tab

le2.

2019

American

GeriatricsSo

cietyBeers

Criteria®

forPo

tentially

Inap

prop

riateM

edicationUse

inOlder

Adu

ltsa

OrganSystem

,TherapeuticCa

tegory,D

rug(s)

Ratio

nale

Recommendatio

nQu

ality

ofEvidence

Strength

ofRe

commendatio

n

Anticho

linergics

b

Firs

t-ge

neratio

nan

tihistamines

Bromph

enira

mine

Carbino

xamine

Chlorph

enira

mine

Clemas

tine

Cyp

rohe

ptad

ine

Dex

brom

phen

iramine

Dex

chlorphe

niramine

Dim

enhy

drinate

Diphe

nhyd

ramine(oral)

Dox

ylam

ine

Hyd

roxyzine

Mec

lizine

Prometha

zine

Pyrilamine

Trip

rolidine

Highlyan

ticho

linergic;

clea

ranc

eredu

cedwith

adva

nced

age,

andtoleranc

ede

velops

whe

nus

edas

hypn

otic;risk

ofco

nfus

ion,

drymou

th,c

onstipation,

andothe

ran

ticho

linergiceffectsor

toxicity

Use

ofdiph

enhy

dram

inein

situations

such

asac

ute

trea

tmen

tofs

evereallergic

reac

tionmay

beap

prop

riate.

Avo

idMod

erate

Stron

g

Antiparkins

onianag

ents

Ben

ztropine

(oral)

Trih

exyp

henidy

l

Not

reco

mmen

dedforprev

entio

nor

trea

tmen

tof

extrap

yram

idal

symptom

swith

antip

sych

otics;

more

effectiveag

ents

availablefortrea

tmen

tofP

arkins

ondise

ase

Avo

idMod

erate

Stron

g

Antispa

smod

ics

Atrop

ine(exclude

sop

htha

lmic)

Bellado

nnaalka

loids

Clidinium-chlordiaz

epox

ide

DicyclomineHom

atropine

(exclude

sop

thalmic)

Hyo

scya

mine

Meths

copo

lamine

Propa

nthe

line

Sco

polamine

Highlyan

ticho

linergic,

unce

rtaineffectiven

ess

Avo

idMod

erate

Stron

g

Antith

rombo

tics

Dipyridam

ole,

oral

shorta

cting(doe

sno

tapp

lyto

theex

tend

ed-relea

seco

mbina

tionwith

aspirin

)

May

caus

eorthos

tatic

hypo

tens

ion;

moreeffective

alternatives

available;

IVform

acce

ptab

leforus

ein

cardiacstress

testing

Avo

idMod

erate

Stron

g

Anti-infec

tive

Nitrofuran

toin

Poten

tialfor

pulm

onarytoxicity,h

epatox

icity,a

ndpe

riphe

raln

europa

thy,

espe

cially

with

long

-term

use;

saferalternatives

available

Avo

idin

individu

alswith

crea

tinine

clea

ranc

e<30

mL/min

orforlong

-term

supp

ression

Low

Stron

g

Cardiov

ascu

lar

Periphe

rala

lpha

-1bloc

kers

fortrea

tmen

tof

hype

rten

sion

Dox

azos

inPrazo

sin

Teraz

osin

Highris

kof

orthos

tatic

hypo

tens

ionan

das

sociated

harm

s,es

pecially

inolde

rad

ults;n

otreco

mmen

dedas

routinetrea

tmen

tfor

hype

rten

sion

;alte

rnativeag

ents

have

supe

riorris

k/be

nefitp

rofile

Avo

idus

eas

anan

tihyp

ertens

ive

Mod

erate

Stron

g

(Con

tinu

ed)


Tab

le2(C

ontd.)

OrganSystem

,TherapeuticCa

tegory,D

rug(s)

Ratio

nale

Recommendatio

nQu

ality

ofEvidence

Strength

ofRe

commendatio

n

Cen

tral

alph

a-ag

onists

Clonidine

forfirst-linetrea

tmen

tof

hype

rten

sion

Other

CNSalph

a-ag

onists

Gua

nabe

nzGua

nfac

ine

Methy

ldop

aRes

erpine

(>0.1mg/da

y)

Highris

kof

adve

rseCNSeffects;

may

caus

ebrad

ycardia

andorthos

tatic

hypo

tens

ion;

notrec

ommen

dedas

routine

trea

tmen

tfor

hype

rten

sion

Avo

idas

first-linean

tihyp

ertens

ive

Avo

idothe

rCNSalph

a-ag

onists

aslisted

Low

Low

Stron

g

Stron

g

Disop

yram

ide

May

indu

cehe

artfailure

inolde

rad

ults

beca

useof

potent

nega

tiveinotropicac

tion;

strong

lyan

ticho

linergic;

othe

ran

tiarrhy

thmic

drug

spreferred

Avo

idLo

wStron

g

Drone

darone

Worse

outcom

esha

vebe

enrepo

rted

inpa

tientstaking

dron

edaron

ewho

have

perm

anen

tatrialfi

brillationor

seve

reor

rece

ntly

deco

mpe

nsated

heartfailure.

Avo

idin

individu

alswith

perm

anen

tatria

lfibrillationor

seve

reor

rece

ntly

deco

mpe

nsated

heartfailure

High

Stron

g

Digox

inforfirst-linetrea

tmen

tofa

trial

fibrillationor

ofhe

artfailure

Use

inatria

lfibrillation:

shou

ldno

tbeus

edas

afirst-line

agen

tinatria

lfibrillation,

beca

usetherearesa

feran

dmoreeffectivealternatives

forrate

controls

uppo

rted

byhigh

-qua

lityev

iden

ce.

Use

inhe

artfailure:e

vide

nceforbe

nefits

andha

rmsof

digo

xinis

confl

ictin

gan

dof

lower

quality;m

ostb

utno

tall

oftheev

iden

ceco

ncerns

usein

HFrEF.T

here

isstrong

eviden

ceforothe

rag

ents

asfirst-linetherap

yto

redu

ceho

spita

lizations

andmortalityin

adults

with

HFrEF.In

heartfailure,h

ighe

rdo

sage

sareno

tassoc

iatedwith

additio

nalb

enefi

tand

may

increa

seris

kof

toxicity.

Dec

reas

edrena

lclearan

ceof

digo

xinmay

lead

toincrea

sedris

kof

toxiceffects;

furthe

rdo

seredu

ctionmay

bene

cessaryin

thos

ewith

stag

e4or

5ch

ronickidn

eydise

ase.

