Equianalgesic Dose Ratios for Opioids: A Critical Review and Proposals for Long-Term Dosing

© U.S. Cancer Pain Relief Committee, 2001 0885-3924/01/$–see front matterPublished by Elsevier, New York, New York PII S0885-3924(01)00294-9

672 Journal of Pain and Symptom Management Vol. 22 No. 2 August 2001

Original Article

Equianalgesic Dose Ratios for Opioids: A Critical Review and Proposals for Long-Term Dosing

Jose Pereira, MB ChB, CCFP, Peter Lawlor, MB, CCFP, Antonio Vigano, MD, Marlene Dorgan, BA, MLS, and Eduardo Bruera, MD

Division of Palliative Medicine, (J.P.) University of Calgary and Tertiary Palliative Care Unit, Calgary, Alberta, Canada; Palliative Care Program (A.V.) and Regional Palliative Care Program (P.L.), Grey Nuns Community Hospital & Health Centre, Edmonton, Alberta, Canada; John W. Scott Health Sciences Library (M.D.), University of Alberta, Edmonton, Alberta, Canada; and Department of Symptom Control and Palliative Care, (E.B.) University of Texas, MD Anderson Cancer Center, Houston, Texas, USA

Abstract

Clinicians involved in the opioid pharmacotherapy of cancer-related pain should be acquainted with a variety of opioids and be skilled in the selection of doses when the type of opioid or route of administration needs changing. The optimal dose should avoid under-dosing or overdosing, both associated with negative outcomes for the patient. Although equianalgesic dose tables are generally used to determine the new doses in these circumstances, the evidence to support the ratios indicated in these tables largely refers to the context of single dose administration. The applicability of these ratios to the setting of chronic opioid administration has been questioned. A systematic search of published literature from 1966 to September 1999 was conducted to critically appraise the emerging evidence on equianalgesic dose ratios derived from studies of chronic opioid administration. There were six major findings: 1) there exists a general paucity of data related to long-term dosing and studies are heterogeneous in nature; 2) the ratios exhibit extremely wide ranges; 3) methadone is more potent than previously appreciated; 4) the ratios related to methadone are highly correlated with the dose of the previous opioid; 5) the ratio may change according to the direction the opioid switch; and 6) discrepancies exist with respect to both oxycodone and fentanyl. Overall, these findings have important clinical implications for clinicians and warrant consideration in the potential revision of current tables. The complexity of the clinical context in which many switches occur must be recognized and also appreciated in the design of future studies.

J Pain Symptom Manage 2001;22:672–687.

© U.S. Cancer Pain Relief Committee, 2001.

Key Words

Opioid, equianalgesic dose ratios, potency, rotation, switching, chronic pain, opioid toxicity,

tolerance

Introduction

Opioids are the mainstay of pharmacologicalmanagement of moderate to severe cancer pain,and morphine, administered orally, is generally

Address reprint requests to:

Jose Pereira, MB ChB, CCFP,Palliative Care Unit, Foothills Medical Centre, PalliativeCare Office, Room 710, South Tower, 1403–29 StreetN.W., Calgary, AB, T2N 2T9 Canada.

Accepted for publication: November 7, 2000.

Vol. 22 No. 2 August 2001 Equianalgesic Dose Ratios for Opioids 673

considered to be the drug and route of choicerespectively.

1

However, it is recognized that theroute of administration and the type of opioidmust, on occasion, be changed.

2,3

The variousreasons for these changes include: dysphagia,the practical inconvenience when large quanti-ties of a particular formulation are required,poor response to a particular opioid,

4

and asan emerging strategy in the management ofopioid-related adverse effects, particularly neu-rotoxicity.

4–9

When switching routes or opioids,the goal is to achieve optimal analgesia andavoid the toxicity associated with overdosingand the inadequate pain control that accompa-nies under-dosing.

In order to assist physicians in the calcula-tion of doses when switching opioids, guide-lines in the form of equianalgesic dose tablesare available.

10–14

Recent reports have high-lighted deficiencies in these tables.

15,16

Currenttables are derived from the results of earlierstudies of relative potency ratios using singledose crossover designs.

17–23

These studies, al-though elegant in design, generally utilizedsingle doses and involved subjects who had lim-ited opioid exposure, both in duration anddose. They did not reflect the clinical realitiesof chronic opioid administration in the treat-ment of cancer pain, including switches madein the context of opioid-related neurotoxicityor lack of response to a particular opioid. Theextrapolation of the results from these studiesto the context of opioid tolerant patients may,therefore, be invalid. Recognizing the limita-tions of early studies, we conducted a reviewwith the aim of critically appraising the evi-dence supporting the ratios related to chronicadministration and determining limitations incurrently quoted ratios.

Methods

For the purpose of this review an “equianal-gesic dose ratio” refers to the ratio of the doseof two opioids required to produce the sameanalgesic effect. In this article, an equianalge-sic dose ratio of morphine to hydromorphoneof 5:1, for example, indicates that 10 mg oforally administered morphine is equianalgesicto 2 mg of orally administered hydromorphone;in other words, hydromorphone is 5 times morepotent than morphine. The literature some-

times refers to the “relative potency ratio.”From a pragmatic perspective, this is the in-verse of the equianalgesic dose ratio. For exam-ple, the “relative potency ratio” of morphine tohydromorphone is 1:5. Preference will begiven to the use of the term “equianalgesicdose ratio” (EDR) in this article. The reportedEDR in this article will also imply the sameroute of administration for the opioids beingcompared, unless otherwise stated. However,some of the studies identified in this reviewcompared relative opioid consumption ratherthan EDR per se, and inferred an EDR fromthat comparison in the absence of applying anyformal measure of analgesia such as a visual an-alog score. Other studies refer to the term “rel-ative potency” which can apply to any opioideffect, including analgesia. The “relative po-tency” between morphine and hydromor-phone might therefore be reported as 1:5.

The review was restricted to the equianalge-sic dose ratios between morphine and the morecommonly utilized “strong” opioids, namely hy-dromorphone, oxycodone, methadone, andfentanyl. Although the major focus of the re-view was to compare equianalgesic dose ratiosbetween opioids rather than between routes,literature related to ratios between the morefrequently used routes of administration wasalso evaluated. These included the oral (PO),subcutaneous (SC), intravenous (IV), intramus-cular, and rectal routes.

A systematic search of the literature was con-ducted. The following databases were searched:MEDLINE from 1966 to September 1999; CAN-CERLIT from 1983 to September 1999; and EM-BASE from 1988 to September 1999. The searchterms included:

opioid/s, opiate/s, narcotic/s, mor-phine, hydromorphone, methadone, fentanyl, oxy-codone, equianalges*, dose ratio, potency, potency ratio,

and

tables.

A broad free-text search method wasutilized and included combinations of thesewords. The inclusion criteria consisted of retro-spective and prospective human studies that em-ployed multiple opioid dosing models ratherthan single doses. Hand searches of 9 major text-books addressing opioid therapy and the man-agement of cancer pain were conducted. Meet-ing abstracts of the International Association forthe Study of Pain and the American Society forClinical Oncology were searched. Abstracts thatwere later published in full were not included.The reference lists of publications identified by

674 Pereira et al. Vol. 22 No. 2 August 2001

the database and hand searches were examinedfor further references.

The following data were extracted from eli-gible studies: 1) the opioids being compared,2) the study design, 3) the dosing period ornumber of doses, 4) the number of subjects inthe study, 5) the types of patients (i.e., cancerpatients, perioperative patients, non-cancer pain,etc.), 6) the reported dose ratio findings, and 7)adverse effects or complications reported.

Results

Eighteen studies met the inclusion criteria.The limited number of studies and the heteroge-neity of patient groups, methods, and clinicalsettings precluded the conduct of a formal meta-analysis. The relative lack of studies examiningratios related to changes in routes of administra-tion did not allow for a detailed assessment.

Hydromorphone

The studies identifying the equianalgesicdose ratios between hydromorphone and otheropioids are presented in Table 1.

