British Journal of Anaesthesia, 121 (5): 1052e1058 (2018)

doi: 10.1016/j.bja.2018.07.017
Advance Access Publication Date: 31 August
2018 Clinical Practice

C L I N I C A L P R A C T I CE

Agitation in adults in the post-anaesthesia care unit

after general anaesthesia
A. Fields1, J. Huang2, D. Schroeder3, J. Sprung2 and T. Weingarten2,*
1
Department of Internal Medicine, 2Department of Anesthesiology and Perioperative Medicine and
3
Division of Biomedical Statistics and Informatics, Mayo Clinic, Rochester, MN, USA

*Corresponding author. E-mail: weingarten.toby@mayo.edu

Abstract
Background: Agitation after general anaesthesia can lead to self-harm, violence against staff, and increased resource
utilisation. We aimed to assess patient and procedural characteristics associated with this complication in adults.
Methods: We identified cases of agitation (Richmond AgitationeSedation Scale score þ3 or þ4, or administration of
haloperidol) in patients after general anaesthesia in the PACU from July 1, 2010 to September 30, 2016. The cases were
matched 1:1 with control patients without agitation by age, sex, and procedure. Potential clinical associations were
assessed with a multivariable analysis.
Results: We identified agitation in 510 patients [incidence: 2.5 cases/1000 patients; 95% confidence interval (CI): 2.3e2.7].
Variables associated with agitation were substance misuse [odds ratio (OR): 6.77; 95% CI: 1.23e37.2; P¼0.03], cognitive
impairment (OR: 4.66; 95% CI: 1.79e12.1; P¼0.002), obesity (OR: 2.49; 95% CI: 1.66e3.73; P<0.001), psychiatric problems (OR:
2.05; 95% CI: 1.32e3.19; P¼0.002), fall risk (OR: 1.66; 95% CI: 1.02e2.70; P¼0.04), postoperative presence of a tracheal tube
(OR: 16.6; 95% CI: 7.25e38.2; P<0.001), urine catheter (OR: 7.25; 95% CI: 4.31e12.2; P<0.001), nasogastric tube (OR: 4.06; 95%
CI: 1.51e10.9; P¼0.006), or chest tube (OR: 3.46; 95% CI: 1.07e11.2; P¼0.006). Compared with control patients, more
agitated patients had postoperative delirium (16.1% vs 6.3%; P<0.001) and pulmonary complications (9.8% vs 4.7%;
P¼0.002).
Conclusions: Agitation after general anaesthesia was associated with postoperative indwelling catheters, tracheal
intubation and patient features suggestive of pre-existing mental health problems. Anticipation of high-risk patients
could allow allocation of staffing resources to provide a safe environment for anaesthetic recovery.

Keywords: recovery period; anaesthesia, general; anaesthesia, emergence delirium; complication, postoperative;
complications, pulmonary

Editorial decision: 21 July 2018; Accepted: 21 July 2018

© 2018 British Journal of Anaesthesia. Published by Elsevier Ltd. All rights reserved.
For Permissions, please email: permissions@elsevier.com