Avo

idthis

rate

controla

gent

asfirst-

linetherap

yforatria

lfibrillation

Avo

idas

first-linetherap

yforhe

art

failure

Ifus

edforatria

lfibrillationor

heart

failure,a

void

dosa

ges>0.12

5mg/da

y

Atrialfi

brillation:

low

Hea

rtfailure:

low

Dos

age

>0.12

5mg/da

y:mod

erate

Atrialfibrillation:

strong

Hea

rtfailure:

strong

Dos

age

>0.12

5mg/da

y:strong

Nife

dipine

,immed

iate

releas

ePoten

tialfor

hypo

tens

ion;

riskof

prec

ipita

tingmyo

cardial

isch

emia

Avo

idHigh

Stron

g

Amioda

rone

Effe

ctiveformaintaining

sinu

srhythm

buth

asgrea

ter

toxicitie

sthan

othe

ran

tiarrhy

thmicsus

edin

atria

lfibrillation;

may

bereas

onab

lefirst-linetherap

yin

patie

nts

with

conc

omita

nthe

artfailure

orsu

bstantialleft

ventric

ular

hype

rtroph

yifrhythm

controlispreferredov

errate

control

Avo

idas

first-linetherap

yforatria

lfibrillationun

less

patie

ntha

she

art

failure

orsu

bstantialleftv

entricular

hype

rtroph

y

High

Stron

g

Cen

tral

nervou

ssystem

Antidep

ressan

ts,a

lone

orin

combina

tion

Amitriptyline

Amox

apine

Clomipramine

Des

ipramine

Dox

epin

>6mg/da

yIm

ipramine

Highlyan

ticho

linergic,

seda

ting,

andca

useorthos

tatic

hypo

tens

ion;

safety

profi

leof

low-dos

edo

xepin

(≤6mg/da

y)co

mpa

rableto

that

ofplac

ebo

Avo

idHigh

Stron

g


Tab

le2(C

ontd.)

OrganSystem

,TherapeuticCa

tegory,D

rug(s)

Ratio

nale

Recommendatio

nQu

ality

ofEvidence

Strength

ofRe

commendatio

n

Nortriptyline

Parox

etine

Protriptyline

Trim

ipramine

Antipsych

otics,

first(co

nven

tiona

l)an

dse

cond

(atypica

l)ge

neratio

nIncrea

sedris

kof

cerebrov

ascu

larac

cide

nt(strok

e)an

dgrea

terrate

ofco

gnitive

declinean

dmortalityin

person

swith

demen

tiaAvo

idan

tipsych

oticsforb

ehav

ioralproblem

sof

demen

tiaor

delirium

unless

nonp

harm

acolog

icalop

tions

(eg,

beha

vioral

interven

tions

)hav

efailedor

areno

tpos

siblean

dtheolde

rad

ultisthreaten

ingsu

bstantialharm

tose

lfor

othe

rs

Avo

id,e

xcep

tinsc

hizo

phreniaor

bipo

lardiso

rder,o

rforsh

ort-term

use

asan

tiemetic

durin

gch

emothe

rapy

Mod

erate

Stron

g

Barbiturates

Amob

arbital

Butab

arbital

Butalbital

Mep

hoba

rbita

lPen

toba

rbita

lPhe

noba

rbita

lSec

obarbital

Highrate

ofph

ysical

depe

nden

ce,toleran

ceto

slee

pbe

nefits,g

reater

riskof

overdo

seat

low

dosa

ges

Avo

idHigh

Stron

g

Ben

zodiaz

epines

Sho

rtan

dinterm

ediate

actin

g:

Alprazo

lam

Estaz

olam

Loraze

pam

Oxa

zepa

mTem

azep

amTria

zolam

Long

actin

g:

Chlordiaz

epox

ide(alone

orin

combina

tion

with

amitriptylineor

clidinium)

Clona

zepa

mCloraze

pate

Diaze

pam

Fluraze

pam

Qua

zepa

m

Older

adults

have

increa

sedse

nsitivity

tobe

nzod

iaze

pine

san

dde

crea

sedmetab

olism

oflong

-ac

tingag

ents;inge

neral,allb

enzo

diaz

epines

increa

seris

kof

cogn

itive

impa

irmen

t,de

lirium,falls,fractures

,and

motor

vehiclecras

hesin

olde

rad

ults

May

beap

prop

riate

forse

izurediso

rders,

rapidey

emov

emen

tsleep

beha

vior

diso

rder,b

enzo

diaz

epine

with

draw

al,e

than

olwith

draw

al,s

everege

neralized

anxietydiso

rder,a

ndpe

riproce

durala

nesthe

sia

Avo

idMod

erate

Stron

g

Mep

roba

mate

Highrate

ofph

ysical

depe

nden

ce;s

edating

Avo

idMod

erate

Stron

gNon

benz

odiaze

pine

,ben

zodiaz

epine

rece

ptor

agon

isth

ypno

tics(ie

,“Z-drugs

”)Eszop

iclone

Zalep

lon

Zolpide

m

Non

benz

odiaze

pine

benz

odiaze

pine

rece

ptor

agon

ist

hypn

otics(ie

,Zdrug

s)ha

vead

verseev

ents

similarto

thos

eof

benz

odiaze

pine

sin

olde

rad

ults

(eg,

delirium,

falls,fractures

);increa

sedem

erge

ncyroom

visits/

hosp

italizations

;motor

vehiclecras

hes;

minim

alim

prov

emen

tinslee

platenc

yan

ddu

ratio

n

Avo

idMod

erate

Stron

g

Ergoloidmes

ylates

(deh

ydroge

natedergo

talkaloids

)Isox

suprine

Lack

ofeffica

cyAvo

idHigh

Stron

g

(Con

tinu

ed)


Tab

le2(C

ontd.)

OrganSystem

,TherapeuticCa

tegory,D

rug(s)

Ratio

nale

Recommendatio

nQu

ality

ofEvidence

Strength

ofRe

commendatio

n

End

ocrin

eAnd

roge

nsMethy

ltestos

terone

Tes

tosteron

e

Poten

tialfor

cardiacprob

lems;

contraindica

tedin

men

with

pros

tate

canc

erAvo

idun

less

indica

tedforco

nfirm

edhy

pogo

nadism

with

clinical

symptom

sMod

erate

Wea

k

Des

icca

tedthyroid

Con

cernsab

outc

ardiac

effects;

saferalternatives

available

Avo

idLo

wStron

g

Estroge

nswith

orwith

outp

roge

stins

Evide

nceof

carcinog

enic

potential(brea

stan

den

dometriu

m);lack

ofca

rdioprotec

tiveeffect

and

cogn

itive

protec

tionin

olde

rwom

enEvide

nceindica

testhat

vagina

lestroge

nsforthe

treatmen

tof

vagina

ldryne

ssaresa

fean

deffective;

wom

enwith

ahistoryof

brea

stca

ncer

who

dono

tres

pond

tono

nhormon

altherap

iesaread

vise

dto

discus

stheris

ksan

dbe

nefitsof

low-dos

eva

gina

lestroge

n(dos

ages

ofes

tradiol

8wee

ksun

less

forh

igh-riskpa

tients(eg,

oral

cortico

steroids

orch

ronicNSAID

use),

eros

ivees

opha

gitis,B

arrettes

opha

gitis,

patholog

icalhy

persec

retoryco

ndition

,or

demon

stratedne

edform

ainten

ance

treatmen

t(eg

,bec

ause

offailure

ofdrug

discon

tinua

tiontrialor

H2-rece

ptor

antago

nists)

High

Stron

g


Tab

le2(C

ontd.)

