GENERAL ARTICLES
The Role of Postoperative Analgesia in Delirium and
Cognitive Decline in Elderly Patients: A Systematic Review
Harold K. Fong, MD, Laura P. Sands, PhD, and Jacqueline M. Leung,
School of Medicine, Department of Anesthesia and Perioperative Care, University of California, San Francisco, California;
Purdue University, West Lafayette, Indiana
Postoperative delirium and cognitive decline are ad-
verse events that occur frequently in elderly patients.
Preexisting patient factors, medications, and various
intraoperative and postoperative causes have been im-
plicated in the development of postoperative delirium
and cognitive decline. Despite previous studies identi-
fying postoperative pain as a risk factor, relatively few
clinical studies have compared the effect of common
postoperative pain management techniques (IV and
epidural) or opioid analgesics on postoperative cogni-
tive status. A systematic search of the PubMed and CI-
NAHL databases identified six studies comparing dif-
ferent opioid analgesics on postoperative delirium and
cognitive decline and five studies comparing IV and
D
elirium is an acute confusional state with al-
terations in attention and consciousness (1). In
contrast, cognitive decline is defined as
changes in one or more neuropsychological do-
mains and is often more subtle than delirium (2,3).
Health care professionals frequently under-estimate
cognitive decline because its determination requires
administration of neuropsychological tests. Conse-
quences of cognitive status changes include poor
functional recovery and increased morbidity (4 – 6).
Elderly patients in particular are more susceptible to
developing postoperative cognitive decline (POCD)
(7,8) and delirium (9 –12). After noncardiac surgery,
the incidence of POCD ranges from 7%–26%
(8,13,14) and the incidence of delirium ranges from
10%– 60% (11). After cardiac surgery, POCD and
Supported, in part, by institutional funds and the National Insti-
tute of Aging, National Institutes of Health, Grant #1K24
AG00948-05 awarded to Dr. Leung.
Accepted for publication October 28, 2005.
Address correspondence and reprint requests to Jacqueline M.
Leung, MD, MPH, University of California, San Francisco, Depart-
ment of Anesthesia and Perioperative Care, 521 Parnassus, San
Francisco, CA 94143-0648. Address e-mail to leungj@anesthesia.
ucsf.edu.
DOI: 10.1213/01.ane.0000198602.29716.53
©2006 by the International Anesthesia Research Society
0003-2999/06
MD, MPH
epidural routes of administering analgesia. Meperidine
was consistently associated with an increased risk of
delirium in elderly surgical patients, but the current ev-
idence has not shown a significant difference in postop-
erative delirium or cognitive decline among other more
frequently used postoperative opioids such as mor-
phine, fentanyl, or hydromorphone. The available
studies also suggest that IV or epidural techniques do
not influence cognitive function differently. However,
future investigations of sufficient study size and more
standardized methods of defining outcomes are neces-
sary to confirm the current findings.
(Anesth Analg 2006;102:1255–66)
delirium occur in 24%– 80% (2,15,16) and 3%– 47%
(2) of patients, respectively. The varied incidence
rates are likely a result of differences in study meth-
odology and patient population characteristics such
as age. These statistics will likely increase because of
an aging population and more elderly patients pre-
senting for major surgery. Given their common oc-
currence and evidence for their long-term effects on
functioning (8,17–19), POCD and delirium are thus
clinically important issues in the perioperative man-
agement of the elderly patient.
The etiology of POCD and delirium is still not
well understood. Conclusions from previous studies
have not always been in complete agreement, and
the distinction between delirium and POCD has
often been blurred. However, studies do agree that
a variety of medications can precipitate delirium
and acute cognitive decline in both medical and
surgical patients (Table 1). Among elderly surgical
patients in particular, preexisting patient factors as
well as intraoperative and postoperative causes are
also involved (Table 2). Some studies have also
highlighted important intraoperative factors that
have not been shown to increase POCD or delirium.
Anesth Analg 2006;102:1255–66 1255
1256 FONG ET AL. ANESTH ANALG
POSTOPERATIVE ANALGESIA AND COGNITIVE IMPAIRMENT 2006;102:1255–66

Table 1. Medications Associated With Delirium and Cognitive Decline

Opioid analgesics (6,9,33) Meperidine, fentanyl, morphine, hydromorphone
Sedative-hypnotics (6,23,33,52) Benzodiazepines, barbiturates
Antihistamines (53,54) Diphenhydramine, hydroxyzine, chlorpheniramine
Nonsteroidal antiinflammatory drugs (55–57) Naproxen, ibuprofen, indomethacin
Drugs affecting cholinergic transmission in the CNS:
Anticholinergics (5,6,23,58) Atropine, scopolamine
Antiparkinsonian agents (58,59) Benztropine, trihexyphenidyl, levodopa
Neuroleptics (9,60) Clozapine, thioridazine, chlorpromazine
Tricyclic antidepressants (58,60) Amitryptiline, imipramine
Class 1A antiarrhythmics (61,62) Disopyramide, quinidine, procainamide
Other cardiac medications (6,23,58) Digoxin, beta-adrenergic antagonists, methyldopa
Gastrointestinal H2-antagonists (63,64) Cimetidine, ranitidine
CNS ⫽ central nervous system.

Table 2. Nonpharmacologic Risk Factors of Postoperative Cognitive Decline or Delirium

Patient-relatedrisk factors
Prior cognitive impairment
(7)
Intraoperative factors
Postoperative factors
Nonsignificant factors
Postoperative cognitive decline
Increasing age (7,8)
Prior cognitive impairment
(4,5,12,32,65)
Depression (7)
Low education level (not
completed high school) (7,8)
Cardiac surgery (69–71)
Longer duration of anesthesia (8) Noncardiac thoracic surgery (12)
Pain (27,73)
Respiratory complications (8)
Infections (8)
Second operation within 1 wk (8) Electrolyte and metabolic derangements (74)
General vs. regional anesthesia
(3,14,20,21)
Postoperative delirium
Increasing age (especially ⬎70 yr) (4,5,12,65)
Depression (5,66,67)
Poor functional status (4,12,65)
Abnormal blood pressure (32)
Heart failure on admission (32)
Abnormal serum electrolytes, or glucose (12,68)
Self-reported alcohol abuse (10,12,65)
Narcotic use before admission (65)
Cardiac surgery (72)
Aortic aneurysm surgery (12)
Bilateral versus unilateral knee replacement (10)
Blood loss (24)
Pain (9,25,32)
Hematocrit ⬍ 30% (24)
Blood transfusions (24)
General vs. regional anesthesia (22–24)

Perioperative hypoxemiaor Intraoperative hemodynamic complications suchas

hypotension (8)* hypotension, bradycardia, tachycardia (24)

*Small studies showed an association between hypoxemia and postoperative confusion (75,76).