24–27

Two ofthese studies

24,26

originated from the same cen-ter but related to different patient groups. Bothof these studies involved switches between mor-phine and hydromorphone. The previous opi-oid was administered for at least 48 hours be-fore the opioid switch. The time to reachstabilization was a minimum of 24 hours

24

and48 hours

26

respectively, followed by a 48-hourperiod of stable dosing. The earlier study

24

in-volved opioid switches that also involved achange of route of administration, whereas themore recent one

26

involved opioid switcheswithout a change in route of administration. Al-though the latter study included both oral andsubcutaneous routes, there were no switches be-tween routes. No significant difference wasfound in the EDR between morphine to hydro-morphone involving oral to oral and subcutane-ous to subcutaneous switches. Both studies re-vealed marked interindividual variability and nocorrelation between the previous opioid doseand the relative equianalgesic dose ratio. Themorphine to hydromorphone (M:HM) EDR de-rived from these two studies were 5:1 and 5.3:1respectively. However, the EDRs (expressed asM:HM) for the HM to M switches were 3.5:1 and3.7:1, differing significantly from the ratio for Mto HM switches (

P

�

0.0001). Lawlor and col-

leagues calculated a unified overall dose ratio ofM:HM of 4.3:1 (3.3–4.8, lower quartiles).

Dunbar et al. studied the consumption of mor-phine and hydromorphone by patient controlledanalgesia for the treatment of bone marrowtransplantation-related mucositis.

25

Two differentgroups of patients over a seven-day period (be-ginning 7 days after transplantation) were com-pared. All patients were on an analgesic regi-men for at least 4 days prior to the start of thestudy. There was no randomization of patientsto treatment groups. An initial EDR of mor-phine to hydromorphone of 5:1 was assumed.The mean daily drug use yielded a morphineto hydromorphone dose ratio of 3.55:1. Al-though patient satisfaction was similar betweenthe two groups, resting pain scores were higherin the hydromorphone group. The investiga-tors noted that although a ratio of 7:1, which isoccasionally quoted, would have sufficed in thefirst 2 days of the study, a ratio of 3:1 was moreappropriate by the end of the study (7 dayslater). They postulate that the difference be-tween short term and longer term administra-tion could be the result of: 1) pharmacokineticvariability inherent in short term studies; 2)opioid metabolite accumulation; or 3) patientsdeveloping drug tolerance at different levels.

A recent study by Miller et al. compared theanalgesic efficacy and adverse effects of hydro-morphone and morphine delivered by contin-uous SC infusion in 74 patients with cancer-related pain admitted to a palliative care unit.

27

The study incorporated a double-blind, ran-domized, controlled design and excluded cog-nitively impaired patients. The most commonreason for the switches was the inability to takemedications orally. The initial PO to SC EDRused for morphine was 2:1 and the EDR of sub-cutaneous morphine to hydromorphone as5:1. Although the study was not specifically de-signed to establish equianalgesic dose ratios, itwas reported that, when using a dose ratio ofmorphine to hydromorphone of 5:1, the hy-dromorphone group (

n

�

33) required signifi-cantly more opioid rescue doses in the first 24hours than the morphine group (

n

�

41).Over the course of the study, both groupsneeded an increase in dose (52% and 63% ofpatients in the morphine and hydromorphonegroups respectively—the difference was notsignificant). The dose increases were 14% inthe hydromorphone group and 17% in the


Tab

le 1

Hyd

rom

orph

one:

Equ

iana

lges

ic D

ose

Rat

ios

Bet

wee

n H

ydro

mor

phon

e an

d O

ther

Opi

oids

Aut

hor

Opi

oids

(‘t

o’de

not

es d

irec

tion

of

sw

itch

) (r

oute

s)D

esig

n (

Stud

y D

urat

ion

)

n

Dia

gnos

is(r

easo

ns

for

swit

chin

g)E

quia

nal

gesi

cD

ose

Rat

ioO

ther

Fin

din

gsan

d C

omm

ents

Bru

era

et a

l 19

96 [

24]

M to

HM

,

HM

to M

(PO

), S

C, P

R)

SC to

PO

/PR

n

�

37

SC to

PO

n

�

29

SC to

SC

n

�

23

Ret

rosp

ecti

ve (

48 h

ours

of

sta

bilit

y af

ter

opio

idsw

itch

req

uire

d)

M to

HM

36

HM

to M

12

Can

cer

(Tox

icit

y [3

2],

esca

lati

ng

dose

[4]

,bo

th to

xici

ty a

nd

esca

lati

ngd

ose

[5],

ot

her

[6]

M to

HM

a

5.3

:1(4

.9–6

.4:1

) H

M to

M

a

3.6

:1(3

.3–5

.0:1

)

PO: S

C E

DR

of

2:1

use

d.C

ompl

icat

ion

s re

port

edfo

r sw

itch

es to

met

had

one.

Dun

bar

et a

l.19

96 [

25]

M &

HM

PC

A (

IV to

IV

)Pr

ospe

ctiv

e, o

pen

(7

day

s)M

36

HM

21

Bon

e-m

arro

w tr

ansp

lan

tm

ucos

itis

St

able

pai

n

3.5:

1

b

No

com

plic

atio

ns

repo

rted

.

Ass

umed

5:1

rat

io a

tin

itia

tion

of s

tudy

.L

awlo

r et

al

1997

[26

]M

to H

MH

M to

M(S

C to

SC

n

�

69)

PO to

PO

n

�

22)

Ret

rosp

ecti

ve (

48 h

ours

of

stab

ility

aft

er o

pioi

dsw

itch

req

uire

d)

91 s

wit

ches

in 7

4 co

nse

cuti

ve s

ubje

cts

M to

HM

44

HM

to M

47

Can

cer

(Opi

oid

toxi

city

in m

ajor

ity

of c

ases

)M

:HM

of 5

.0:1

(4

.2–5

.9:1

)

a

HM

:M (

expr

esse

d as

M:H

M)

3.7:

1 (2

.9–4

.5)

Dos

e ra

tio

not

dep

ende

nt

on th

e do

se o

f th

epr

evio

us o

pioi

d.E

DR

of P

O :

SC o

f 2:1

use

d or

adv

erse

eff

ects

.M

iller

et a

l 19

99 [

27]

To

M &

HM

(C

I SC

to S

C)

R, D

B, c

ontr

olle

d tr

ial

(72

hou

rs)

74/1

33

M 4

1 H

M 3

3

Palli

ativ

e pa

tien

ts

(87%

) h

ad c

ance

r)E

DR

of m

orph

ine

toh

ydro

mor

phon

e of

5:1

assu

med

an

d PO

: SC

of

2:1

. Pat

ien

tsin

HM

gro

upre

quir

ed m

ore

brea

kth

roug

hs

Ass

esse

d op

ioid

con

sum

ptio

n.

Hig

h a

ttri

tion

.Si

de e

ffec

ts r

epor

ted.

M

�

mor

phin

e; H

M

�

hyd

rom

orph

one;

C

�

sub

cuta

neo

us r

oute

; ED

R

�

Equ

ian

alge

sic

dose

rat

ion

; PO

�

ora

l rou

te; I

V

�

intr

aven

ous

rout

e; P

CA

�

pat

ien

t con

trol

led

anal

gesi

a; C

I

�

con

tin

uous

infu

sion

;R

�

ran

dom

ized

; DB

�

Dou

ble

Blin

d; X

�

Cro

ssov

er.

a

Exp

ress

ed a

s m

edia

n (

firs

t-th

ird

quar

tile

s) in

the

dire

ctio

n o

f M to

HM

.

b

Mea

n v

alue

.


morphine group (

P

�

0.3). The authors sug-gest that the morphine to hydromorphone

po-tency ratio

of 1:5 was too low for almost one-halfof the patients (stated otherwise: the morphineto hydromorphone EDR of 5:1 was too high).Twenty-eight percent of patients died beforecompletion of the study. Unfortunately, be-cause of progressive disease, only 8 of the pa-tients who completed the study were capable ofrecording treatment side effects and completeself-assessments of their pain using visual ana-logue scales. Pain control and adverse effectswere assessed by proxy rating in the remainderof those completing the study. This introduceda potential limitation in the interpretation ofthe study findings.

Generally, these ratios are consistent withthose found in the equianalgesic tables of mostcurrent resources offering guidelines on themanagement of chronic pain. However, it isimportant to recognize the wide ranges inequianalgesic dose ratios, reflecting the largedegree of interindividual variation.