1052

Editor’s key points
Agitation after general anaesthesia is rare but increases
the care burden for staff.
This case-controlled study assessed the risk factors for
post-anaesthesia agitation and associated subsequent
outcomes.
The characteristics that were associated with the
development of agitation included preoperative cogni-
tive impairment, psychiatric disorders, and obesity.
The presence of invasive devices (e.g. urine catheter,
chest tube, and tracheal tube) was also associated with
agitation.
Agitated patients stayed longer in the recovery area and
were more likely to have subsequent delirium and
respiratory complications.
Early post-anaesthesia recovery is a dynamic period during
which vital organ systems recover from the effects of anaes-
thesia and surgery.1 The return of mental status to pre-
procedural levels is a critical discharge criterion from the
PACU to lower levels of care. Delays in any element of
patient recovery can lead to PACU inefficiencies, resulting in
a bottleneck of patient flow.1 Profound sedation in the PACU
has been associated with increased postoperative adverse
events, including respiratory complications.2,3 In contrast,
post- operative agitation can lead to patient or healthcare team
injury. Agitation that occurs during recovery from general
anaesthesia, emergence agitation, could be partly responsible for
high levels of violence reported against healthcare workers in
recovery rooms.4,5 Furthermore, postoperative agitation could
represent hyperactive delirium; postoperative delirium is
associated with longer hospital stays and increased morbidity,
mortality, and need for institutionalisation.6 Emergence
agitation in adult surgical patients has not been as well stud-
ied as in children. 7 Therefore, we aimed to (i) evaluate the
prevalence of emergence agitation in adults after general
anaesthesia, (ii) assess for potential associations between pa-
tient and perioperative factors and agitation, and (iii) describe
the postoperative course of adult patients who have had
agitation episodes.
Methods
Study design and patient selection
This retrospective, matched, case-controlled study was con-
ducted at a major tertiary academic centre, Mayo Clinic
Hos- pital, Rochester, MN, USA. The two campuses of Mayo
Clinic HospitaldSaint Marys Campus and Methodist
Campusdboth have a PACU. The study was approved by
the Mayo Clinic Institutional Review Board (identification
number: 16-009558). Consistent with Minnesota Statute
144.295, patients who declined authorisation for their
medical records to be used in research were excluded from
this study.
We retrospectively searched our anaesthesia database for
the records of adult patients (≥18 yr) who underwent surgical
procedures with general anaesthesia and were admitted to a
PACU from July 1, 2010 to September 30, 2016, and provided
authorisation for research use of their medical records. The
electronic health records of patients were searched to identify
those who experienced emergence agitation. Emergence agita-
tion was defined as a Richmond AgitationeSedation Scale
(RASS) score8 (Supplementary Appendix) of þ3 or þ4 or the
Postoperative agitation in adult patients ■ 1053
administration of haloperidol during the PACU stay.
Patients identified as having emergence agitation (cases)
were then matched 1:1 with patients without emergence
agitation (con- trols) on the basis of age, sex, and procedure
type. Automated and manual review of patient records was
performed.
Study setting and clinical practice
Before operation, all patients undergo nursing review of
their medical records, home medications, fall risk
assessment with a validated risk model (Hendrich II Fall Risk
Model, which as- sesses risk of fall using history of
confusion, depression, ver- tigo, and difficulties with
elimination; use of anticonvulsants or benzodiazepines;
male sex; and ability to rise up from a chair),9 current use of
tobacco and alcohol, and current and past substance misuse
history. In addition, all patients are questioned regarding a
history of obstructive sleep apnoea; because of the high
prevalence of undiagnosed obstructive sleep apnoea, those
who respond in the negative undergo additional screening.10
Patients are also assessed by the attending
anaesthesiologist. Anaesthetic management was predicated
on patient and procedural characteristics, at the discretion of
the attending anaesthesiologist. All study pa- tients
underwent Phase 1 recovery (the recovery phase during which
close monitoring is required) in one of the two PACUs. These
PACUs were staffed by registered nurses trained in Phase 1
recovery, and the Saint Marys Campus PACU staff also
included an anaesthesiology resident. The attending anaes-
thesiologist was immediately available if advanced
expertise was required. Phase 1 discharge criteria are based on
standard criteria11 and adequate control of pain and
postoperative nausea and vomiting (PONV). As an added
layer of safety, patients are also continually assessed for
episodes of respi- ratory depression (e.g. apnoeic spells). 12
Disposition from the PACU to an ambulatory area,
postoperative ward, or ICU is decided by the attending
anaesthesiologist on the basis of patient and procedural
characteristics.
Data abstraction
Medical, surgical, and anaesthesia records were electronically
abstracted using previously described proprietary software. 13
Pre-surgical variables included patient age; sex; BMI; pres-
ence of co-morbid conditions (cardiovascular disease, pulmo-
nary disease, dementia/cognitive impairment, or diabetes
mellitus); history of psychiatric problems (defined as a previ-
ous psychiatric diagnosis or consultation or chronic use of
antidepressants or antipsychotic medications); results of pre-
operative nursing assessments; and home use of benzodiaze-
pines, gabapentinoids, and opioids. Anaesthesia variables
included anaesthetic technique and medications administered
(from time of admission to dismissal from the PACU).
Surgical records were reviewed for type of procedure, blood
product transfusions, fluid resuscitation, and surgical duration.
The continued postoperative use of indwelling catheters (e.g.
urine catheter, nasogastric tube, chest tube, or surgical drain)
or tracheal tube, which could contribute to patient
discomfort and potentially trigger agitation, was also
abstracted.
The patient records were reviewed for evidence of medical
or anaesthesia complications. These included PACU
administra- tion of medications to treat agitation, pain, PONV,
and hyper- tension; evidence of respiratory depression during
Phase 1 recovery; postoperative complications {pulmonary,
 1054 ■ Fields et al.
myocardial infarction, thromboembolic, or delirium [all
inpatients
 Postoperative agitation in adult patients ■ 1055