For example, comparisons between general and
regional anesthesia as potential risk factors for
POCD (14,20,21) and delirium (22–24) have not re-
vealed a significant difference between the two tech-
niques. A systematic review of intraoperative anes-
thesia techniques on POCD has also not shown a
difference (3).
Among the multitude of factors associated with
POCD and delirium, are there management options
that can reduce the incidence of these events? Because
preoperative patient-related factors are not modifi-
able, intraoperative anesthetic types and management
have not been shown to influence outcomes, and the
postoperative period has received little attention in
previous studies, we chose to investigate management
strategies pertaining to the postoperative setting. This
review will focus on the use of opioids and postoper-
ative analgesic modalities. The rationale for choosing
these two areas is several-fold. First, opioids are the
drugs most commonly given to patients after major
ANESTH ANALG
2006;102:1255–66
surgery. Second, few reports have investigated the
role of postoperative analgesia in cognitive decline
and delirium. Furthermore, because postoperative
pain has been shown to be associated with delirium
(25,26) and cognitive decline (27), and opioids rou-
tinely used to treat postoperative pain can have cen-
tral effects, it is of particular interest to ascertain if
there are any analgesic drugs or methods that reduce
patients’ risk for these events. Accordingly, we con-
ducted a literature search to identify studies that com-
pared: 1) the effects of different opioid analgesics and
2) IV versus epidural modes of analgesia on POCD or
delirium.
Methods
A broad search of the PubMed database of the Na-
tional Library of Medicine and CINAHL database
from 1966 to 2005 was conducted using the following
MESH terms: “Pain, postoperative OR analgesia, epidural
OR analgesia, patient-controlled OR analgesics, non-
narcotic OR analgesics, opioid OR anesthesia, epidural OR
anesthesia, spinal” AND “cognition OR cognition disor-
ders OR confusion OR delirium OR memory/drug effects
OR mental processes/drug effects OR psychological tests.”
Although our focus was on opioids, non-opioid anal-
gesics were included in the search so that we could
assess the overall available evidence for postoperative
analgesia. We also looked for studies primarily com-
paring the effect of type of anesthesia (general, epi-
dural, or spinal) on cognitive decline or delirium, as
the intraoperative techniques may continue into the
postoperative period for analgesia, thus allowing a
secondary comparison of epidural or spinal and IV
methods. We also searched for articles not indexed for
Medline, using the “pubmednotmedline” qualifier in
the search command. All references within relevant
papers were further investigated for additional
studies.
For this systematic review, clinical trials and ob-
servational (cohort and case-control) studies were
included. Studies must have compared different
modes of analgesia delivery (IV versus epidural or
spinal) or types of opioid analgesics in the postop-
erative period for an association with cognitive de-
cline or delirium and defined methods to assess
cognitive function or delirium. Studies were in-
cluded even if POCD or delirium was not the pri-
mary outcome of interest.
Eight-hundred-twenty-one articles were re-
trieved. Studies that did not assess delirium or cog-
nitive decline during the postoperative period were
excluded (724 articles). Studies that investigated
postoperative delirium or cognitive decline but did
FONG ET AL. 1257
POSTOPERATIVE ANALGESIA AND COGNITIVE IMPAIRMENT
Table 3. Assessing Quality of Randomized Clinical Trials
Study described as randomized ⫹1 point
Method of randomization was described ⫹1 point
and was appropriate (table of random
numbers, computer generated, etc.)
Method of randomization was described -1 point
and was inappropriate (patients allocated
alternately, hospital number, etc.)
Study described as double-blind ⫹1 point
Method of double-blinding was described ⫹1 point
and was appropriate (identical placebo,
active placebo, dummy, etc.)
Method of double-blinding was described -1 point
and was inappropriate (e.g. comparison
of tablet versus injection with no double
dummy)
Description of withdrawals and dropouts ⫹1 point
was given
Adapted from (28).
not compare different opioid analgesic drugs or
modalities were also excluded (85 articles). Because
this review focuses on evaluating potential differ-
ences in contemporary postoperative analgesia tech-
niques, older studies that compared IV versus IM or
subcutaneous analgesia techniques were not in-
cluded (2 articles).
The first reviewer (HKF) screened all abstracts per
inclusion and exclusion criteria. A second reviewer
(JML) validated all the entry criteria for the remain-
ing 10 studies before the data extraction process and
also validated any articles to be excluded when
uncertainty arose. Per established guidelines on
study design, these studies were stratified into lev-
els of evidence that ranged from I to IV where level
I: systematic review of randomized controlled trials,
level II: randomized controlled trials, level III: non-
randomized controlled trials or from cohort or case-
control analytical studies, and level IV: expert opin-
ion. Furthermore, randomized controlled trials were
assigned a quality score on a 1–5 scale (Table 3) (28).
For each study included in this report, the following
information was recorded: study sample character-
istics (number of subjects, type of surgery, age range
of subjects), interventions being compared, criteria
for defining POCD or delirium, methods of neuro-
psychological assessment, methods of measuring
pain relief between groups, outcomes or findings
related to the primary research question, and any
statistical information given in the studies. Where
possible, study power was calculated by comparing
the proportion of the group of patients with the
higher rate of delirium against a hypothesized de-
crease of 10% given the sample sizes presented in
the study.
1258 FONG ET AL. ANESTH ANALG
POSTOPERATIVE ANALGESIA AND COGNITIVE IMPAIRMENT 2006;102:1255–66

Table 4. Studies Comparing Type of Analgesic

Outcome
measures and
Study type period of
Type of Age of study and quality cognitive data
Study N surgery subjects score if RCT Predictors collection Outcome Conclusions
Herrick et al., 96 Hip or knee ⱖ65 yr RCT, 2 Morphine Confusion 7 of 49 patients in Confusion /
1996 (29) (PCA) (delirium) morphine group delirium: no
Fentamyl bypresence of developed confusion significant
(PCA) disorientation, 2 of 47 patients in difference
hallucination, fentanyl group
inappropriate developed confusion
behavior Morphine group with
Cognitive greater decrease in
function by: cognitive performance
(1) MMSE on POD1 per MMSE
2) SPMSQ scores, but fentanyl
Preop, then group with greater
POD 1-5 decrease on PODS per
Pain by VAS SPMSQ scores
Pain scores not
significantly different