Methadone

The studies exploring the equianalgesic doseratios between methadone

24,28–30

and other opi-oids are presented in Table 2. Bruera et al.

24

re-ported a retrospective study of 65 switches fromsubcutaneous hydromorphone to oral (

n

�

37) or rectal (

n

�

28) methadone. An overallsubcutaneous hydromorphone to oral metha-done EDR 1.14:1 was calculated, in turn sug-gesting a morphine to methadone EDR of 10:1using an EDR of M:HM of 5:1. Patients wereswitched because of poor pain control (

n

�

24), opioid-related neurotoxicity (

n

�

28) orboth (

n

�

15). After the switch, the EDR wascalculated when patients were deemed to bestable. Stability was defined as 48 consecutivehours without an increase in pain intensity, asmeasured by visual analog scales, and the re-quirement of 3 or less rescue doses. The proto-col used for switching to methadone entailed agradual elimination of the previous opioid andintroduction of methadone over a 3-day pe-riod. Stabilization occurred within a range of 1to 6 days following the switch. There was a pos-itive correlation between the dose of the previ-ous opioid and the calculated EDR. The me-dian EDR of subcutaneous hydromorphone tooral methadone was found to be 1.6:1 in pa-tients who had previously been on more than

330 mg of hydromorphone per day, comparedto a median EDR of 0.95:1 in those patientstaking less than 330 mg per day. Eight patientswere reported to have developed respiratorydepression during the switch. This was success-fully reversed in all cases.

Two follow-up retrospective studies

28,29

con-firmed the association between the equianalge-sic dose ratio and the dose of the previous opi-oid. Ripamonti et al. combined data from twocenters and compared dose ratios in patientsswitched from low (

n

�

37) and high (

n

�

51)doses of previous opioid.

29

To facilitate com-parison, opioid doses were standardized to sub-cutaneous hydromorphone equivalents using adose ratio of 5:1 for morphine to hydromor-phone. The median (range) dose for the high-and low-dose groups were 236 mg (36–1080 mg)and 3 mg (1–60mg) per day, respectively. Themedian (lower and upper quartiles) EDR of hy-dromorphone (SC) to methadone (PO) for pa-tients switched from low doses of opioid (i.e.,doses equivalent to

�

3 mg of SC hydromor-phone) to PO methadone was 0.17:1 (0.14:1–0.25:1), compared to 1.5:1 (1.1:1–2.58:1) forpatients switched from higher opioid doses(equivalent to

�

300 mg of hydromorphoneSC). There was a highly positive correlation be-tween the equianalgesic dose ratio and thedose of the previous opioid prior to switchingto methadone (R

2

�

0.19,

P

�

0.0001). Side ef-fects were not reported.

Although the retrospective study by Lawloret al.

28

originated from the same center as theone by Bruera et al.,

24

a different cohort of pa-tients was studied. Consecutive switches fromPO, SC, or IV morphine to PO methadone in14 patients were evaluated. Patients were switchedfrom doses ranging from 85 mg to 24027 mg oforal morphine per day (for PO to SC morphinean EDR of 2:1 was used). Six additional switchesoccurred from methadone to morphine. Thesewere included in the analysis only to determinethe overall ratio. This study provided furtherevidence of methadone’s high potency relativeto morphine and the presence of a changingEDR as determined by the dose of the previousopioid. The main limitation of this study wasthe relatively small sample size. Side effectswere not reported.

In a prospective study, Ripamonti et al. haveprovided the most compelling evidence to dateconfirming that the equianalgesic ratio is not


Tab

le 2

Met

hado

ne: E

quia

nalg

esic

Dos

e R

atio

s B

etw

een

Met

hado

ne, M

orph

ine

and

Oth

er O

pioi

ds

Aut

hor

Opi

oids

(‘t

o’

den

otes

dir

ecti

onof

sw

itch

) (r

oute

s)

Des

ign

(Stu

dy D

urat

ion

)

n

Dia

gnos

is

(rea

son

s fo

r sw

itch

ing)

Equ

ian

alge

sic

Dos

e R

atio

Oth

er F

indi

ngs

an

dC

omm

ents

Bru

era

et a

l19

96 [

24]

(See

Tab

le 1

)

HM

to M

eth

HM

(SC

) to

an

dfr

om M

eth

(PO

,PR

)

Ret

rosp

ecti

ve (

48 h

ours

of

sta

bilit

y af

ter

opio

id

swit

ch r

equi

red)

65C

ance

r (O

pioi

d n

euro

toxi

city

[8]

, poo

r re

spon

se [

24],

bot

h

neu

roto

xici

ty a

nd

poor

re

spon

se [

15],

an

d [2

])

Ove

rall:

1.1

4:1

(0.5

2–2.

04:1

)

a

Acc

ordi

ng

to d

ose

ran

ges

c

:

⇒

13–3

00 m

g: 0

.95:

1(0

.2–1

2.3:

1

b

⇒

�

300

mg:

1.6:

1 (0

.3–1

4.4:

1)

b

⇒

13–2

076

mg:

1.14

:1 (

0.5–

2.04

:1)

b

Res

pira

tory

dep

ress

ion

repo

rted

in 8

cas

es (

all

reve

rsed

). C

orre

lati

on

wit

h p

revi

ous

dose

not

ed.

Var

iabi

lity

of E

DR

not

ed.

Law

lor

et a

l 19

98 [

28]

M to

ME

M

(IV

, PO

, SC

)to

an

d fr

omM

eth

(PO

, PR

)

Ret

rosp

ecti

ve (

48 h

ours

of s

tabi

lity

afte

r op

ioid

swit

ch r

equi

red)

20C

ance

r (O

pioi

d to

xici

ty,

poor

res

pon

se)

Ove

rall:

11.

20:1

(5.

06–1

3.24

:1)

a

Acc

ordi

ng

to d

ose

ran

ges

c

:

⇒

42–1

2012

mg

d

: 11.

4:1

(5.9

–16.

3)

a

⇒

3–24

0 m

g

d

: 8.

2:1

(4.3

7–11

.3)

a

Cor

rela

tion

wit

h p

revi

ous

dose

: r

�

0.8

6. V

aria

bilit

y of

dos

e ra

tios

not

ed. A

dver

seef

fect

s/co

mpl

icat

ion

s n

ot r

epor

ted.

Rip

amon

ti e

t al

1998

[29

]H

M to

Met

h

M, O

xy, C

od, B

up(S

C/

PO to

PO

)

Ret

rosp

ecti

ve (

48 h

ours

of

sta

bilit

y af

ter

opio

idsw

itch

req

uire

d)

88C

ance

rO

vera

ll: 0

.51:

1 (0

.20–

1.38

)

a,e

Acc

ordi

ng

to d

ose

ran

ges

c

:

⇒

36–

1080

mg:

1.

47:1

(0.

81–2

.47:

1)

a

⇒

oth

ers

0.25

:1(0

.17–

0.44

:1)

e

Adv

erse

eff

ects

/co

mpl

icat

ion

s n

ot r

epor

ted.

Rip

amon

ti e

t al

1998

[30

]M

to M

eth

(PO

to P

O)

Cro

ss-s

ecti

onal

, Pro

spec

tive

(D

ays

to s

tabi

lity

afte

rsw

itch

. See

text

)

38C

ance

rO

vera

ll: 7

.75:

1 (5

–10:

1)

b

Acc

ordi

ng

to d

ose

ran

ges

c

:

⇒

30–9

0 m

g: 3

.7:1

(2.

5–8.

8:1)

b

⇒

90–3

00 m

g: 7

.75:

1 (4

–10:

1)

b

⇒

300

mg:

12.

25:1

(10

–14.

3:1)

b

Stro

ng

corr

elat

ion

(r

�

0.9

1)be

twee

n p

revi

ous

opio

id

dose

an

d E

DR

. Adv

erse

effe

cts/

com

plic

atio

ns

not

repo

rted

.