OrganSystem

,TherapeuticCa

tegory,D

rug(s)

Ratio

nale

Recommendatio

nQu

ality

ofEvidence

Strength

ofRe

commendatio

n

Painmed

ications

Mep

eridine

Orala

nalges

icno

teffe

ctivein

dosa

gesco

mmon

lyus

ed;

may

have

high

erris

kof

neurotox

icity,inc

luding

delirium,

than

othe

rop

ioids;

saferalternatives

available

Avo

idMod

erate

Stron

g

Non

–cycloo

xyge

nase-selectiveNSAIDs,oral:

Asp

irin>32

5mg/da

yDiclofena

cDiflun

isal

Etodo

lac

Fen

oprofen

Ibup

rofen

Ketop

rofen

Mec

lofena

mate

Mefen

amic

acid

Melox

icam

Nab

umeton

eNap

roxe

nOxa

proz

inPiro

xica

mSulinda

cTolmetin

Increa

sedris

kof

gastrointestinal

blee

ding

orpe

ptic

ulce

rdise

asein

high

-riskgrou

ps,inc

luding

thos

e>75

yearsor

taking

oral

orpa

renteral

corticos

teroids,

anticoa

gulants,

oran

tiplateleta

gents;

useof

proton

-pum

pinhibitoror

misop

rostol

redu

cesbu

tdoe

sno

telim

inateris

k.Upp

erga

strointestinal

ulce

rs,g

ross

blee

ding

,orpe

rforation

caus

edby

NSAIDsoc

curin

~1%

ofpa

tientstrea

tedfor

3-6mon

thsan

din

~2%

-4%

ofpa

tientstrea

tedfor1ye

ar;

thes

etren

dsco

ntinue

with

long

erdu

ratio

nof

use.

Also

canincrea

sebloo

dpres

sure

andindu

cekidn

eyinjury.

Risks

aredo

serelated.

Avo

idch

ronicus

e,un

less

othe

ralternatives

areno

teffe

ctivean

dpa

tient

cantake

gastroprotec

tive

agen

t(proton

-pum

pinhibitoror

misop

rostol)

Mod

erate

Stron

g

Indo

metha

cin

Ketorolac

,inc

lude

spa

renteral

Increase

drisk

ofgastrointestinalbleed

ing/pep

ticulcer

diseas

eand

acute

kidney

injury

inolder

adu

ltsIndo

metha

cinis

morelikelythan

othe

rNSAID

sto

have

adve

rseCNSeffects.

Ofallthe

NSAID

s,indo

metha

cin

hasthemos

tad

verseeffects.

Avo

idMod

erate

Stron

g

Ske

letalm

usclerelaxa

nts

Carisop

rodo

lChlorzo

xazo

neCyc

lobe

nzap

rine

Metax

alon

eMetho

carbam

olOrphe

nadrine

Mos

tmus

clerelaxa

ntspo

orly

toleratedby

olde

rad

ults

beca

useso

meha

vean

ticho

linergicad

verseeffects,

seda

tion,

increa

sedris

kof

frac

tures;

effectiven

essat

dosa

gestoleratedby

olde

rad

ults

ques

tiona

ble

Avo

idMod

erate

Stron

g

Gen

itourinary

Des

mop

ressin

Highris

kof

hypo

natrem

ia;s

afer

alternativetrea

tmen

tsAvo

idfortrea

tmen

tofn

octuria

orno

cturna

lpolyu

riaMod

erate

Stron

g

Abb

reviations:C

NS,

centraln

ervo

ussystem

;HFrEF,

heartfailu

rewithredu

cedejection

fraction

;NSA

ID,n

onsteroida

lanti-inflam

matorydrug

;SIA

DH,syn

drom

eof

inap

prop

riatean

tidiuretic

horm

onesecretion.

a The

prim

arytarget

audience

isthepracticing

clinician.

The

intentions

ofthecriteria

includ

e(1)im

prov

ingtheselectionof

prescription

drug

sby

clinicians

andpa

tients;(2)evalua

ting

patterns

ofdrug

usewithin

popu

lation

s;(3)educatingclinicians

andpa

tients

onprop

erdrug

usage;an

d(4)evalua

ting

health-outcome,

quality-of-care,

cost,a

ndutilization

data.

bSeealso

criterionon

high

lyan

ticholinergican

tidepressants.


Tab

le3.

2019

American

GeriatricsSo

cietyBeers

Criteria®

forPo

tentially

Inap

prop

riateM

edicationUse

inOlder

Adu

ltsDue

toDrug-Disease

orDrug-Sy

ndrome

Interactions

Tha

tM

ayExa

cerbatetheDisease

orSy

ndromea

Diseaseor

Syndrome

Drug(s)

Ratio

nale

Recommendatio

nQu

ality

ofEvidence

Strength

ofRe

commendatio

n

Cardiov

ascu

lar

Hea

rtfailure

Avo

id:C

ilostaz

ol

Avo

idin

heartfailure

with

redu

ced

ejec

tionfrac

tion:

Non

dihy

drop

yridineCCBs(diltiaze

m,

verapa

mil)

Use

with

cautionin

patie

ntswith

heart

failure

who

areas

ymptom

atic;a

void

inpa

tientswith

symptom

atic

heartfailure:

NSAIDsan

dCOX-2

inhibitors

Thiaz

olidined

ione

s(pioglita

zone

,rosiglita

zone

)

Drone

darone

Poten

tialtoprom

otefluidretention

and/or

exac

erba

tehe

artfailure

(NSAIDs

andCOX-2

inhibitors,n

ondihy

drop

yridine

CCBs,thiazo

lidined

ione

s);p

oten

tialto

increa

semortalityinolde

radu

ltswith

heart

failure

(cilostaz

olan

ddron

edaron

e)

Asno

ted,

avoid

orus

ewith

caution

Cilostaz

ol:low

Non

dihy

drop

yridine

CCBs:

mod

erate

NSAIDs:

mod

erate

COX-2

inhibitors:low

Thiaz

olidined

ione

s:high

Drone

darone

:high

Cilostaz

ol:s

tron

g

Non

dihy

drop

yridine

CCBs:

strong

NSAIDs:

strong

COX-2

inhibitors:stro

ng

Thiaz

olidined

ione

s:strong

Drone

darone

:stron

g

Syn

cope

AChE

Is

Non

selectivepe

riphe

rala

lpha

-1bloc

kers

(ie,d

oxaz

osin,p

razo

sin,

terazo

sin)

Tertia

ryTCAs

Antipsych

otics:

Chlorprom

azine

Thioridaz

ine

Olanz

apine

AChE

Isca

usebrad

ycardiaan

dsh

ould

beav

oide

din

olde

rad

ults

who

sesync

ope

may

bedu

eto

brad

ycardia.