Rapp et al., 61 Abdominal Mean 40 yr for RCT, 4 Morphine “Cognitive Hydromorphone No significant
1996 (30) morphine (PCA) changes” group with greater difference
group Hydromorphone defined by decrease in mean
Mean 48 yr (PCA) 1) DSST scores in DSST
for hydro- 2) Trail B (nonsignificant) and
morphone 3) POMS Trail B (significant)
group Preop, then compared to the
POD 1–2 morphine group
Pain by 0–10 after surgery
score Did not provide
number of patients
with cognitive
changes
Pain scores not
significantly different

Silvasti et al., 43 Microvascular Mean, 51 yr RCT, 5 Morphine Cognitive Only 14 of 18 No significant

2000 (31) breast recon- (PCA) function by tramadol patients, difference
struction Tramadol 1) DSST and 17 of 25
(PCA) Preop, then morphine patients
POD 1–2 completed test.
Pain by VAS Dropouts due to
and 0–10 sedation or visual
verbal rating difficulties.
scale Mean changes in DSST
scores between the
two groups not
significantly different
Pain scores not
significantly different

Adunsky et al., 92 Hip 65-97 yr Retrospective Morphine Delirium by 13 of 38 patients in Meperidine

2002 (34) Mean 81.9 yr cohort study Meperidine CAM morphine group with signif-
for meperidine MMSE scores developed delirium icantly greater
group also obtained 19 of 44 patients incidence of
Mean 82.8 yr for Preop, then in meperidine delirium
morphine group POD 1-10 group MMSE scores
Did not developed were not
comment on delirium significantly
pain relief different
ANESTH ANALG FONG ET AL. 1259
2006;102:1255–66 POSTOPERATIVE ANALGESIA AND COGNITIVE IMPAIRMENT

Table 4. Continued
Outcome
measures and
Study type period of
Type of Age of study and quality cognitive data
Study N surgery subjects score if RCT Predictors collection Outcome Conclusions

Morrison et al., 541 Hip ⬍70 yr (N ⫽ 49) Prospective Meperidine vs. Delirium Meperidine Meperidine with
2003 (32) 70–79 yr (N ⫽ 141) cohort morphine and defined by versus significantly
ⱖ80 yr (N ⫽ 351) study “other opioids” 1) CAM other opioid greater
2) chart (RR 2.4, CI incidence of
review 1.3–4.5, delirium
Preop, then P 0.004)
POD 1-
discharge
Did not
comment on
pain relief

Marcantonio et 91 cases; General, ⱖ50 yr; mean, Case-control Meperidine Delirium Meperidine use Meperidine with
al., 1994 (33) 154 gynecologic, 73 yr study Morphine defined by (OR 2.7, CI significantly
controls and Fentanyl 1) CAM 1.3–5.5) greater incidence
orthopedic Oxycodone 2) chart Morphine of delirium
surgeries Codeine review (OR 1.2, when all five
Also compared: Preop, then CI 0.6–2.4) drugs compared
meperidine POD 1–5 Fentanyl Meperidine with
(epidural) to Did not (OR 1.5, significantly
fentanyl comment on CI 0.6–4.2) greater incidence
epidural), and pain relief Epidural of delirium
meperidine meperidine compared to
(PCA) to (OR 24, CI fentanyl within
morphine (PCA) 1.3–44) the epidural
versus analgesia group
epidural Meperidine with
fentanyl (OR greater incidence
0.9 CI 0.3–2.7) of delirium
PCa merperidine compared to
(OR 2.1, CI morphine within
0.4–10.7) intravenous
versus PCA analgesia group,
morphine but difference
(OR, 0.9, was not
CI 0.4–1.9) significant

N ⫽ number of study subjects; RCT ⫽ randomized controlled trial; MMSE ⫽ Mini Mental Status Examination; SPMSQ ⫽ Short Portable Mental Status
Questionnaire; Preop ⫽ preoperative; POD ⫽ postoperative day; DSST ⫽ Digit Symbol Substitution Test; Trail B ⫽ Trail Making B Test; POMS ⫽ Profile of Mood
States; CAM ⫽ Confusion Assessment Method; PCA ⫽ patient-controlled analgesia; RR ⫽ relative risk; OR ⫽ odds ratio; CI ⫽ confidence interval; VAS ⫽ visual
analog scale.