M

�

mor

phin

e; H

M �

hyd

rom

orph

ine;

Met

h �

met

had

one;

Oxy

� o

xyco

don

e B

u �

bup

ren

orph

ine;

SC

� s

ubcu

tan

eous

rou

te; P

O �

ora

l rou

te; I

V �

intr

aven

ous

rout

e; P

R �

rec

tal r

oute

; ED

R �

equ

ian

al-

gesi

c do

se r

atio

. a M

edia

n d

ose

rati

o (1

st-3

rd q

uart

iles)

b M

edia

n (

ran

ge)

c Dai

ly d

ose

in m

g of

the

prev

ious

opi

oid

(th

e op

ioid

sub

ject

s w

ere

on p

rior

to s

wit

chin

g to

met

had

one)

d E

xpre

ssed

as

mor

phin

e eq

uiva

len

t sub

cuta

neo

usly

(sc

)e D

oses

of i

nit

ial o

pioi

d (M

, Oxy

, Cod

, Bup

) w

ere

con

vert

ed to

a h

ydro

mor

phon

e eq

uiva

len

t


fixed when switching to methadone from otheropioids.30 Thirty-eight cancer patients were in-cluded in an open study. For data analysis, thepatients were grouped according to the totaldaily dose of morphine prior to the switch. Theequianalgesic dose ratios determined (Table2) varied according to the previous opioiddose. Wide EDRs (2.5:1 to 14.3:1) were notedin this prospective study. Although pain was as-sessed, adverse effects or complications werenot reported.

OxycodoneThe studies addressing the equianalgesic

dose ratios between oxycodone31–36 and otheropioids are presented in Table 3. Kalso andVainio switched 20 patients with uncontrolledcancer pain who were taking so-called “weak”opioids (and nonsteroidal anti-inflammatorydrugs in some cases) to intravenous morphineor oxycodone in an elaborate randomized, dou-ble-blinded, crossover design.31 Patients first ti-trated themselves pain-free using IV patient-controlled analgesia for 48 hours. After 48hours they were switched to the PO form of ei-ther morphine or oxycodone. In switching tothe PO route the researchers assigned theirsubjects to two different groups, each with 10subjects (Group 1 and Group 2). In Group 1,the bioavailabilities of morphine and oxycodonewere assumed to be 44% and 66% respectivelyand for Group 2, the bioavailabilities were as-sumed to be 33% and 50% respectively. It is notclear why the investigators chose these levels ofbioavailability, but in their original manuscriptthey refer to the wide ranges of bioavailabilityreported in other studies. Following the switchin route, patients were able to adjust their oraldoses to improve pain control. This phaselasted 48 hours and was followed by the cross-over phase to the alternative opioid. The rou-tine of first establishing pain control with theintravenous formulation of the opioid followedby a switch to the oral form was repeated. Inthis study, the investigators reported “potencyratios”. The median calculated PO to IV po-tency ratios (giving comparable analgesia)were 0.31:1 for morphine and 0.70:1 for oxyc-odone. An equianalgesic dose ratio (EDR) ofPO oxycodone to PO morphine of 3:4 wasnoted (giving an inferred morphine to oxy-codone EDR of 1.48:1). Patients receiving in-travenous oxycodone required approximately

30% more opioid to produce the same allevia-tion of pain as with intravenous morphine.This gives an inferred IV morphine to IV oxyc-odone equianalgesic dose ratio of 0.7:1. Wideranges were noted, particularly during oral ad-ministration of the opioids. Morphine causedmore nausea than oxycodone and hallucina-tions (5 cases) occurred only during morphinetreatment. There were no other major differ-ences in the side effects between the two opio-ids. This result was consistent with previousstudies by Beaver et al.17,18

Glare and Walsh enrolled 24 patients withcancer-related pain32 in an open, non-random-ized study with the aim of primarily assessingthe safety and efficacy of oxycodone. The ma-jority of patients (n � 19) were already takingoxycodone, in combination with acetaminophenor aspirin, upon enrollment. Twenty patientscompleted the study. Fifteen patients required adose adjustment. The median number of doseadjustments required following the initial switchwas 1 (range 0–6) and the median number ofdays to reach the effective dose were 3 (range,1–10), indicating relatively stable patients withrespect to pain control. Ten patients had occa-sion to be switched from PO oxycodone tomorphine administered orally or parenterally.Limited information is given about these pa-tients and the number of patients switched toPO or parenteral morphine is not indicated inthe original paper. Using potency ratios of POoxycodone to PO morphine of 1:1 and PO oxy-codone to parenteral morphine of 3:1, they ob-served that analgesia was effectively and safelymaintained in this small number of subjects.

Silvasti et al. compared the consumption ofmorphine and oxycodone in two different post-operative patient groups over a 24-hour period.35

Either morphine or oxycodone were adminis-tered in IV boluses using a patient controlled an-algesic pump. No difference in pain scores wasnoted between the two patient groups through-out the 24-hour period. The consumption ofmorphine and oxycodone was similar in thetwo groups and the authors concluded that thetwo opioids were equipotent.

Three recent studies have suggested that oxy-codone is relatively more potent than mor-phine.33,34,36 All three studies involved cancerpatients, two incorporated a double-blind, ran-domized, crossover design.33,34 In the study byHeiskanen and Kalso, 45 patients were en-


Tab

le 3

Oxy

codo

ne: E

quia

nalg

esic

Dos

e R

atio

s of

Oxy

codo

ne R

elat

ive

to O

ther

Opi

oids

Aut

hor

Opi

oids

(‘t

o’de

not

es d

irec

tion

of

sw

itch

) (r

oute

s)D

esig

n(S

tudy

Dur

atio

n)

nD

iagn

osis

(rea

son

s fo

r sw

itch

ing)

Equ

ian

alge

sic

Dos

e R

atio

aO

ther

Fin

din

gs

and

Com

men

ts

Kal

so a

nd

Vai

nio

19

90 [

31]

M &

Oxy

(O

xy I

V to

Oxy

PO

M I

V to

M P

O

M P

O to

Oxy

PO

M

IV

to O

xy I

VO

xy I

V to

M I

V)

R, D

B, X

(96

hrs

� 9

6 h

rs)

20C

ance

rO

xy I

V to

Oxy

PO

0.7

:1

(0.4

9–0.

78:1

) M

IV

toM

PO

0.3

1:1

(0.1

7–0.

35:1

)M

PO

to O

xy P

O 1

.48:

1b

MIV

to O

xy I

V 0

.7:1

(0

.6–0

.88:

1)

Gla

re &

Wal

sh19

93 [

32]

Oxy

PO

to M

PO

or I

V (

PO to

PO

or I

V)

Pros

pect

ive,

ope

n,

not

ran

dom

ized

(a

t lea

st 5

day

s)

10 o

ut o

f 24

Can

cer

Oxy

PO

to M

PO

1:1

Oxy

PO

to M

IV

3:1

Des

ign

ed to

ass

ess

the

effi

cacy

an

d sa

fety

of

oxy

codo

ne.

Hei

skan

en &

Kal

so P

ain

1993

[33

]M

to O

xy

Oxy

to M

(C

R P

O to

PO

)

R, D

B, X

(48

hrs

to s

tabi

lize

then

3–6

days

)27

eva

luab

le o

utof

45

enro

lled

Can

cer

(Sta

ble

pain

)O

xy to

M 2

:3c

M to

Oxy

4:3

c

Bru

era

et a

l19

98 [

34]

M to

Oxy

Oxy

to M

(PO

)R

, DB

, X (

7day

s)23

out

of 3

1C

ance

r (S

tabl

e pa

in)

M to

Oxy

1.5

:1O

xy to

M (

1–2.