Non

selective

perip

herala

lpha

-1bloc

kers

caus

eorthos

tatic

bloo

dpres

sure

chan

gesan

dsh

ould

beav

oide

din

olde

rad

ults

who

sesync

opemay

bedu

eto

orthos

tatic

hypo

tens

ion.

Tertiary

TCAsan

dthe

antipsych

oticslistedincrea

setheriskof

orthos

tatic

hypo

tens

ionor

brad

ycardia.

Avo

idAChE

Is,T

CAs,

and

antip

sych

otics:

high

Non

selectivepe

riphe

ral

alph

a-1bloc

kers:h

igh

AChE

Isan

dTCAs:

strong

Non

selective

perip

herala

lpha

-1bloc

kers

and

antip

sych

otics:

wea

k

Cen

tral

nervou

ssystem

Delirium

Anticho

linergics

(see

Tab

le7an

dfull

crite

riaav

ailableon

www.

geria

tricscareo

nline.org.)

Antipsych

oticsb

Ben

zodiaz

epines

Corticos

teroids(orala

ndpa

renteral)c

H2-rece

ptor

antago

nists

Cim

etidine

Fam

otidine

Nizatidine

Ran

itidine

Mep

eridine

Non

benz

odiaze

pine

,ben

zodiaz

epine

rece

ptor

agon

isth

ypno

tics:

eszo

piclon

e,za

leplon

,zolpide

m

Avo

idin

olde

rad

ults

with

orat

high

risk

ofde

lirium

beca

useof

potentialo

findu

cing

orworse

ning

delirium

Avo

idan

tipsych

oticsforbe

havioral

prob

lemsof

demen

tiaan

d/or

delirium

unless

nonp

harm

acolog

ical

optio

ns(eg,

beha

vioral

interven

tions

)ha

vefailedor

areno

tpos

siblean

dtheolde

rad

ultis

threaten

ingsu

bstantialh

arm

tose

lfor

othe

rs.A

ntipsych

oticsareas

sociated

with

grea

terris

kof

cerebrov

ascu

lar

accide

nt(strok

e)an

dmortalityin

person

swith

demen

tia.

Avo

idH2-rece

ptor

antago

nists:

low

Allothe

rs:m

oderate

Stron

g

Dem

entia

orco

gnitive

impa

irmen

tAnticho

linergics

(see

Tab

le7an

dfull

crite

riaav

ailableon

www.

geria

tricscareo

nline.org)

Ben

zodiaz

epines

Non

benz

odiaze

pine

,ben

zodiaz

epine

rece

ptor

agon

isth

ypno

tics

Eszop

iclone

Avo

idbe

caus

eof

adve

rseCNSeffects

Avo

idan

tipsych

oticsforbe

havioral

prob

lemsof

demen

tiaan

d/or

delirium

unless

nonp

harm

acolog

ical

optio

ns(eg,

beha

vioral

interven

tions

)ha

vefailedor

areno

tpos

siblean

dtheolde

rad

ultis

threaten

ingsu

bstantialh

arm

tose

lfor

Avo

idMod

erate

Stron

g


http://www.geriatricscareonline.orghttp://www.geriatricscareonline.orghttp://www.geriatricscareonline.orghttp://www.geriatricscareonline.org

Tab

le3(C

ontd.)

Diseaseor

Syndrome

Drug(s)

Ratio

nale

Recommendatio

nQu

ality

ofEvidence

Strength

ofRe

commendatio

n

Zalep

lon

Zolpide

mAntipsych

otics,

chronican

das

-nee

ded

useb

othe

rs.A

ntipsych

oticsareas

sociated

with

grea

terris

kof

cerebrov

ascu

lar

accide

nt(strok

e)an

dmortalityin

person

swith

demen

tia.

History

offalls

orfrac

tures

Antiepileptics

Antipsych

oticsb

Ben

zodiaz

epines

Non

benz

odiaze

pine

,ben

zodiaz

epine

rece

ptor

agon

isth

ypno

tics

Eszop

iclone

Zalep

lon

Zolpide

m

Antidep

ressan

tsTCAs

SSRIs

SNRIs

Opioids

May

caus

eatax

ia,impa

iredps

ycho

motor

func

tion,

sync

ope,

additiona

lfalls;sho

rter-

actingbe

nzod

iaze

pine

sareno

tsafer

than

long

-actingon

es.

Ifon

eofthedrug

smustbeused

,con

side

rredu

cing

useofothe

rCNS-active

med

ications

thatincrea

seriskoffalls

and

fractures

(ie,antiepileptics,op

ioid-re

ceptor

agon

ists,antipsychotics,an

tidep

ressan

ts,

nonb

enzodiazep

inean

dbe

nzod

iazepine

receptor

agon

isthypno

tics,othe

rseda

tives/hypno

tics)an

dimplem

entother

strategies

toredu

cefallrisk.Datafor

antidep

ressan

tsaremixed

butnocompe

lling

eviden

cethatcerta

inan

tidep

ressan

tsconfer

less

fallriskthan

othe

rs.

Avo

idun

less

safer

alternatives

areno

tav

ailable;

avoid

antie

pilepticsex

cept

for

seizurean

dmoo

ddiso

rders

Opioids

:avo

idex

cept

for

pain

man

agem

entinthe

setting

ofse

vere

acute

pain

(eg,

rece

ntfrac

tures

orjointrep

lace

men

t)

Opioids

:mod

erate

Allothe

rs:h

igh

Stron

g

Parkins

ondise

ase

Antiemetics

Metoc

lopram

ide

Proch

lorperaz

ine

Prometha

zine

Allan

tipsych

otics(excep

tque

tiapine

,cloz

apine,

pimav

anse

rin)

Dopam

ine-receptor

antagonistswith

potentialtoworsenparkinsonian

symptom

s

Excep

tions

:Pim

avan

serin

andcloz

apineap

pear

tobe

less

likelyto

prec

ipita

teworse

ning

ofParkins

ondise

ase.

Que

tiapine

has

only

been

stud

iedin

low-qua

lityclinical

trials

with

effica

cyco

mpa

rableto

that

ofplac

eboin

five

trials

andto

that

ofcloz

apinein

twoothe

rs.

Avo

idMod

erate

Stron

g

Gas

trointes

tinal

History

ofga

stric

ordu

oden

alulce

rsAsp

irin>32

5mg/da

yNon

–COX-2–se

lectiveNSAIDs

May

exac

erba

teex

istin

gulce

rsor

caus

ene

w/add

ition

alulce

rsAvo

idun

less

othe

ralternatives

areno

teffectivean

dpa

tient

can

take

gastroprotec

tive

agen

t(ie,p

roton-pu

mp

inhibitoror

misop

rostol)

Mod

erate

Stron

g

Kidne

y/urinarytrac

tChron

ickidn

eydise

asestag

e4or

high

er(creatinine

clea

ranc

e<30

mL/min)

NSAIDs(non

-COXan

dCOXse

lective,

oral

andpa

renteral,n

onac

etylated

salicylates

)

May

increa

seris

kof

acutekidn

eyinjury

andfurthe

rde

clineof

rena

lfun

ction

Avo

idMod

erate

Stron

g

(Con

tinu

ed)


was lowered to 70 years or older from 80 years or older. Thiscriterion was also expanded to cover use of aspirin as primaryprevention of colorectal cancer. Note that this criterion doesnot apply to use of aspirin for secondary prevention of eitherdisease.