Results From the literature search, 3 randomized trials

Studies Comparing Different Opioid Analgesics
in the Postoperative Period
Our search of the literature demonstrates that there
has not been a systematic review of randomized con-
trolled trials (level I evidence) comparing different
analgesic drugs and methods for an association with
postoperative delirium and POCD. From existing
studies, the best evidence is provided by randomized
controlled trials (level II evidence).
(29 –31) and 3 observational studies (32–34) (level III
evidence) compared different opioid analgesics (Ta-
ble 4). With 2 exceptions, studies enrolled mainly
elderly patients. Three of these studies investigated
postoperative delirium, 2 assessed POCD, and 1
examined both. For all studies, assessments of cog-
nitive status were conducted for at least 2 days into
the postoperative period or until patient discharge.
Comparisons of analgesic drugs used as well as the
methods of defining POCD or delirium were quite
1260 FONG ET AL.
POSTOPERATIVE ANALGESIA AND COGNITIVE IMPAIRMENT
heterogeneous across the 5 studies. Given these lim-
itations, particularly the nonstandardized method
of defining outcomes, a meta-analysis of available
data could not be conducted.
Three randomized controlled trials compared opi-
oid analgesics in the postoperative period, with scores
ranging from 2 to 5 on the quality scale. In the first
study, Herrick et al. (29) compared morphine with
fentanyl in a group of 96 patients undergoing hip or
knee arthroplasty. Pain relief was not significantly
different between groups. Although 14.3% of patients
receiving morphine developed confusion, compared
with 4.3% of patients receiving fentanyl, this was
not statistically significant. The sample size limited
the power of the study to only 0.26. The researchers
also investigated cognitive decline using the Mini
Mental Status Examination (MMSE) (35) and the
Short Portable Mental Status Questionnaire
(SPMSQ) (36). The MMSE is the most widely used
screening tool for cognitive disorders in clinical and
epidemiological settings. The SPMSQ is also a com-
monly used sensitive and specific screening test of
cognitive impairment, but it is less comprehensive
and does not assess language and motor functions.
Patients who received morphine had a significantly
larger reduction in MMSE scores than did patients
who received fentanyl on postoperative day 1. How-
ever, this association was reversed for the SPMSQ
results because patients who received fentanyl per-
formed significantly more poorly on postoperative
day 5. The second study by Rapp et al. (30) com-
pared the effects of morphine versus hydromor-
phone on the development of POCD in 61 patients
undergoing lower abdominal surgery. Pain scores
were not different between groups. The group
treated with hydromorphone performed signifi-
cantly more poorly on the Trail Making B Test but
not on the Digit Symbol Substitution Test (37,38).
The third study by Silvasti et al. (31) compared
morphine with tramadol in 43 women undergoing
breast reconstruction, in which pain relief was not
significantly different between the two groups. No
significant difference in performance on the Digit
Symbol Substitution Test was found between
groups. The latter two studies did not provide the
statistics needed to calculate power.
The other three studies with level III evidence com-
pared meperidine to one or more other opioid drugs,
and all showed that meperidine was associated with a
significantly more frequent incidence of postoperative
delirium. However, whether there is a difference in
the efficacy of pain relief between groups cannot be
ascertained from the presented data. The first study by
Adunsky et al. (34) compared morphine and meperi-
dine in a group of 92 patients undergoing hip surgery.
A similar comparison of morphine and meperidine
ANESTH ANALG
2006;102:1255–66
was made in the second study by Morrison et al. (32),
who studied 541 elderly patients with hip fracture.
The third investigation is a case-control study by Mar-
cantonio et al. (33) that compared meperidine, mor-
phine, fentanyl, oxycodone, and codeine, although
cases and controls mainly received morphine and me-
peridine. Of the five medications, meperidine was the
only one associated with a significantly more frequent
incidence of postoperative delirium. In all three stud-
ies, delirium was identified by the Confusion Assess-
ment Method (CAM) (39), a commonly used screening
tool for acute confusional states adapted from the
Diagnostic and Statistical Manual of Mental Disorders
(40). One study (34) also used the MMSE in addition to
using CAM, but MMSE scores were not significantly
different between the two groups.
Studies Comparing Mode of Postoperative
Analgesia Delivery
The literature search retrieved 4 randomized trials
(level II evidence) and 1 case-control study (level III)
that compared IV and epidural analgesia in the post-
operative period (Table 5) (10,20,33,41,42). Study pa-
tients were mainly older subjects. Three studies iden-
tified delirium as the outcome, and 2 studies
investigated cognitive impairment. Similar to the
studies comparing analgesic drugs, some of these in-
vestigations were limited by small sample sizes and
heterogeneous criteria to define POCD or delirium.
In all 4 randomized controlled trials, no significant
difference was found between IV and epidural anal-
gesia with respect to delirium (10,41) or POCD (20,42).
The studies’ quality scores ranged from 2 to 4. Only
one trial was double-blind. The study by Mann et al.
(41) compared IV patient-controlled analgesia with
morphine to patient-controlled epidural analgesia
with bupivacaine and sufentanil in 70 patients under-
going major abdominal surgery. Delirium was defined
by DSM criteria and an Abbreviated Mental Test (43),
and incidence rates were similar between the IV (24%)
and epidural (26%) groups despite significantly better
analgesia in patients receiving epidurals. Williams-
Russo et al. (10) compared IV fentanyl and epidural
fentanyl and bupivacaine in 51 patients undergoing
bilateral knee replacement. Pain scores were not sig-
nificantly different between the two groups. Using
DSM criteria, the overall incidence of delirium was
41%, with no significant difference between the epi-
dural (38%) and IV (44%) groups. An earlier study by
Riis et al. (20) primarily investigated the effects of the
type of anesthesia (general versus epidural) on ‘intel-
lectual impairment,‘ or cognitive decline, among 30
patients having hip replacement. Study subjects who
ANESTH ANALG FONG ET AL. 1261
2006;102:1255–66 POSTOPERATIVE ANALGESIA AND COGNITIVE IMPAIRMENT

Table 5. Studies Comparing Mode of Analgesia (Intravenous or Epidural)

Outcome
measures and
Type of study period of
Type of Age of study and quality cognitive data
Study N surgery subjects score if RCT Predictors collection Outcome Findings
Mann et al., 70 Abdominal ⱖ70 yr RCT, 3 Intravenous Delirium by 8 of 33 patients in No significant
2000 (41) morphine 1) AMT intravenous group difference
(PCA) 2) DSM-III developed delirium between
Epidural criteria 8 of 31 patients in intravenous
bupivacaine Diagnosis of epidural group and epidural
and sufentanil postoperative developed delirium analgesia
(PCEA) delirium Patients receiving
required DSM-III epidural analgesia
criteria and had significantly
decrease in AMT better pain relief
score of 2 or
more points
Preop, then POD
1-discharge
Pain by VAS

Eriksson-Mjoberg 40 Hysterectomy Mean, 46 yr RCT, 4 Intravenous Cognitive function Only 6 of 20 patients No significant
et al. morphine ⫹ by in intravenous group difference
1997 (42) patient- 1) grammatical and 11 of 20 patients
controlled reasoning test in epidural group
intravenous 2) letter completed tests, due
morphine cancellation to lethargy in large
Epidural test number of subjects.
morphine ⫹ Preop, then 2, 8, Patients receiving
patient- 18 hours epidural analgesia
controlled postoperatively had significantly
intravenous Pain by VAS better pain relief
morphine
(patient-
controlled
intravenous
morphine
available to
both groups
as needed)