3:1)

Silv

asti

et a

l19

98 [

35]

Oxy

IV

(PC

A)

M I

V (

PCA

)R

, DB

, (24

hou

rs)

24 (

Oxy

IV

] 25

(M

IV

)Po

st-o

pera

tive

M to

Oxy

1:1

d

Gag

non

et a

l19

99 [

36]

M &

HM

to O

xy (

SC)

Ret

rosp

ecti

ve

(min

imal

24

hou

rs

of s

tabi

lity)

19 (

8 M

to O

xy)

(11

HM

to O

xy)

Can

cer

(opi

oid-

rela

ted

toxi

city

, poo

r re

spon

se)

Ove

rall:

M to

Oxy

1.4

:1

M to

Oxy

1:1

HM

to

Oxy

0.4

:1

M �

mor

phin

e; M

� h

ydro

mor

phin

e; O

xy �

oxy

codo

ne;

SC

� s

ubcu

tan

eous

rou

te; P

O �

ora

l rou

te; I

V �

intr

aven

ous

rout

e; C

R �

con

trol

led

rele

ase

form

ulat

ion

; PC

A �

pat

ien

t co

ntr

olle

d an

alge

sia;

R �

ran

dom

ized

; DB

� d

oubl

e-bl

ind;

X �

cro

ssov

er.

a Rat

ios

expr

esse

d in

the

dire

ctio

n o

f th

e sw

itch

un

less

oth

erw

ise

stat

ed. R

atio

s ar

e qu

oted

as

per

orig

inal

pap

ers

- med

ian

(an

d ra

nge

) w

her

e av

aila

ble.

b Rat

io w

as in

ferr

ed fr

om m

edia

n d

oses

giv

en in

the

orig

inal

pap

er.

c Bas

ed o

n c

onsu

mpt

ion

.d E

xpre

ssed

as

mor

phin

e eq

uiva

len

t (m

orph

ine

to h

ydro

mor

phon

e ra

tio

of 5

:1 u

sed)

.


rolled to compare the steady-state pharmaco-dynamic profiles, safety, and efficacy of con-trolled-release PO formulations of oxycodoneand morphine.33 Patients were first random-ized to receive either oxycodone or morphinein an open-label titration phase. Based partlyon the premise that oxycodone has beenshown to have a higher and less variable oralbioavailability than morphine,31,37 an EDR ofoxycodone to morphine of 2:3 was selected toguide the initial switches. When a stable dailydose was reached, the double-blind crossoversequence was randomized. Using a doubleblind approach, patients were administeredmedication for a minimum of 3 days to ensurethat a steady state was reached. Twelve patientswere titrated with oxycodone and 15 with mor-phine. Since there was a difference in the painscores and the amount of rescue analgesicsused between the oxycodone and morphinestable phases, the dosing cannot be consideredequianalgesic. Patients receiving oxycodone inthe first phase, most of them after oxycodonetitration, had better pain control than those re-ceiving oxycodone in the second phase. Theoxycodone to morphine total opioid consump-tion ratio was lower in patients receiving oxy-codone first (2:3) as compared to patients re-ceiving oxycodone in the second phase (3:4).The authors speculate that this is likely relatedto the crossover design and accompanying pe-riod effect (they propose that crossover designsshould, in the future, include titration to effectwith each opioid.). The authors postulate thatthe relative equianalgesic doses of various opi-oids are different at the initiation of opioidtreatment when compared to longer term doseexposure. There was wide individual variationin both pain control and the incidence of ad-verse effects. There were trends towards lessvomiting and nightmares with oxycodone ascompared to morphine.

Gagnon et al. reviewed their program’s ex-perience with parenteral oxycodone in a retro-spective review.36 Sixty-three patients with ad-vanced cancer had been switched to oxycodoneby SC administration from a variety of otherstrong opioids. Most of the patients had beenswitched to oxycodone because of opioid neu-rotoxicity. A subgroup of 19 patients who wereswitched from morphine (n � 8) and hydro-morphone (n � 11) were selected to deter-mine the EDR between morphine and oxy-

codone. To facilitate the comparison, patientswho received hydromorphone had their dosesconverted to morphine SC equivalents, using amorphine to hydromorphone EDR of 5:1. Themean (�SD) and median combined morphineto oxycodone EDR were 1.9:1 � 1.5:1 and 1.4:1,respectively. In effect, this translates to 15 mg ofmorphine SC being equivalent to 10 mg of oxy-codone SC. Local skin/site reactions to the oxy-codone administration were observed but other-wise no adverse effects were recorded.

FentanylThe equianalgesic dose ratios between par-

enteral fentanyl38–41 and other opioids arelisted in Table 4. All four studies identified arerelatively small.38–41 The only prospective studyidentified in this group aimed to compare theefficacy and adverse effects between subcutane-ously administered fentanyl and morphine,rather than to specifically determine the equi-analgesic ratio between the two opioids.41 Ofthe 50 hospice patients with cancer-relatedpain who were approached to take part in thisstudy, 30 participated and 23 completed it. Pa-tients entered in the study were generally sta-ble, lucid and tolerating morphine well. Usinga double-blind, randomized, crossover design,patients were switched to morphine or fentanylusing a conversion EDR of morphine 10 mg:fentanyl 150 �g (i.e., a potency ratio of fenta-nyl to morphine of 66:1).41 Three days laterthe opioids were switched. Because of difficul-ties blinding the breakthrough medication,meperidine (pethidine) was used for rescuedoses. This may have affected the interpreta-tion of the results. Patients on fentanyl re-ported significantly more pain on the secondday. Overall, the morphine-first group had alower dose of opioid throughout the studythan those patients who received fentanyl first(most patients were receiving infusion dosesequivalent to less than 100 mg of morphineper day). The number of rescue doses was sig-nificantly greater in the fentanyl-first group ondays 2 and 3 than the morphine-first group.The authors postulate that this may suggest aninadequate ratio for converting PO morphineto SC fentanyl, or it may be that the steady stateof SC fentanyl was not reached before PO mor-phine levels diminished (15 patients had beenon PO controlled release formulations of mor-phine prior to the study). No patients experi-


Tab

le 4

Fent

anyl

: Rel

ativ

e A

nalg

esic

Pot

ency

Rat

ios

Bet

wee

n Fe

ntan

yl a

nd M

orph

ine

Aut

hor

Opi

oids

(‘t

o’

den

otes

dir

ecti

onof

sw

itch

) (r

oute

s)D

esig

n (

Stud

y D

urat

ion

)n

Dia

gnos

is(r

easo

ns

for

swit

chin

g)Po

ten

cy R

atio

aO

ther

Fin

din

gs a

nd

Com

men

ts

Paix

et a

l19

95 [

38]

M to

Fen

t (C

SCI)

(H

M

to M

) (M

to S

uf)

Ret

rosp

ecti

ve9

out o

f 11

Can

cer

(Tox

icit

y)SC

Fen

t to

SC M

68:

1(r

ange

15–

100:

1) (

SD �

) 23

:1W

oodh

ouse

A e

t al

1996

[39

]M

, pet

hid

ine

(mep

erid

ine)

, Fen

tPr

ospe

ctiv

e 48

hrs

55 M

orph

ine

(19)

Fe

nta

nyl

(18

)Po

stop

erat

ive

(var

iety

of

indi

cati

ons)

Use

d E

DR

of 1

mg

M �

0.0

1 m

g Fe

nt N

oted

that

rat

io m

ay

hav

e be

en to

o h

igh

sin

ce

pati

ents

rec

eivi

ng

fen

tan

yl

requ

ired

sig

nif

ican

tly

mor

e op

ioid

than

mor

phin

e at

48

hrs

.W

atan

abe

et a

l19

98 [

40]

M to

Fen

t CSC

I H

M

to F

ent C

SCI

(PO

/ SC

to C

SCI)

Ret

rosp

ecti

ve (

24 h

ours

of

stab

ility

aft

er o

pioi

d sw

itch

re

quir

ed, n

o m

ore

than

2

resc

ue d

oses

)

17 o

ut o

f 22

Can

cer

(Opi

oid

neu

roto

xici

ty p

oor

resp

onse

SC F

ent t

o SC

M 8

5.4:

1 (6

5–11

2.5:

1) S

C F

ent t

o SC

HM

23:

1 (1

0.7–

29.7

:1)

Hun

t et a

l19

99 [

41]

M to

Fen

t (C

SCI

to

CSC

I) F

ent t

o M

(CSC

I to

CSC

I)

R, D

B, X

(6

days

)23

Can

cer

pati

ents

SC F

ent:

SC M

of 6

6:1

Mep

erid

ine

(pet

hid

ine)

w

as u

sed

for

brea

kth

roug

h a

nal

gesi

a.

Th

e au

thor

s re

port

that

th

e ra

tio

of 6

6:1

appe

ared

app

ropr

iate

.