• In addition to the existing caution about dabigatran, theupdated criteria highlight caution about use of rivaroxaban fortreatment of venous thromboembolism or atrial fibrillation inadults 75 years or older.

• Tramadol was added to the list of drugs associated withhyponatremia or syndrome of inappropriate antidiuretic hor-mone secretion. The chemotherapeutic agents carboplatin, cyclo-phosphamide, cisplatin, and vincristine were removed from thislist because the panel thought the prescribing of these highly spe-cialized drugs fell outside the scope of the criteria.

• Vasodilators were removed, because syncope is not unique toolder adults.

• The combination dextromethorphan/quinidine was added tothe “use with caution” table on the basis of limited efficacy inpatients with behavioral symptoms of dementia without pseu-dobulbar affect while potentially increasing the risk of falls anddrug-drug interactions.

• The combination trimethoprim-sulfamethoxazole (TMP-SMX)should be used with caution by patients with reduced kidneyfunction and taking an angiotensin-converting enzyme inhibi-tor (ACEI) or angiotensin receptor blocker (ARB) because ofan increased risk of hyperkalemia.

Drug-Drug Interactions

Table 5 contains potentially clinically important drug-druginteractions to be avoided in older adults. New recommen-dations include avoiding use of opioids concurrently withbenzodiazepines and avoiding use of opioids concurrentlywith gabapentinoids (except when transitioning from theformer to the latter). Other additions to the table are inter-actions involving TMP-SMX, macrolide antibiotics, andciprofloxacin. TMP-SMX in combination with phenytoinor warfarin increases the risk of phenytoin toxicity andbleeding, respectively. Macrolides, excluding azithromycin,or ciprofloxacin in combination with warfarin increasesbleeding risk. Ciprofloxacin in combination with theophyl-line increases risk of theophylline toxicity. The concurrentuse of a combination of three or more central nervous sys-tem (CNS) agents (antidepressants, antipsychotics, benzodi-azepines, nonbenzodiazepine benzodiazepine receptoragonist hypnotics, antiepileptics, and opioids) and increasedfall risk have been collapsed into one recommendationinstead of separate recommendations for each drug class.The recommendation on avoiding concurrent use of medi-cations that increase serum potassium has been expandedto encompass a broader range of these medications.

PIMs Based on Kidney Function

Table 6 contains a list of medications that should beavoided or have their dosage reduced based on kidney func-tion. Two antibiotics have been added, ciprofloxacin andTMP-SMX, over concerns of increased CNS effects and ten-don rupture, and worsening renal function and hyperkale-mia, respectively. Dofetilide was also added because ofconcerns of corrected QT interval prolongation and torsadede pointes. The creatinine clearance lower limit at which toavoid edoxaban has been reduced to less than 15 mL/min.T

able

3(C

ontd.)

Diseaseor

Syndrome

Drug(s)

Ratio

nale

Recommendatio

nQu

ality

ofEvidence

Strength

ofRe

commendatio

n

Urin

aryinco

ntinen

ce(alltype

s)in

wom

enEstroge

noral

andtran

sdermal

(exclude

sintrav

aginal

estrog

en)

Periphe

rala

lpha

-1bloc

kers

Dox

azos

inPrazo

sin

Teraz

osin

Lack

ofeffica

cy(orale

stroge

n)an

dag

grav

ationof

inco

ntinen

ce(alpha

-1bloc

kers)

Avo

idin

wom

enEstroge

n:high

Periphe

rala

lpha

-1bloc

kers:m

oderate

Estroge

n:strong

Periphe

rala

lpha

-1bloc

kers:s

tron

g

Lower

urinarytrac

tsymptom

s,be

nign

pros

tatic

hype

rplasia

Stron

glyan

ticho

linergicdrug

s,ex

cept

antim

usca

rinicsforurinaryinco

ntinen

ce(see

Tab

le7an

dfullcrite

riaav

ailableon

www.geriatricscareo

nline.org)

May

decrea

seurinaryflow

andca

use

urinaryretention

Avo

idin

men

Mod

erate

Stron

g

Abb

reviations:AChE

I,acetylcholinesterase

inhibitor;

CCB,calcium

chan

nelblocker;

CNS,

centralnervou

ssystem

;COX,cycloo

xygena

se;NSA

ID,no

nsteroidal

anti-infl

ammatory

drug

;SN

RI,

serotonin-

norepineph

rine

reup

take

inhibitor;SSRI,selectiveserotoninreup

take

inhibitor;TCA,tricyclic

antidepressant.

a The

prim

arytarget

audience

isthepracticing

clinician.

The

intentions

ofthecriteria

includ

e(1)im

prov

ingtheselectionof

prescription

drug

sby

clinicians

andpa

tients;(2)evalua

ting

patterns

ofdrug

usewithin

popu

lation

s;(3)educatingclinicians

andpa

tients

onprop

erdrug

usage;an

d(4)evalua

ting

health-outcome,

quality-of-care,

cost,a

ndutilization

data.

bMay

berequ

ired

totreatconcurrent

schizoph

renia,

bipo

lardisorder,a

ndotherselected

mentalh

ealthcond

itions

butshou

ldbe

prescribed

inthelowesteffectivedo

sean

dshortest

possible

duration

.c Excludesinha

ledan

dtopicalforms.Orala

ndpa

renteral

corticosteroidsmay

berequ

ired

forcond

itions

such

asexacerba

tion

ofchronicob

structivepu

lmon

arydiseasebu

tshou

ldbe

prescribed

inthelowesteffective

dose

andfortheshortest

possible

duration

.


http://www.geriatricscareonline.org

DISCUSSION

The 2019 AGS Beers Criteria® update contributes to thecritically important evidence base and discussion of medica-tions to avoid in older adults and the need to improve medi-cation use in older adults. The 2019 AGS Beers Criteria®

include 30 individual criteria of medications or medicationclasses to be avoided in older adults (Table 2) and 16 cri-teria specific to more than 40 medications or medicationclasses that should be used with caution or avoided in cer-tain diseases or conditions (Tables 3 and 4). As in past

updates, there were several changes to the 2019 AGS BeersCriteria®, including criteria that were modified or dropped,a few new criteria, and some changes in the level of evi-dence grading and clarifications in language and rationale(Tables 8–10).