regardless of level of care are assessed with the Confusion

Assessment Method for the ICU (CAM-ICU)14 twice daily or
more often if mental status is noted to fluctuate]}; the need for
behavioural emergency response team15 or emergency
response team16 activation; ICU admission; and in-hospital
death.

Statistical analysis
Data were summarised as mean (standard deviation) or
median [inter-quartile range (IQR)] for continuous variables
and fre- quency (percentage) for nominal variables. Analyses
to assess the characteristics potentially associated with
emergence agitation were performed using multivariable
conditional lo- gistic regression and reported as odds ratio
(OR) and 95% con- fidence interval (CI). Outcomes were
assessed with univariate comparisons, with Phase 1 agitation
as the dependent variable; continuous variables were assessed
with appropriate two- sample methods (t-test or rank-sum
test), and categorical var- iables were assessed with Fisher’s
exact test. Two-tailed P- values are reported, and P<0.05 was
considered significant. No adjustments were made for
multiple comparisons. Analyses were performed using JMP
Pro (version 13.0.0; SAS Institute, Inc., Cary, North Carolina,
USA).

Results
During the study period, 207 569 patients were admitted to the
PACU, of whom 510 patients (0.25%) had emergence agitation,
for an incidence of 2.5 (95% CI: 2.3e2.7) cases per 1000
patients. Among the 510 patients, 446 (87.5%) had a RASS
score of þ3
or þ4 (IQR: þ3 to þ3) and 160 (31.4%) received haloperidol; 96
patients (18.8%) had both an increased RASS score and
received haloperidol. For these cases, 108 patients (21.2%)
required additional healthcare staff involvement and 29
(5.7%) required physical restraints. These episodes resulted
in six self-inflicted dermatological injuries and the removal of
six i.v. lines and three nasogastric tubes; there was no
documented self-extubation or removal of urine catheters.
Fifteen patients
(2.9%) were documented to inflict violence upon healthcare
staff, including hitting (n¼9), kicking (n¼6), grabbing (n¼1),
and spitting (n¼1).

Table 1 Patient characteristics and surgery types used to

match cases and controls. Values are mean (standard devia-
tion) or no. of patients (%)

Characteristic Cases: agitation Controls: no

(n¼510) agitation (n¼510)

Age (yr) 63.2 (17.9) 63.1 (17.7)