M �

mor

phin

e; H

M �

hyd

rom

orph

ine;

Fen

t �

fen

tan

yl; S

uf �

suf

enta

nyl

; CSC

I �

con

tin

uous

sub

cuta

neo

us in

fusi

on; S

C �

sub

cuta

neo

us r

oute

; PO

� o

ral r

oute

; R �

ran

dom

ized

; DB

� d

oubl

e-bl

ind;

X �

cros

sove

r; E

DR

� e

quia

nal

gesi

c do

se r

atio

.a E

xpre

ssed

as

mea

n (

ran

ge).


enced an obvious withdrawal syndrome whenswitching to SC fentanyl. Despite these differ-ences, they concluded that the conversion ra-tio they used of SC morphine 10 mg to SC fent-anyl 150 �g is adequate but cautioned thatadditional breakthrough doses may be re-quired in the days following a substitution oforal morphine to SC fentanyl.

Two retrospective studies report on the equi-analgesic dose ratios relative to subcutaneouslyadministered fentanyl by continuous infusion.38,40

Paix et al. report 11 palliative patients with cancer-related pain who, with the exception of one whowas on PO codeine and one switched frommorphine administered epidurally, were switchedfrom morphine by continuous SC infusion to fen-tanyl continuous SC infusions because of signifi-cant adverse effects and opioid-related neurotox-icity.38 They compared the mean daily dose ofmorphine and fentanyl required to give stableanalgesia (stability criteria were not clearly de-fined in the original article) before and afterthe change of drugs. The mean (�SD) fenta-nyl to morphine relative potency ratio was 68:1 �23:1 (range, 15:1–100:1). Stated in terms of anEDR, this translates to 1 mg of SC morphinebeing equianalgesic to 15 �g of SC fentanyl.Watanabe et al.40 reviewed the charts of 22 pal-liative patients with cancer-related pain whohad been switched from a variety of opioids tosc fentanyl by continuous infusion. Five of the22 patients were switched to fentanyl fromtransdermal fentanyl (due to poor pain con-trol) and 17 from other opioids (because ofopioid-related neurotoxicity). The patientswere therefore relatively unstable in terms ofpain control and the presence of opioid-relatedtoxicity. In 13 patients who were switched frommorphine or hydromorphone and reached dosestability (defined as no dose changes and nomore than 2 rescue doses per day for 48 hours),the median relative potency ratio of SC fentanylto SC morphine (n � 4 patients who stabi-lized) was 85.4:1 (range 65:1–112.5:1). Thistranslates to an EDR of 1 mg of SC morphinebeing equivalent to 11.8 �g of SC fentanyl. Therelative potency ratio of SC fentanyl to SC hydro-morphone (n � 6) was 23:1 (range 10.7:1–29.7:1). The authors noted a wide range ofdose ratios and suggested that a more cautiousfentanyl to morphine potency ratio of 100:1 maybe advisable but acknowledged that the smallsize of the study was a major limitation.

In a randomized, double-blinded study, 55postsurgical, opioid-naïve patients were random-ized to receive morphine, meperidine (pethi-dine), or fentanyl intravenously by patient-con-trolled analgesia.39 This study did not aim tospecifically determine the equianalgesic ratiosbetween the opioids, but compared opioidconsumption among the 3 groups. The re-searchers noted that (when using ratios of 1 mgmorphine � 10 mg meperidine/pethidine �0.01 mg fentanyl) equivalent amounts of opi-oid were used by the fentanyl and morphinegroups at the end of the first day, but the fenta-nyl group used significantly more than themorphine group by the end of the two-daystudy period. The authors suggested that thisdifference was related either to fentanyl havinga shorter duration of action than the other opi-oids (because of its higher metabolic clear-ance) or that the doses selected were not equi-potent. They dismissed the latter reason on thegrounds that the demand ratios (the numberof successful to total demands made) were sim-ilar for each of the drugs.

DiscussionOur systematic review of the literature re-

sulted in a number of both general and spe-cific findings. Based on these findings, we pro-pose a number of revisions in relation to thecurrent equianalgesic tables. Furthermore, wewish to highlight some important methodolog-ical issues in relation to future research in thisarea.

General FindingsThere are a surprisingly small number of

studies that have assessed opioid equianalgesicdose ratios in the context of chronic pain man-agement and repeated opioid dosing. Thosethat do exist are heterogeneous in terms ofsize, subjects, settings, specific aims, and studymethods, particularly those relating to the cal-culation of relative potency ratios. In many ofthe selected studies, the determination of anequianalgesic dose was often not a primary out-come measure and thus the study design lackedthe required elaboration to capture many ofthe multiple factors known to influence pain in-tensity and opioid efficacy. Most of the studiesinvolved small sample sizes. Given the interin-dividual variation in dose ratios, small sample


size is therefore quite a major limitation in theinterpretation of study findings. The validity ofcomparing ratios derived from postoperativesurgical patients with those from patients withadvanced cancer is also questionable. Unlikemost surgical study patients, many patientswith advanced cancer have varying degrees ofrenal impairment, hepatic impairment, differ-ential tolerance to different opioid effects andpsychosocial distress. Also, the duration of ex-posure to opioid and the number of doses var-ies across studies. Although tolerance is ahighly complex and controversial issue,42 thedegree to which it does occur is likely to varygreatly across these studies. Pain is also re-garded to be a multidimensional construct.43

The degree to which the multidimensionalityof pain is assessed and addressed across variousstudy settings and studies in similar settings islikely to be highly variable. Heterogeneity there-fore exists among these studies in many aspects, afactor that warrants recognition, particularly whenone is attempting to compare their findings.

Specific Findings

Wide ranges. Most of the studies report verywide ranges in EDR (see Tables). This reflectsthe marked observed inter- and intra-individ-ual variability among patients’ responses to dif-ferent opioids. Numerous factors contribute tothis variability, including the route of adminis-tration, a drug’s half-life, bioavailability, druginteractions, the pathophysiology of the painstate, clearance by the liver and/or kidneys, ac-cumulation of opioid metabolites, access to thereceptors and binding affinity to the receptors,to mention but a few. Other factors include apatient’s prior opioid exposure and the clinicalcontext of the switch, e.g. opioid-related neu-rotoxicity and confusion.

High potency of methadone. All the studies indi-cate a higher potency of methadone than is of-ten accepted or indicated in the equianalgesicdose tables of most standard sources. Clearly,an EDR of PO morphine to PO methadone of1:1 underestimates the potency of methadonewhen given long-term. An EDR of 3:1 or 4:1can only be relied upon for patients who aretreated with low morphine doses. Although anEDR of 10:1 is likely more appropriate whenpatients are switched from relatively higher

doses of a specific opioid to methadone, it isalso recognized that the ratios depend on thedose of the previous opioid (see below).

Ratios depend on the dose of the previous opioid.The assumption that the same relative potencyratio operates irrespective of the level of opioiddose reached prior to an opioid switch is flawed,specifically with respect to methadone. Two rel-atively large retrospective series24,29 and a pro-spective study30 have demonstrated a significantcorrelation between the dose of the previousopioid and the dose ratio between methadoneand the previous opioid. In the case of a mor-phine to methadone switch, for example, thehigher the morphine dose, the higher the mor-phine to methadone equianalgesic dose ratiowill likely be. This phenomenon has not beenwell studied in the case of other opioids. Twostudies24,36 failed to identify the phenomenonin the case of switches between morphine andhydromorphone.

Inconsistencies related to oxycodone and parenteralfentanyl. Earlier studies suggested that oxy-codone may be either slightly less potent31 orequipotent to morphine.32,35 Recent studiessuggest that PO oxycodone appears to be 1.5to 2 times relatively more potent than PO mor-phine.33,34,36 Wide ranges in EDR are howeverreported. Similarly, wide ranges are reportedin the limited studies addressing the relativepotency ratio of parenteral fentanyl relative toother opioids.38,40,41 The potency ratios of SC fen-tanyl to SC morphine reported are in the orderof 68:1 (mean) to 85.4:1 (median):1. The man-ufacturers suggest a SC fentanyl to SC mor-phine potency ratio of 100:1.

Possibility of directional inequality of cross toler-ance. The equianalgesic dose ratio for a switchfrom morphine to hydromorphone, for exam-ple, is generally considered to be the same as aswitch from hydromorphone to morphine, i.e.,a single ratio serves bidirectional switching.However, recent studies suggest that ratios mayindeed change according to the direction ofthe switch.24,26,28,36 The clinical relevance ofthis is not yet clear. Although it is unlikely tohave major impact in switches at lower doses, itis possible that failure to recognize the direc-tional difference in ratio may result in negativeconsequences when switching at higher doses.