The 2019 AGS Beers Criteria® is the third such updateby the AGS and the fifth update of the AGS Beers Criteria®

since their original release.1,2,10–12 The criteria was firstpublished almost 30 years ago in 1991, making them thelongest running criteria for PIMs in older adults.

Table 4. 2019 American Geriatrics Society Beers Criteria® for Potentially Inappropriate Medications: Drugs To BeUsed With Caution in Older Adultsa

Drug(s) Rationale RecommendationQuality ofEvidence

Strength ofRecommendation

Aspirin for primary preventionof cardiovascular diseaseand colorectal cancer

Risk of major bleeding from aspirinincreases markedly in older age. Severalstudies suggest lack of net benefit whenused for primary prevention in older adultwith cardiovascular risk factors, but evidenceis not conclusive. Aspirin is generallyindicated for secondary prevention in olderadults with established cardiovasculardisease.

Use with caution inadults ≥70 years

Moderate Strong

DabigatranRivaroxaban

Increased risk of gastrointestinal bleedingcompared with warfarin and reported rateswith other direct oral anticoagulants whenused for long-term treatment of VTE or atrialfibrillation in adults ≥75 years.

Use with cautionfor treatment ofVTE or atrialfibrillation in adults≥75 years

Moderate Strong

Prasugrel Increased risk of bleeding in older adults;benefit in highest-risk older adults (eg, thosewith prior myocardial infarction or diabetesmellitus) may offset risk when used for itsapproved indication of acute coronarysyndrome to be managed with percutaneouscoronary intervention.

Use with caution inadults ≥75 years

Moderate Weak

AntipsychoticsCarbamazepineDiureticsMirtazapineOxcarbazepineSNRIsSSRIsTCAsTramadol

May exacerbate or cause SIADH orhyponatremia; monitor sodium level closelywhen starting or changing dosages in olderadults

Use with caution Moderate Strong

Dextromethorphan/quinidine

Limited efficacy in patients with behavioralsymptoms of dementia (does not apply totreatment of PBA). May increase risk of fallsand concerns with clinically significant druginteractions. Does not apply to treatment ofpseudobulbar affect.

Use with caution Moderate Strong

Trimethoprim-sulfamethoxazole

Increased risk of hyperkalemia when usedconcurrently with an ACEI or ARB inpresence of decreased creatinine clearance

Use with caution inpatients on ACEIor ARB anddecreasedcreatinineclearance

Low Strong

Abbreviations: ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; PBA, pseudobulbar affect; SIADH, syndrome of inappro-priate antidiuretic hormone secretion; SNRI, serotonin-norepinephrine reuptake inhibitor; SSRI, selective serotonin reuptake inhibitor; TCA, tricyclic antide-pressant; VTE, venous thromboembolism.aThe primary target audience is the practicing clinician. The intentions of the criteria include (1) improving the selection of prescription drugs by cliniciansand patients; (2) evaluating patterns of drug use within populations; (3) educating clinicians and patients on proper drug usage; and (4) evaluating health-outcome, quality-of-care, cost, and utilization data.


Tab

le5.

2019

American

GeriatricsSo

cietyBeers

Criteria®

forPo

tentially

Clin

ically

Impo

rtan

tDrug-DrugInteractions

Tha

tSh

ould

BeAvo

ided

inOlder

Adu

lts

ObjectDrug

andClass

InteractingDrug

andClass

Risk

Ratio

nale

Recommendatio

nQu

ality

ofEvidence

Strength

ofRe

commendatio

n

RASinhibitor(ACEIs,A

RBs,

aliskiren)

orpo

tassium-spa

ring

diuretics(amilorid

e,triamterene

)

Ano

ther

RASinhibitor

(ACEIs,A

RBs,

aliskiren)

Increa

sedris

kof

hype

rkalem

iaAvo

idroutineus

ein

thos

ewith

chronickidn

eydise

asestag

e3a

orhigh

er

Mod

erate

Stron

g

Opioids

Ben

zodiaz

epines

Increa

sedris

kof

overdo

seAvo

idMod

erate

Stron

gOpioids

Gab

apen

tin,p

rega

balin

Increa

sedris

kof

seve

rese

datio

n-relatedad

verse

even

ts,inc

luding

resp

iratory

depres

sion

andde

ath

Avo

id;e

xcep

tions

arewhe

ntran

sitio

ning

from

opioid

therap

yto

gaba

pentin

orpreg

abalin,o

rwhe

nus

ingga

bape

ntinoids

toredu

ceop

ioid

dose

,alth

ough

cautionsh

ould

beus

edin

all

circum

stan

ces.

Mod

erate

Stron

g

Anticho

linergic

Anticho

linergic

Increa

sedris

kof

cogn

itive

decline

Avo

id;m

inim

izenu

mbe

rof

anticho

linergicdrug

s(Tab

le7)

Mod

erate

Stron

g

Antidepressants(TCAs,SSRIs,and

SNRIs)

Antipsy

chotics

Antiepileptics

Ben

zodiaz

epines

and

nonb

enzo

diaz

epine,

benz

odiaze

pine

rece

ptor

agon

isth

ypno

tics

(ie,“Z-drugs

”)Opioids

Any

combina

tionof

three

ormoreof

thes

eCNS-activedrug

sa

Increa

sedris

kof

falls

(all)

and

offrac

ture

(ben

zodiaz

epines

andno

nben

zodiaz

epine,

benz

odiaze

pine

rece

ptor

agon

isth

ypno

tics)

Avo

idtotalo

fthree

ormore

CNS-activedrug

sa;m

inim

ize

numbe

rof

CNS-activedrug

s

Com

bina

tions

includ

ing

benzod

iazepine

san

dno

nben

zodiazep

ine,

benzod

iazepine

receptor

agon

isthypno

ticsor

opioids:high

Allothe

rco

mbina

tions

:mod

erate

Stron

g

Corticos

teroids,

oral

orpa

renteral

NSAIDs

Increa

sedris

kof

peptic

ulce

rdise

aseor

gastrointestinal

blee

ding

Avo

id;ifn

otpo

ssible,p

rovide

gastrointestinal

protec

tion

Mod

erate

Stron

g

Lithium

ACEIs

Increa

sedris

kof

lithium

toxicity

Avo

id;m

onito

rlithium

conc

entrations

Mod

erate

Stron

g

Lithium

Loop

diuretics

Increa

sedris

kof

lithium

toxicity

Avo

id;m

onito

rlithium

conc

entrations

Mod

erate

Stron

g

Periphe

ralα

-1bloc

kers

Loop

diuretics

Increa

sedris

kof

urinary

inco

ntinen

cein

olde

rwom

enAvo

idin

olde

rwom

en,u

nles

sco

ndition

swarrant

both

drug

sMod

erate

Stron

g

Phe

nytoin

Trim

etho

prim

-sulfametho

xazo

leIncrea

sedris

kof

phen

ytoin

toxicity

Avo

idMod

erate

Stron

g

The

ophy

lline

Cim

etidine

Increa

sedris

kof

theo

phylline

toxicity

Avo

idMod

erate

Stron

g

The

ophy

lline

Ciproflox

acin

Increa

sedris

kof

theo

phylline

toxicity

Avo

idMod

erate

Stron

g

Warfarin

Amioda

rone

Increa

sedris

kof

blee

ding

Avo

idwhe

npo

ssible;ifu

sed

toge

ther,m

onito

rINR

clos

ely

Mod

erate

Stron

g

Warfarin

Ciproflox

acin

Increa

sedris

kof

blee

ding

Avo

idwhe

npo

ssible;ifu

sed

toge

ther,m

onito

rINR

clos

ely

Mod

erate

Stron

g

Warfarin

Mac

rolides

(excluding

azith

romycin)