Williams-Russo 51 Bilateral knee 48–80 yr; mean, RCT, 3 Intravenous Delirium by 11 of 25 patients in No significant
et al., replacement 68 yr fentanyl 1) DSM-III intravenous group difference
1992 (10) Epidural criteria developed delirium
bupivacaine Preop, then POD 10 of 26 patients in
and fentanyl 1-7 epidural group
Pain by VAS developed delirium
epidural vs. IV:OR
0.87, CI 0.45-169
Pain relief not signi-
ficantly different

received intraoperative epidural anesthesia were also
given postoperative epidural analgesia, and, pre-
sumptively, patients who received general anesthe-
sia during surgery were given postoperative opioids
parenterally. No significant difference in the inci-
dence of POCD was observed between the two
groups. Data on pain relief were not available from
this study. We determined that the power of these 3
trials to be 0.10 (41), 0.06 (10), and 0.04 (20), respec-
tively. Finally, a study by Eriksson-Mjoberg et al.
(42) comparing the effects of epidural to IV mor-
phine on cognitive function among 40 patients un-
dergoing hysterectomy did not find a significant
difference between groups, although patients re-
ceiving epidural morphine had significantly lower
pain scores. The study did not include the necessary
1262 FONG ET AL. ANESTH ANALG
POSTOPERATIVE ANALGESIA AND COGNITIVE IMPAIRMENT 2006;102:1255–66

Table 5. Continued
Outcome
Type of measures and
study and period of
Age of study quality score cognitive data
Study N Type of surgery subjects if RCT Predictors collection Outcome Findings

Riis et al., 30 Hip replacement ⱖ60 yr RCT, 2 Patients initially Intellectual 5 of 20 patients No significant
1983 (20) randomized impairment in epidural difference
to 3 different (cognitive group had
anesthetic decline) by intellectual
regimens for 1) Trail A and impairment
surgery B 3 of 10 patients
(general 2) DSST in non-
anesthesia 3) other epidural
{G}, epidural memory group had
anesthesia and intellectual
{E}, or abstraction impairment
combined tests
general- Preop, then
epidural POD
{E⫹G}). 2-7
Postoperatively, Did not
E and E⫹G comment on
groups pain relief
received
epidural
analgesia and
G group
received
“opiate drugs
for pain
relief”
Specific an-
algesics were
not named

Marcantonio 91 cases, General, ⱖ50 yr; mean, Case-control Intravenous Delirium by PCA group Epidural with
et al., 154 gynecologic, 73 yr study (PCA) 1) CAM OR 2.1, CI significantly
1994 (33) controls and analgesia 2) chart review 0.4-10.7 greater
orthopedic (either Preop, then POD Epidural incidence
surgeries meperidine or 1-5 group of deli-
morphine) Did not OR 24, CI rium, How-
Epidural comment on 13-4.4 ever, mere-
analgesia pain relief pidine ac-
(either counted for
meperidine or 85% of
fentanyl) epidural
analgesia,
and
multivariate
analysis
not
performed

RCT ⫽ randomized controlled trial; AMT ⫽ Abbreviated Mental Test; DSM ⫽ Diagnostic and Statistical Manual of Mental Disorders; DSST ⫽ Digit Symbol
Substitution Test; Trail A and B ⫽ Trail Making A Test and Trail Making B Test; CAM ⫽ Confusion Assessment Method; PCA – patient-controlled analgesia;
PCEA ⫽ patient-controlled epidural analgesia; Preop ⫽ preoperative; POD ⫽ postoperative day; OR ⫽ odds ratio; CI ⫽ confidence interval; VAS ⫽ visual analog
scale.