The mechanisms underlying this phenomenonare not clear but could relate to the generationof active metabolites.26,42

Proposals for Updating Current Equianalgesic Dose Tables

Equianalgesic dose tables should displayprominent footnotes that indicate the following:

1. Relevancy for long-term dosing. Some guide-lines and textbooks quote ratios that aremore applicable to acute pain manage-ment. These tables may, for example, in-dicate hydromorphone to be up to 10times more potent that morphine.44 Sincethe ratios differ depending on whetheran opioid is being administered on a onetime basis (or only 2 or 3 doses) or multi-ple (long-term) dosing, tables should in-dicate clearly what they are intended for—short term administration versus chronic,long term administration.

2. Equinalgesic dose ratio tables are guidelines. Itis important to emphasize that equianal-gesic conversion schema are only meantto act as guidelines. Although current ta-bles can serve as useful guidelines whenswitching an opioid type or route of ad-ministration, they do not recognize thewide variation between individuals. Thebasis for the observed inter- and intra-individual variability in sensitivity to opi-oid analgesia and adverse effects is multi-factorial, complicated and still poorlyunderstood.15,42 The relative potency ofsome opioids may increase with repetitivedosing.45 The dose, therefore, needs tobe individualized and carefully titrated toeffect. Despite the variabilities in opioidconversion and responsiveness, the litera-ture does not describe many serious ad-verse events when switching between mor-phine, hydromorphone, and oxycodone.However, absence in the literature doesnot necessarily indicate absence of com-plications since such events may go unre-ported. In the case of methadone, seriousadverse events have been reported.

3. Titrate to effect. See #2 above.

4. Monitor clinically. See #2 above. Following aswitch, patients should be monitored regu-larly until stable. This entails monitoring

for reported pain intensity in addition tothe signs and symptoms of opioid neuro-toxicity (e.g., myoclonus, hallucinations,hyperalgesia, cognitive dysfunction), otheropioid adverse effects (e.g., somnolence,nausea, and constipation) and acute opi-oid overdosing. Knowledge of the opioid’selimination half-life may guide the moni-toring protocol. Methadone, for example,has a variable, long half-life and complica-tions may only manifest a day or two later.Clinical prudence is, therefore, advised.

5. Recognition of the lack of complete cross-toler-ance between opioids. The phenomenon ofincomplete cross tolerance can lead togreater than anticipated potency in a newopioid, even when from the same generalclass of opioid analgesic.42,46 When switch-ing opioids, clinicians should convert thedosage based upon consideration of arange of factors including available equi-analgesic dose data, clinical factors, con-cerns for patient safety, and incompletecross tolerance. To accommodate these, afurther decrease in the predicted equian-algesic dose by a further 30–50% is rec-ommended.

6. Caution in the setting of renal impairment: Re-nal impairment results in the accumula-tion of certain opioid metabolites, therebyincreasing the risk for opioid-related toxic-ity and potentially altering the relativeequianalgesic dose ratio.8 It is suggestedthat some opioids such as methadone andfentanyl are free of active metabolites, andtheir metabolism and elimination are in-dependent of renal function, thereby con-stituting a lower risk of developing neuro-toxicity. However, this premise needs to bestudied. Although their metabolites maybe analgesically inactive, some caution isadvised as they may perhaps exert somecentral nervous system toxicity. Morphine-3-glucuronide, an analgesically inactivemetabolite, is implicated in neurotoxicity.8

We suggest that the body of the current equianal-gesic tables be updated to reflect the following:

1. Methadone is more potent than previously ac-cepted. The higher potency of methadonerelative to other opioids and the correla-tion of the ratio with the dose of the opi-


oid prior to switching have considerableclinical implications and should be repre-sented in updated tables. Foley and Houdecomment: “The observed correlation be-tween dose ratio and previous opioid’sdose, coupled with the observed wide varia-tion in dose ratio, suggests the need for ahighly individualized approach in theprocess of reaching optimal dose of meth-adone.”15 An equianalgesic dose ratio ofPO morphine to PO methadone of 1:1 inthe management of chronic pain is inappro-priate, while equianalgesic dose ratios of 3:1or 4:1 apply only when patients are switchedfrom very low doses of morphine or otheropioids to methadone (see #2 below).

2. The ratios relative to methadone depend on thedose of the previous opioid. There are someprospective data to provide guidelines.30 Ri-pamonti and colleagues suggest the follow-ing morphine to methadone equianalgesicdose ratios: i) if the daily oral morphinedose prior to the switch is 30 mg to 90 mg,an EDR of 4:1 should be adequate; ii) if theoral morphine dose is 90 mg to 300 mg perday then an EDR of 8:1 is recommended;and iii) if the dose is greater than 300 mg oforal morphine per day, the EDR should be12:1 and even higher if the patient is onvery high doses prior to the switch.

3. Oxycodone is more potent than morphine. Anequianalgesic dose ratio of PO morphineto PO oxycodone of 1.5 to 2:1 is recom-mended as a guideline.

4. A SC fentanyl to SC morphine potency ratiorange of 80–100:1. This range is likely safeand should suffice as a starting point,with the proviso that patients are moni-tored and assessed regularly and the dosetitrated accordingly.

Methodological Issues and Future ResearchAlthough randomized controlled trials are

considered the “gold standard” of evidence,the retrospective nature of some of the studiesidentified in this review is balanced to some ex-tent in that neither physicians nor patientswere aware that the dose ratio was going to bestudied at the time of the opioid switch. None-theless, both observer and patient bias couldhave influenced the assessment of pain inten-sity level and the appearance of adverse effects.

In some of the retrospective studies, switchingtook place because of opioid-related neurotox-icity, the reason many patients undergo switch-ing. Opioid switching in patients with goodpain control usually is not necessary, hence aprospective study of opioid switching in suchpatients may have its own limitations. Further-more, prospective study of switching versusnon-switching in patients with opioid-related tox-icity may raise ethical and methodological con-cerns. However, randomized controlled trials inthe context of repeated dosing, and incorporat-ing a crossover feature, may provide useful infor-mation. To incorporate a crossover feature in theclinical context of opioid neurotoxicity, anticipa-tory consent would have to be obtained from thepatient (if cognition is intact) or alternatively aproxy (if the patient is incapable of giving con-sent). Future studies should be designed to spe-cifically clarify equianalgesic dose ratios, ratherthan use “side data” to infer these ratios.

ConclusionWhen switching opioids the goal is to achieve

optimal analgesia, hence avoiding the toxicityassociated with overdosing and the inadequatepain control that accompanies underdosing. Areview of the evidence related to equianalgesicdose and relative potency ratios in the contextof multiple opioid dosing reveals that: 1) thereexists a general paucity of data and consider-able heterogeneity in the nature of the studies;2) the ratios exhibit extremely wide ranges; 3)methadone is far more potent than previouslyappreciated; 4) the ratios related to metha-done are highly correlated with the dose of theprevious opioid; 5) the ratio may change ac-cording to the direction the opioid switch; and6) discrepancies exist with respect to both oxy-codone and fentanyl. Overall, these findingshave important clinical implications for clini-cians and warrant consideration in the poten-tial revision of current tables. The complexityof the clinical context in which many switchesoccur needs to be recognized and this needs tobe appreciated in the design of future studies.

References1. World Health Organization. Cancer pain relief

(ed 2). Geneva, Switzerland: World Health Organi-zation, 1996.


2. Cherny NJ, Chang F, Frager G, et al. Opioidpharmacotherapy in the management of cancerpain: a survey of strategies used by pain physiciansfor the selection of analgesic drugs and routes of ad-ministration. Cancer 1995;76:1283–1293.

3. Bruera E, Macmillan K, Hanson J. Palliative carein a cancer center: results in 1984 versus 1987. J PainSymptom Manage 1990;5:83–93.

4. Galer B, Coyle N, Pasternak W, Portenoy R. In-dividual variability in the response to different opio-ids: report of five cases. Pain 1992;49:87–91.

5. Coyle N, Adelhardt J, Foley KM, Portenoy RK.Character of terminal illness in the advanced cancerpatient: pain and other symptoms during the last fourweeks of life. J Pain Symptom Manage 1990;5:83–93.