Increa

sedris

kof

blee

ding

Avo

idwhe

npo

ssible;ifu

sed

toge

ther,m

onito

rINR

clos

ely

Mod

erate

Stron

g


The 2019 update has a similar number of changes tothe 2015 update but fewer changes than the 2012 update.This is likely because, with the support of the AGS and theexpert panel, the criteria have been regularly updated aboutevery 3 years since 2012. In 2019, 25 medications or medi-cation classes to be avoided outright or in a disease condi-tion were dropped from the AGS Beers Criteria® (Table 8).A few were also moved to a new table category or modified(Table 10). For medications to be removed from the AGSBeers Criteria®, the panel had to have new evidence or astrong rationale, for reasons such as the literature showed achange in evidence that cast new doubt on their “avoid”status. Finally, some drugs or drug-disease combinationswere omitted because they are not disproportionately rele-vant to the older adult population; this included the criteriaon drugs to avoid in adults with chronic seizures or epilepsyand in adults with insomnia.

Four new medications or medication classes wereadded to the list of drugs to be used with caution (Table 4;additions are also summarized in Table 9). Dextromethor-phan/quinidine was added because of its limited efficacy,concerns for clinically significant drug interactions, andpotentially increased risk of falls in older adults. TMP-SMXwas placed in the “use with caution table” because ofincreased risk of hyperkalemia when used concurrently withan ACEI or ARB in the presence of decreased creatinineclearance.13,14 Rivaroxaban was also added to the use withcaution table for adults 75 years or older. Other importantchanges in the use with caution table included lowering theage threshold in the aspirin for primary prevention recom-mendation from 80 years or younger to 70 years or youn-ger on the basis of emerging evidence of a major increase inthe risk of bleeding at a lower age.15 The Aspirin in Reduc-ing Events in the Elderly (ASPREE) trial, which was pub-lished outside the window of our literature search, foundthat low-dose aspirin used for primary prevention in olderadults did not confer a reduction in mortality, disability-freesurvival, or cardiovascular events.16,17 In a few instances,the level of evidence was revised based on new literature andthe improved modified grading method. For instance,H2-receptor antagonists were removed from the list of drugsto avoid in dementia, and the evidence level for H2-receptorantagonists was decreased to low (from moderate in 2015)for drugs to avoid in delirium.18 Again in 2019, the panelclarified the language for sliding-scale insulin because thiscontinued to be an area of confusion for clinicians.

Importantly, several drugs were added to the drug-disease and drug-drug interactions tables (Tables 3 and 5).Notably, SNRIs were added to the list of antidepressantdrug classes to avoid in persons with a history of falls orfractures.19,20 For this criterion, the level of evidence foropioids was changed to “moderate”; all other drugs remainat high. Two new drug-drug interactions involving opioidswere added, reflecting evidence of substantial harms thatcan occur when opioids are used concurrently with benzodi-azepines or gabapentinoids. Though these drug interactionsinvolving opioids are problematic in all persons, they aregrowing increasingly common and may lead to greater harmin vulnerable older adults. These concerns need to be balancedwith the need to treat chronic pain. A recent review of deathsfrom opioids concluded that the burden of opioid overdose inolder adults requires special attention, noting the largestT

able

5(C

ontd.)

ObjectDrug

andClass

InteractingDrug

andClass

Risk

Ratio

nale

Recommendatio

nQu

ality

ofEvidence

Strength

ofRe

commendatio

n

Warfarin

Trim

etho

prim

-sulfametho

xazo

leIncrea

sedris

kof

blee

ding

Avo

idwhe

npo

ssible;ifu

sed

toge

ther,m

onito

rINR

clos

ely

Mod

erate

Stron

g

Warfarin

NSAIDs

Increa

sedris

kof

blee

ding

Avo

idwhe

npo

ssible;ifu

sed

toge

ther,m

onito

rclos

elyfor

blee

ding

High

Stron

g

Abb

reviations:ACEI,an

giotensin-conv

erting

enzymeinhibitor;

ARB,an

giotensinreceptor

blocker;

CNS,

centralnervou

ssystem

;IN

R,internationa

lno

rmalized

ratio;

NSA

ID,no

nsteroidal

anti-infl

ammatorydrug

;RAS,

renin-an

giotensinsystem

;SNRI,serotonin-

norepineph

rine

reup

take

inhibitor;SSRI,selectiveserotoninreup

take

inhibitor;TCA,tricyclic

antidepressant.

a CNS-active

drug

s:an

tiepileptics;an

tipsycho

tics;b

enzodiazepines;n

onbenzod

iazepine,b

enzodiazepinereceptor

agon

isthy

pnotics;TCAs;SSRIs;S

NRIs;a

ndop

ioids.


Table 6. 2019 American Geriatrics Society Beers Criteria® for Medications That Should Be Avoided or Have TheirDosage Reduced With Varying Levels of Kidney Function in Older Adults

Medication Classand Medication

Creatinine Clearanceat Which ActionRequired, mL/min Rationale Recommendation

Quality ofEvidence


Anti-infectiveCiprofloxacin

relative increase in opioids occurred in persons 55 to64 (754% increase from 0.2% to 1.7%) and 65 years andolder and the absolute number of deaths in this group ismoderate.21,22

Several drug-drug interactions involving antimicrobialagents were also added to Table 5, and the recommendationto avoid concurrent use of three or more CNS-active

medications was reformatted to clarify and bring furtherattention to the increased risk of falls and other harms thatcan occur when multiple CNS-active medications arecombined.23

PIM use continues to be a serious problem in olderadults and especially in vulnerable older adults with multi-ple chronic conditions. Thus, the AGS Beers Criteria® con-tinue to be useful and necessary as a clinical tool, as aneducational tool at the bedside, and as a public health toolto improve medication safety in older adults. The AGSBeers Criteria® can increase awareness of polypharmacyand aid decision making when choosing drugs to avoid inolder adults. In a 2017 study using medical expendituredata (n = 16,588) in adults 65 years and older, poor healthstatus was associated with increased PIM use. In anotherstudy, the use of PIMs, as measured by the 2015 criteria, inpersons with dementia was 11% higher after diagnosis thanin the year of diagnosis.24,25 Benzodiazepine use remainscommon in older adults, especially in older women, despitethe fact that older adults are highly vulnerable to harmsassociated with use of these drugs.26 The challenge ofdecreasing PIM use and improving the overall quality ofmedication prescribing in older adults remains, and theAGS Beers Criteria® are one part of the solution.