statistics to allow calculation of study power. Fur-
ther, a large percentage of patients did not complete
neuropsychological testing in the last study.
Marcantonio et al.’s case-control study (33) was the
only observational study that compared epidural and
IV methods. In this investigation, the epidural route
was associated with a significant increase in delirium.
However, the researchers explain that meperidine was
the medication used in 85% of the epidurals, while the
remaining 15% consisted of fentanyl. In this study,
meperidine was associated with a more frequent inci-
dence of postoperative delirium when compared with
ANESTH ANALG
2006;102:1255–66
either morphine or fentanyl. The authors did not per-
form multivariate analysis to determine the relative
importance of medication type versus route of drug
delivery on the occurrence of postoperative delirium.
A comparison of pain relief between groups was also
not conducted.
Discussion
Studies Comparing Different Opioid Analgesics
in the Postoperative Period
Our review focuses on opioid analgesics administered
to patients postoperatively because these are the drugs
most commonly given to patients after major surgery.
Furthermore, elderly patients are especially suscepti-
ble to adverse drug effects resulting from the use of
multiple drugs, drug-drug interactions, reduced renal
clearance, and possible impaired hepatic function.
Opioid analgesics commonly used in the postoper-
ative period include morphine, fentanyl, and hydro-
morphone. The 3-glucuronide metabolites of both
morphine and hydromorphone can lead to neuroexci-
tation, whereas fentanyl undergoes hepatic biotrans-
formation to inactive metabolites. Patients with renal
insufficiency may have impaired elimination of active
metabolites, resulting in undesirable side effects such
as delirium (44 – 46). Morphine also undergoes hepatic
conjugation to form morphine-6-glucuronide, an opioid
agonist. Although some studies have examined the rel-
ative analgesic efficacy and adverse effects profile of
morphine compared to morphine-6-glucuronide, none
has evaluated their impact on postoperative delirium
(47). Meperidine, a now infrequently used opioid anal-
gesic, has a relatively long half-life and its metabolite,
normeperidine, is a central nervous system stimulant
with anticholinergic properties that may induce seizures
and delirium (48,49). Accumulation of normeperidine
will occur in patients with renal dysfunction. Meperi-
dine is sometimes used in a smaller dosage to treat
postoperative shivering. The effect of this smaller dose of
meperidine on delirium and cognition has not been
investigated.
Of the six studies that compared specific opioid
analgesics, all three studies that compared meperi-
dine with another opioid confirmed that meperidine
is associated with a significantly increased risk of
delirium in elderly surgical patients. However,
there is a significant limitation in these three studies
because they did not discuss whether analgesia was
similar between the groups of patients. Of the re-
maining studies that investigated other opioids, one
did not show a significant difference between fent-
anyl and morphine on the occurrence of postopera-
tive confusion, while the other showed that patients
FONG ET AL. 1263
POSTOPERATIVE ANALGESIA AND COGNITIVE IMPAIRMENT
receiving hydromorphone performed more poorly
on tests of cognitive function compared with those
who received morphine, but only one of the two
tests used showed a significant difference. In both of
these studies, pain relief between the two patient
groups was not significantly different.
Despite the differences in pharmacokinetics
among the different opioid analgesics, there are no
convincing data to indicate which opioids, exclud-
ing meperidine, increase the risk of POCD or delir-
ium.
The lack of a standardized approach to measuring
delirium and cognitive decline presents an addi-
tional difficulty in comparing results across studies.
Cognitive decline involves changes in several neu-
ropsychological domains, including memory, exec-
utive function, perceptual organization, language,
psychomotor function, attention, and concentration
(3,50). Among the three randomized controlled tri-
als reviewed (Table 4), the cognitive tests included
the MMSE, the SPMSQ, the Digit Symbol Substitu-
tion Test, and the Trail Making B Test. However,
these tests do not all assess the same cognitive do-
mains (3) and they may not be equally sensitive in
detecting changes in cognition or differences be-
tween study groups. For example, the MMSE and
SPMSQ are considered screens for the presence of
dementia, as used in Herrick et al.’s study (29), but
they are not as sensitive as other cognitive tests for
detecting the cognitive effects of drugs. This differ-
ence in test sensitivity could explain why test results
were discordant when more than one neuropsycho-
logical assessment was used in the same study.
Studies Comparing Mode of Analgesia Delivery
The mode of analgesia delivery is also an important
consideration in managing postoperative pain. For
acute postoperative pain control, current widely used
methods involve IV or epidural modalities, with or
without a patient-controlled device. IV and epidural
methods differ in their efficacy in relieving pain and
profile of drug-related effects (51).
Mechanistically, epidural analgesia allows opioids
to be delivered in close proximity to opioid receptors
on the spinal cord, resulting in a smaller dosage of
opioid used than IV administration to reach a similar
level of pain relief. Theoretically, by having less of a
systemic effect than parenteral analgesia, epidurals
should be associated with a decreased incidence of
postoperative delirium or POCD. The site-specific na-
ture of epidural drug delivery also enables local anes-
thetics to be given epidurally and therefore allows an
opioid-sparing effect.
Despite the beneficial effects of epidural analgesia,
the four randomized trials in this review did not find
1264 FONG ET AL.
POSTOPERATIVE ANALGESIA AND COGNITIVE IMPAIRMENT
a significant difference between epidural and paren-
teral analgesia with respect to postoperative delirium
or cognitive decline. However, the small number of
available studies and study power (3%– 4%) signifi-
cantly limits the validity of this conclusion. In addi-
tion, there are variations in the spinal level at which an
epidural catheter may be inserted (thoracic or lumbar)
as well as the medication used (local anesthetic only,
opioid only, or both). These options further reinforce
the necessity for more comprehensive clinical trials
comparing epidural and IV analgesia.
Only 3 of the 5 studies compared pain relief be-
tween groups. The studies by Mann et al. (41) and
Eriksson-Mjoberg et al. (42) that found significantly
less pain in patients receiving an epidural did not find
a concurrent improvement in rates of delirium. A low
study power in the Mann et al. study may have pro-
hibited such a finding, whereas study power could not
be ascertained from the investigation by Eriksson-
Mjoberg et al.
The few available studies also differ methodologi-