6. MacDonald N, Dere L, Allan S, Champion P.Opioid hyperexcitability: the application of alter-nate opioid therapy. Pain 1993;53:353–355.

7. Maddocks I, Somogyi A, Abbott F, et al. Attenua-tion of morphine-induced delirium in palliative careby substitution with infusion of oxycodone. J PainSymptom Manage 1996;12:182–189.

8. Bruera E, Pereira J. Neuropsychiatric toxicity ofopioids. In: Proceedings of the 8th world congresson pain. Progress in pain research and manage-ment, vol 8. Jensen TS, Turner JA, Wiesenfeld-Hal-lin Z, eds. Seattle: IASP Press, 1997;717–758.

9. De Stoutz N, Bruera E, Suarez-Almazor M. Opi-oid rotation for toxicity reduction in terminal cancerpatients. J Pain Symptom Manage 1995;10:378–384.

10. Cancer pain: a monograph on the managementof cancer pain. (h42-2/5). Ottawa,Canada: Health &Welfare Canada: Minister of Supply and Services,1984.

11. Jacox A, Carr DB, Payne R, et al. Pharmacologicmanagement, clinical practice guideline: manage-ment of cancer pain. Rockville, MD: U.S. Depart-ment of Health and Human Services, 1994;39–74.

12. Hanks GWC, Cherny NI. Opioid analgesic ther-apy. In: Doyle D, Hans GWC, Macdonald N, eds.The oxford textbook of palliative medicine. Oxford:Oxford University Press, 1998;331–355.

13. Foley KM. Supportive care and quality of life. In:DeVita VT Jr, Hellman S, Rosenberg SA, eds. Can-cer: principles & practice of oncology. Philadelphia:Lippincott-Raven, 1997;2807–2841.

14. Levy MH. Pharmacologic treatment of cancerpain. N Engl J Med 1996;335:1124–1132.

15. Foley KM, Houde RW. Methadone and cancerpain management: individualize dose and titrate toeffect. J Clin Oncol 1998;16:3213–3215.

16. Zhukovsky DS, Walsh D, Doona M. The relativepotency between high dose oral oxycodone and in-travenous morphine: a case illustration. J PainSymptom Manage 1999;18:53–55.

17. Beaver WT, Wallenstein SL, Rogers A, Houde R.

Analgesic studies of codeine and oxycodone in pa-tients with cancer. I. Comparisons of oral and intra-muscular codeine and of oral and intramuscular ox-ycodone. J Pharmacol Exper Ther 1978;207:92–100.

18. Beaver WT, Wallenstein SL, Rogers A, et al. An-algesic studies of codeine and oxycodone in patientswith cancer. II. Comparisons of oral and intramus-cular codeine and of oral with intramuscular oxy-codone. J Pharmacol Exper Ther 1978;207:101–108.

19. Sawe J, Daylstrom B, Paalzow L, Rane A. Mor-phine kinetics in cancer patients. Clin Pharm Ther1981;30:629–635.

20. Houde R, Wallenstein S, Beaver W. Evaluationof analgesics in patients with cancer pain. ClinPharm 1966;1:59–97.

21. Houde R, Wallenstein S, Beaver W: Clinicalmeasurement of pain. In: de Stevens G, ed. Analge-sics. New York: Academic Press, 1965:75–122.

22. Houde RW. Systemic analgesics and relateddrugs: narcotic analgesics. In: Bonica JJ, VentafriddaV, eds. Advances in pain research and therapy, vol 2.New York: Raven Press, 1979;263–273.

23. Hanks GW, Hoskin PJ, Aherne GW, Turner P.Explanation for potency of repeated oral doses ofmorphine? Lancet 1987;2:723–724.

24. Bruera E, Pereira J, Watanabe S, Hanson J. Opi-oid rotations in patients with cancer pain: a retro-spective comparison of dose ratios between metha-done, hydromorphone and morphine. Cancer 1996;78:852–857.

25. Dunbar PJ, Chapman CR, Buckley FP, GavrinJR. Clinical analgesic equivalence for morphine andhydromorphone with prolonged PCA. Pain 1996;68:265–270.

26. Lawlor P, Turner K, Hanson J, Bruera E. Doseratio between morphine and hydromorphone in pa-tients with cancer pain: a retrospective study. Pain1997;72:79–85.

27. Miller MG, McCarthy N, Ca OB, Kearney M.Continuous subcutaneous infusion of morphine vshydromorphone: a controlled trial. J Pain SymptomManage 1999;18:9–16.

28. Lawlor P, Turner K, Hanson J, Bruera E. Doseratio between morphine and methadone in patientswith cancer pain. Cancer 1998;82:1167–1173.

29. Ripamonti C, De Conno F, Groff L, et al. Equi-analgesic dose ratio between methadone and otheropioid agonists in cancer pain: comparison of twoclinical experiences. Ann Oncol 1998;9:79–83.

30. Ripamonti C, Groff L, Brunelli C, et al. Switch-ing from morphine to oral methadone in treatingcancer pain. What is the equianalgesic dose ratio? JClin Oncol 1998;16:3216–3221.

31. Kalso E, Vainio A. Morphine and oxycodone hy-drochloride in the management of cancer pain.Clin Pharm Ther 1990;4:639–646.


32. Glare PA, Walsh D. Dose-ranging study of oxy-codone for chronic pain in advanced cancer. J ClinOncol 1993;5:973–978.

33. Heiskanen T, Kalso E. Controlled-release oxy-codone and morphine in cancer-related pain. Pain;1997;73:37–45.

34. Bruera E, Belzile M, Pituskin E, et al. Random-ized, double-blind, crossover trial comparing safetyand efficacy of oral controlled-release oxycodonewith controlled release morphine in patients withcancer pain. J Clin Oncol 1998;16:3222–3229.

35. Silvasti M, Rosenberg P, Seppala T, et al. Com-parison of analgesic efficacy of oxycodone andmorphine in postoperative intravenous patient-con-trolled analgesia. Acta Anaesth Scand 1998;42:576–580.

36. Gagnon B, Bielech M, Belzile M, et al. The useof intermittent subcutaneous injections of oxy-codone for opioid rotation in patients with cancerpain. Support Care Cancer 1999;7:265–270.

37. Leow KP, Smith MT, Williams B, Crammond T.Single-dose and steady-state pharmacokinetics andpharmacodynamics of oxycodone in patients withcancer. Clin Pharmacol Ther 1992;52:487–495.

38. Paix A, Coleman A, Lees J, et al. Subcutaneousfentanyl and sufentanil infusion substitution formorphine intolerance in cancer pain management.Pain 1995;63:263–269.

39. Woodhouse A, Hobbes A, Mather L, Gibson M.A comparison of morphine, pethidine and fentanylin the postsurgical patient-controlled analgesia envi-ronment. Pain 1996;64:115–121.

40. Watanabe S, Pereira J, Hanson J, et al. Fentanylby continuos subcutaneous infusion for the manage-ment of cancer pain. J Pain Symptom Manage 1998;16:323–326.

41. Hunt R, Fazekas B, Thorne D, Brooksbank M. Acomparison of subcutaneous morphine and fenta-nyl in hospice cancer patients. J Pain Symptom Man-age 1999;18:111–119.

42. Lawlor P, Pereira J, Bruera E. Dose ratios amongdifferent opioids: update on the use of the equianal-gesic table. In: Bruera E, Portenoy R, eds. Topics inpalliative care, volume 5. New York: Oxford Univer-sity Press 2001;247–276.

43. Lawlor PG. Multidimensional assessment: painand palliative care. In: Bruera E, Portenoy RK, eds.Cancer pain. New York: Cambridge University Press(in press).

44. Mahler DL, Forest WH. Relative analgesic po-tencies of morphine and hydromorphone in postop-erative pain. Anesthesiology 1975;42:602–607.

45. Foley K. The treatment of cancer pain. N Engl JMed 1985;313:842001 95.

46. Johnson S, Faull C. The absence of ‘cross-toler-ance’ when switching from oral morphine to trans-dermal fentanyl. Palliat Med 1997;11:4942001 495.

Documents

Equianalgesic Dose Ratios for Opioids: A Critical Review and Proposals for Long-Term Dosing