The AGS Beers Criteria® are an essential evidence-based tool that should be used as a guide for drugs to avoidin older adults. However, they are not meant to supplantclinical judgment or an individual patient’s preferences,values, care goals, and needs, nor should they be used puni-tively or to excessively restrict access to medications. Thesecriteria were developed to be used in conjunction with aperson-centered team approach (physicians, nurses, phar-macists, other clinicians, the older adult, family, and others)to prescribing and monitoring adverse effects.27 A compan-ion article published to the 2015 updated AGS BeersCriteria®, entitled “How to Use the Beers Criteria: A Guidefor Patients, Clinicians, Health Systems, and Payors,”remains an important guide for using the AGS BeersCriteria®. It reminds clinicians that medications listed in theCriteria are potentially inappropriate, rather than definitelyinappropriate for all older adults, and encourages users toread the rationale and recommendation statements for eachmedication to avoid because these statements provideimportant guidance.3 Moreover, the criteria should not beinterpreted as giving license to steer patients away fromPIMs to even worse choices. For example, the recommenda-tion to avoid chronic, regular use of NSAIDs should not be

Table 6 (Contd.)

Medication Classand Medication

Creatinine Clearanceat Which ActionRequired, mL/min Rationale Recommendation

Quality ofEvidence


HyperuricemiaColchicine

interpreted as an invitation to prescribe opioids in theirplace. For further reference, a 2012 article provides a caseexample on how nurses can use the criteria to improvemedication use in older adults.28

As in previous years, the panel recognizes the need tooffer older adults and their clinicians pharmacological andnonpharmacological alternatives to medications included inthe AGS Beers Criteria®. Alternatives to some of the mostcommonly implicated medications listed in the 2015 update

were published in a companion article that accompaniedthat update. Readers are encouraged to review these sugges-tions, although we acknowledge that further work needs tobe done to keep pace with updates to the criteria and thechanging landscape of drug and nondrug therapies. We alsoencourage readers to research the safety and effectiveness ofpotential alternatives to drugs included in this document.Deprescribing is a concept to eliminate unsafe or unneces-sary drugs from a patient’s regimen. One source for online

Table 8. Medications/Criteria Removed Since 2015American Geriatrics Society Beers Criteria®

Medication/Criterion Reason for Removal

Independent of Diagnosis or Condition (Table 2)Ticlopidine No longer on US market; low

usePentazocine Oral no longer on US marketConsidering Disease and Syndrome Interactions (Table 3)Chronic seizures or epilepsy

BupropionChlorpromazineClozapineMaprotilineOlanzapineThioridazineThiothixeneTramadol

Not unique to older adults

DementiaH2-receptor antagonists

Weak evidence and to avoidoverly restricting therapeuticoptions for older adults withdementia who havegastroesophageal reflux orsimilar issues (given acoexisting criterion advisingagainst chronic use of PPIsexcept in specificcircumstances)

InsomniaOral decongestants

PhenylephrinePseudoephedrine

StimulantsAmphetamineArmodafinilMethylphenidateModafinil

TheobrominesTheophyllineCaffeine


Parkinson diseaseAripiprazole

Removed as a preferredantipsychotic in older adultswith Parkinson diseasebecause of safety and efficacyconcerns

Use With Caution (Table 4)SIADH/hyponatremia

CarboplatinCyclophosphamideCisplatinVincristine

Highly specialized drugs thatfell outside the scope of thecriteria

SyncopeVasodilators


Abbreviations: PPI, proton-pump inhibitor; SIADH, syndrome of inappro-priate antidiuretic hormone secretion.

Table 9. Medications/Criteria Added Since 2015 Ameri-can Geriatrics Society Beers Criteria®

Medication/Criterion Reason for Addition

Independent of Diagnosis or Condition (Table 2)Glimepiride Severe, prolonged

hypoglycemia in older adultsMethscopolaminePyrilamine

Strong anticholinergic

Considering Disease and Syndrome Interactions (Table 3)History of falls or fracturesSNRI

Associated with increased riskin older adults

Parkinson diseasePimavanserin

Unlike most otherantipsychotics, the revisedcriteria consider pimavanserinacceptable for treatment ofpsychosis in Parkinson disease

Use With Caution (Table 4)Rivaroxaban Emerging evidence of

increased risk of seriousbleeding compared with otheranticoagulant options

Tramadol Risk of SIADH/hyponatremiaDextromethorphan/quinidine Limited efficacy in treating

patients with dementiasymptoms disorder in absenceof pseudobulbar affect whilepotentially increasing risk offalls and drug-drug interactions

TMP-SMX Increased risk of hyperkalemiain combination with ACEIs andARBs in patients with reducedkidney function

Clinically Important Drug-Drug Interactions (Table 5)Opioids + benzodiazepines Increased risk of overdoseOpioids +gabapentin/pregabalin

Increased risk of overdose

Phenytoin + TMP-SMX Increased risk of phenytointoxicity

Theophylline + ciprofloxacin Increased risk of theophyllinetoxicity

Warfarin + ciprofloxacin Increased risk of bleedingWarfarin + macrolides(excluding azithromycin)

Increased risk of bleeding

Warfarin + TMP-SMX Increased risk of bleedingMedications That Should Be Avoided or Have Their DosageReduced With Decreased Kidney Function (Table 6)Ciprofloxacin Increased risk of CNS effectsTMP-SMX Increased risk of worsening of

renal function and hyperkalemia

Abbreviations: ACEI, angiotensin-converting enzyme inhibitor;ARB, angiotensin receptor blocker; CNS, central nervous system;SIADH, syndrome of inappropriate antidiuretic hormone secretion;SNRI, serotonin-norepinephrine reuptake inhibitor; TMP-SMX,trimethoprim-sulfamethoxazole.


deprescribing resources for many medications included inthe 2019 AGS Beers Criteria® is https://deprescribing.org.

Of particular note is the potential role for nonpharmacolo-gical approaches to manage common conditions in older adults.The evidence base for specific nonpharmacological approacheswith a person-centered approach to care is small butgrowing.29–32 One example of the growing evidence for non-drug alternatives is in the area of care for persons with dementiaand delirium. Scales and colleagues published a 2019 compre-hensive review of evidence-based nonpharmacologicalapproaches for behavioral and psychological symptoms ofdementia. They evaluated 197 articles that included sensorypractices (eg, massage, light therapy), psychosocial practices (eg,music, pet therapy, reminiscence), and structured care protocols(eg, mouth care, bathing). Though they had reco

American Geriatrics Society 2019 Updated AGS Beers ...American Geriatrics Society 2019 Updated AGS Beers Criteria ... Case reports, case series, letters to the editor, and editorials

Documents