cally in the assessment of delirium. The different tests
used included DSM-III criteria plus Abbreviated Men-
tal Test (41), the CAM (33), and DSM III criteria only
(10).
There are further limitations in the comparisons of
analgesia techniques. One study (41) used morphine
for the IV group but used bupivacaine with sufentanil
for the epidural patients. The authors noted that be-
cause sufentanil is a highly lipid soluble opioid, a
potential reduction in systemic effect via the epidural
route may not have been observed. The Riis et al. (20)
study did not provide the specific analgesics used, and
in the Marcantonio et al. (24) study, meperidine, an
analgesic with a high deliriogenic potential, was used
in patients who received epidural analgesia.
Implications for Clinical Practice and Future
Directions for Research
In addition to the nonpharmacological risk factors that
have been shown to influence POCD or delirium (Ta-
ble 2), our review is the first to systematically evaluate
the literature on whether management of postopera-
tive pain is important to reducing the risk of develop-
ing POCD and delirium. Our review demonstrates
that this important postoperative period has been
largely ignored in previous research on postoperative
cognitive assessment, as there are only a few clinical
trials that compare opioid analgesics or analgesic mo-
dalities. Thus far, meperidine appears to be the only
drug that should clearly be avoided based on results
from existing studies. There is insufficient evidence to
discern significant differences among morphine, fent-
anyl, hydromorphone or other drugs. Similarly, the
ANESTH ANALG
2006;102:1255–66
data did not demonstrate any difference on postoper-
ative cognitive outcomes between IV versus epidural
methods of postoperative analgesia.
It must be reiterated that prior studies have not
always made a clear distinction between delirium and
cognitive decline and have often erroneously used
these terms interchangeably. Diagnostic criteria differ
for these two entities, as do predictive factors and
potential prognostic significance. Only by using ap-
propriate and uniform outcome measures will results
be collectively useful. This review highlights the im-
portance that future investigations should use more
standardized measures of cognitive decline and delir-
ium to allow proper comparison between studies, use
neuropsychological tests to target domains sensitive to
drug effects, and, most importantly, enroll a sufficient
sample size to adequately discern a difference in effect
size. Finally, because both pain relief and neuroexci-
tation may be induced by opioids, which may influ-
ence delirium in opposite directions, future studies on
postoperative delirium and POCD must include the
assessment of postoperative pain.
References
1. Lipowski Z. Delirium (acute confusional states). JAMA 1987;
258:1789–92.
2. Bekker AY, Weeks EJ. Cognitive function after anaesthesia in
the elderly. Best Pract Res Clin Anaesthesiol 2003;17:259–72.
3. Wu CL, Hsu W, Richman JM, Raja SN. Postoperative cognitive
function as an outcome of regional anesthesia and analgesia.
Reg Anesth Pain Med 2004;29:257–68.
4. Marcantonio ER, Flacker JM, Michaels M, Resnick NM. Delir-
ium is independently associated with poor functional recovery
after hip fracture. J Am Geriatr Soc 2000;48:618–24.
5. Gustafson Y, Berggren D, Brannstrom B, et al. Acute confusional
states in elderly patients treated for femoral neck fracture. J Am
Geriatr Soc 1988;36:525–30.
6. Francis J, Martin D, Kapoor WN. A prospective study of delir-
ium in hospitalized elderly. JAMA 1990;263:1097–101.
7. Ancelin ML, de Roquefeuil G, Ledesert B, et al. Exposure to
anaesthetic agents, cognitive functioning and depressive symp-
tomatology in the elderly. Br J Psychiatry 2001;178:360–6.
8. Moller J, Cluitmans P, Rasmussen L, et al. Long-term postoper-
ative cognitive dysfunction in the elderly: ISPOCD1 study. Lan-
cet 1998;351:857–61.
9. Schor JD, Levkoff SE, Lipsitz LA, et al. Risk factors for delirium
in hospitalized elderly. JAMA 1992;267:827–31.
10. Williams-Russo P, Urquhart B, Sharrock N, Charlson M. Post-
operative delirium: predictors and prognosis in elderly ortho-
pedic patients. J Am Geriatr Soc 1992;40:759–67.
11. Parikh S, Chung C. Postoperative delirium in the elderly.
Anesth Analg 1995;80:1223–32.
12. Marcantonio E, Goldman L, Mangione C, et al. A clinical pre-
diction rule for delirium after elective noncardiac surgery.
JAMA 1994;271:134–9.
13. Canet J, Raeder J, Rasmussen LS, et al. Cognitive dysfunction
after minor surgery in the elderly. Acta Anaesthesiol Scand
2003;47:1204–10.
14. Rasmussen LS, Johnson T, Kuipers HM, et al. Does anaesthesia
cause postoperative cognitive dysfunction? A randomised
study of regional versus general anaesthesia in 438 elderly
patients. Acta Anaesthesiol Scand 2003;47:260–6.
ANESTH ANALG
2006;102:1255–66
15. Rasmussen LS, Larsen K, Houx P, et al. The assessment of
postoperative cognitive function. Acta Anaesthesiol Scand 2001;
45:275–89.
16. Dyer C, Ashton C, Teasdale T. Postoperative delirium: a review
of 80 primary data-collection studies. Arch Intern Med 1995;155:
461–5.
17. Johnson T, Monk T, Rasmussen LS, et al. Postoperative cogni-
tive dysfunction in middle-aged patients. Anesthesiology 2002;
96:1351–7.
18. Rohan D, Buggy DJ, Crowley S, et al. Increased incidence of
postoperative cognitive dysfunction 24 hr after minor surgery in
the elderly. Can J Anaesth 2005;52:137–42.
19. Murray AM, Levkoff SE, Wetle TT, et al. Acute delirium and
functional decline in the hospitalized elderly patient. J Gerontol
1993;48:M181–6.
20. Riis J, Lomholt B, Haxholdt O, et al. Immediate and long-term
mental recovery from general versus epidural anesthesia in
elderly patients. Acta Anaesthesiol Scand 1983;27:44–9.
21. Williams-Russo P, Sharrock N, Mattis S, et al. Cognitive effects
after epidural vs general anesthesia in older adults. JAMA 1995;
274:44–50.
22. Urwin SC, Parker MJ, Griffiths R. General versus regional an-
aesthesia for hip fracture surgery: a meta-analysis of random-
ized trials. Br J Anaesth 2000;84:450–5.
23. Rogers MP, Liang MH, Daltroy LH, et al. Delirium after elective
orthopedic surgery: risk factors and natural history. Int J Psy-
chiatry Med 1989;19:109–21.
24. Marcantonio E, Goldman L, Orav E, et al. The association of
intraoperative factors with the development of postoperative
delirium. Am J Med 1998;105:380–4.
25. Lynch E, Lazor M, Gellis J, et al. The impact of postoperative
pain on the development of postoperative delirium. Anesth
Analg 1998;86:781–5.
26. Vaurio L, Sands L, Wang Y, et al. The role of pain and medica-
tions on postoperative delirium [abstract]. Anesthesiology 2004:
A-39.
27. Duggleby W, Lander J. Cognitive status and postoperative pain:
older adults. J Pain Symptom Manage 1994;9:19–27.
28. Jadad AR, Moore RA, Carroll D, et al. Assessing the quality of
reports of randomized clinical trials: is blinding necessary? Con-
trol Clin Trials 1996;17:1–12.
29. Herrick IA, Ganapathy S, Komar W, et al. Postoperative cogni-
tive impairment in the elderly: choice of patient-controlled an-
algesia opioid. Anaesthesia 1996;51:356–60.
30. Rapp SE, Egan KJ, Ross BK, et al. A multidimensional compar-
ison of morphine and hydromorphone patient-controlled anal-
gesia. Anesth Analg 1996;82:1043–8.
31. Silvasti M, Svartling N, Pitkanen M, Rosenberg PH. Comparison
of intravenous patient-controlled analgesia with tramadol ver-
sus morphine after microvascular breast reconstruction. Eur J
Anaesthesiol 2000;17:448–55.
32. Morrison RS, Magaziner J, Gilbert M, et al . Relationship be-
tween pain and opioid analgesics on the development of delir-
ium following hip fracture. J Gerontol A Biol Sci Med Sci 2003;
58:76–81.
33. Marcantonio ER, Juarez G, Goldman L, et al. The relationship of
postoperative delirium with psychoactive medications. JAMA
1994;272:1518–22.
34. Adunsky A, Levy R, Heim M, et al. Meperidine analgesia and
delirium in aged hip fracture patients. Arch Gerontol Geriatr
2002;35:253–9.
35. Folstein M, Folstein S, McHugh P. ‘Mini-mental state.‘ A prac-
tical method for grading the cognitive state of patients for the
clinician. J Psychiatry Res 1975;12:189–98.
36. Pfeiffer E. A short portable mental status questionnaire for the
assessment of organic brain deficit in elderly patients. J Am
Geriatr Soc 1975;23:433–41.
37. Reitan R, Wolfson D. The Halstead-Reitan neuropsychological
test battery. Tucson: Neuropsychology Press, 1985.
38. Wechsler D. Wechsler Adult Intelligence Scale–revised: test
manual. New York: The Psychological Corporation, 1981.
FONG ET AL. 1265
POSTOPERATIVE ANALGESIA AND COGNITIVE IMPAIRMENT
39. Inouye S, van Dyke C, Alessi C, et al. Clarifying confusion: the
confusion assessment method. Ann Intern Med 1990;113:941–8.
40. Diagnostic and Statistical Manual of Mental Disorders, 4th ed.
Washington, DC: American Psychiatric Association, 1994.
41. Mann C, Pouzeratte Y, Boccara G, et al. Comparison of intrave-
nous or epidural patient-controlled analgesia in the elderly after
major abdominal surgery. Anesthesiology 2000;92:433–41.
42. Eriksson-Mjoberg M, Svensson JO, Almkvist O, et al. Extradural
morphine gives better pain relief than patient-controlled i.v.
morphine after hysterectomy. Br J Anaesth 1997;78:10–6.
43. Jitapunkul S, Pillay I, Ebrahim S. The abbreviated mental test: its
use and validity. Age Ageing 1991;20:332–6.
44. Smith MT. Neuroexcitatory effects of morphine and
hydromorphone: evidence implicating the 3-glucuronide me-
tabolites. Clin Exp Pharmacol Physiol 2000;27:524–8.
45. Fainsinger R, Schoeller T, Boiskin M, Bruera E. Palliative care
round: cognitive failure and coma after renal failure in a patient
receiving captopril and hydromorphone. J Palliat Care 1993;9:
53–5.
46. Babul N, Darke AC, Hagen N. Hydromorphone metabolite
accumulation in renal failure. J Pain Symptom Manage 1995;10:
184–6.
47. Lotsch J, Geisslinger G. Morphine-6-glucuronide: an analgesic
of the future? Clin Pharmacokinet 2001;40:485–99.
48. Clark RF, Wei EM, Anderson PO. Meperidine: therapeutic use
and toxicity. J Emerg Med 1995;13:797–802.
49. Eisendrath SJ, Goldman B, Douglas J, et al. Meperidine-induced
delirium. Am J Psychiatry 1987;144:1062–5.
50. Rasmussen LS. Perioperative cognitive decline: the extent of the
problem. Acta Anaesthesiol Belg 1999;50:199–204.
51. Block BM, Liu SS, Rowlingson AJ, et al. Efficacy of postoperative
epidural analgesia: a meta-analysis. JAMA 2003;290:2455–63.
52. Kudoh A, Takase H, Takahira Y, Takazawa T. Postoperative
confusion increases in elderly long-term benzodiazepine users.
Anesth Analg 2004;99:1674–8.
53. Agostini JV, Leo-Summers LS, Inouye SK. Cognitive and other
adverse effects of diphenhydramine use in hospitalized older
patients. Arch Intern Med 2001;161:2091–7.
54. Gengo FM, Dabronzo J, Yurchak A, et al. The relative antihis-
taminic and psychomotor effects of hydroxyzine and cetirizine.
Clin Pharmacol Ther 1987;42:265–72.
55. Wysenbeek AJ, Klein Z, Nakar S, Mane R. Assessment of cog-
nitive function in elderly patients treated with naproxen: a
prospective study. Clin Exp Rheumatol 1988;6:399–400.
56. Simon LS, Mills JA. Drug therapy: nonsteroidal antiinflamma-
tory drugs (first of two parts). N Engl J Med 1980;302:1179–85.
57. Goodwin JS, Regan M. Cognitive dysfunction associated with
naproxen and ibuprofen in the elderly. Arthritis Rheum 1982;
25:1013–5.
58. O’Keeffe ST, Lavan JN. Clinical significance of delirium sub-
types in older people. Age Ageing 1999;28:115–9.
59. Cummings JL. Behavioral complications of drug treatment of
Parkinson’s disease. J Am Geriatr Soc 1991;39:708–16.
60. Schmidt LG, Grohmann R, Strauss A, et al. Epidemiology of
toxic delirium due to psychotropic drugs in psychiatric hospi-
tals. Compr Psychiatry 1987;28:242–9.
61. Feinberg M. The problems of anticholinergic adverse effects in
older patients. Drugs Aging 1993;3:335–48.
62. Gilbert GJ. Quinidine dementia. JAMA 1977;237:2093–4.
63. Cantu TG, Korek JS. Central nervous system reactions to
histamine-2 receptor blockers. Ann Intern Med 1991;114:
1027–34.
64. Slugg PH, Haug MT 3rd, Pippenger CE. Ranitidine pharmaco-
kinetics and adverse central nervous system reactions. Arch
Intern Med 1992;152:2325–9.
65. Litaker D, Locala J, Franco K, et al. Preoperative risk factors for
postoperative delirium. Gen Hosp Psychiatry 2001;23:84–9.
66. Galanakis P, Bickel H, Gradinger R, et al. Acute confusional
state in the elderly following hip surgery: incidence, risk factors
and complications. Int J Geriatr Psychiatry 2001;16:349–55.
1266 FONG ET AL.
POSTOPERATIVE ANALGESIA AND COGNITIVE IMPAIRMENT
67. Leung J, Sands L, Mullen E, et al. Preoperative depression
increases the risk of postoperative delirium in geriatric surgical
patients [abstract]. Anesthesiology 2004.
68. Zakriya KJ, Christmas C, Wenz JF Sr, et al. Preoperative factors
associated with postoperative change in confusion assessment
method score in hip fracture patients. Anesth Analg 2002;94:
1628–32.
69. van Dijk D, Keizer AM, Diephuis JC, et al. Neurocognitive
dysfunction after coronary artery bypass surgery: a systematic
review. J Thorac Cardiovasc Surg 2000;120:632–9.
70. Newman MF, Kirchner JL, Phillips-Bute B, et al. Longitudinal
assessment of neurocognitive function after coronary-artery by-
pass surgery. N Engl J Med 2001;344:395–402.
71. Pugsley W, Klinger L, Paschalis C, et al. The impact of micro-
emboli during cardiopulmonary bypass on neuropsychological
functioning. Stroke 1994;25:1393–9.
ANESTH ANALG
2006;102:1255–66
72. Rolfson DB, McElhaney JE, Rockwood K, et al. Incidence and
risk factors for delirium and other adverse outcomes in older
adults after coronary artery bypass graft surgery. Can J Cardiol
1999;15:771–6.
73. Heyer EJ, Sharma R, Winfree CJ, et al. Severe pain confounds
neuropsychological test performance. J Clin Exp Neuropsychol
2000;22:633–9.
74. Foreman MD. Confusion in the hospitalized elderly: incidence,
onset, and associated factors. Res Nurs Health 1989;12:21–9.
75. Rosenberg J, Kehlet H. Postoperative mental confusion: associ-
ation with postoperative hypoxemia. Surgery 1993;114:76–81.
76. Berggren D, Gustafson Y, Eriksson B, et al. Postoperative con-
fusion after anesthesia in elderly patients with femoral neck
fractures. Anesth Analg 1987;66:497–504.