Age range Min 18.1, Min 18.9,
Max 97.6 Max 94.6
Men 372 (72.9) 372 (72.9)
Surgery type
General 132 (25.9) 132 (25.9)
Orthopaedic 133 (26.1) 133 (26.1)
Neurosurgical 75 (14.7) 75 (14.7)
Thoracic 64 (12.5) 64 (12.5)
Otolaryngologic 53 (10.4) 53 (10.4)
Urologic/ 37 (7.3) 37 (7.3)
gynaecologic
Vascular 11 (2.2) 11 (2.2)
Plastics 4 (0.8) 4 (0.8)
Interventional 1 (0.2) 1 (0.2)
radiology
 1056 ■ Fields et al.
After 1:1 matching of cases and controls, the groups were agitation to a Riker score of 6 or higher (very agitated to
similar in most patient characteristics (Table 1). Among the dangerous agitation), the estimated incidence of emergence
controls, the median highest RASS score during PACU agitation would be 1.4%17
stay to 2.5%,5 somewhat similar to our estimate. If we had liber-
was 0 (IQR: e1 to 0); 53 controls had a RASS of þ1 and two alised our definition to a RASS score of þ1 or higher, 10.8%
controls had a RASS of þ2. Table 2 shows a comparison of our controls would have met these agitation criteria.
of
patient and perioperative characteristics between the cases
and controls and a multivariable assessment of these char-
acteristics as risk factors for agitation in the PACU. The
cases had more evidence of mental health or cognitive
impairment problems than the controls. Specifically,
preoperative patient variables associated with agitation
were substance misuse
history (OR: 6.77; 95% CI: 1.23e37.2; P¼0.03), cognitive
impairment (OR: 4.66; 95% CI: 1.79e12.1; P¼0.002), obesity
(OR:
2.49; 95% CI: 1.66e3.73; P<0.001), psychiatric problems (OR:
2.05; 95% CI: 1.32e3.19; P¼0.002), and fall risk (OR: 1.66; 95%
CI:
1.02e2.70; P¼0.04) (Table 2). In addition, intraoperative char-
acteristics associated with agitation were the use of des-
flurane (OR: 2.02; 95% CI: 1.32e3.10; P¼0.001) and higher doses
of opioids (OR: 1.19; 95% CI: 1.01e1.39; P¼0.03).
Surprisingly,
the intraoperative use of the psychotropic medication keta-
mine, which may be associated with agitation, was not found
to be associated with agitation (OR: 0.63; 95% CI: 0.37e1.05;
P¼0.08), and the use of droperidol was found to be
protective
from agitation (OR: 0.41; 95% CI: 0.23e0.73; P¼0.002) (Table 2).
Strong associations were found between agitation and the
postoperative presence of a tracheal tube (‘delayed extuba-
tion’, OR: 16.6; 95% CI: 7.25e38.2; P<0.001), urine catheter
(OR:
7.25; 95% CI: 4.31e12.2; P<0.001), nasogastric tube (OR: 4.06;
95% CI: 1.51e10.9; P¼0.006), and chest tube (OR: 3.46; 95% CI:
1.07e11.2; P¼0.04).
Table 3 summarises the course of Phase 1 anaesthesia re-
covery. Compared with the controls, the cases had longer
PACU stay (108 vs 71.4 min; P<0.001); had a higher incidence
of PONV (11.6% vs 5.5%; P<0.001); more required treatments
for hypertension (9.2% vs 2.5%; P<0.001); and more
received opi- oids (70.4% vs 48.2%; P<0.001), ketamine
(12.7% vs 7.1%; P¼0.003), and sedatives (midazolam: 29.0%
vs 1.6%; P<0.001).
Although the cases and controls had comparable post-PACU
disposition (outpatients, transfer to wards, or ICU) (P¼0.28),
the cases had higher incidence of subsequent delirium (16.1%
vs 6.3%; P<0.001) and pulmonary complications (9.8% vs
4.7%; P¼0.002) than did the controls (Table 4).

Discussion
In this retrospective analysis, severe emergence agitation
complicated 2.5 per 1000 cases after general anaesthesia.
In prospective observational studies, the incidence of
emer- gence agitation has been much higher, ranging from
4.7%5 to 19%.7 This discrepancy may reflect that our
definition was restricted to patients who were agitated or
combative (RASS
score: þ3 or þ4) or received haloperidol. We believe these
criteria represent patients who are at the greatest risk for
harming themselves or the healthcare staff. Card and col-
leagues7 included patients with a RASS score of þ1 or higher
and reported an incidence of 19%. Two other studies5,17
used
a Riker scale18 score of 5 or higher (anxious or mildly
agitated) and reported incidences of 4.7%5 and 9.8%.17 If
those studies had restricted the definition of emergence
 Postoperative agitation in adult patients ■ 1057

Table 2 Comparison of patient and perioperative characteristics: cases vs matched controls. IVME, i.v. morphine equivalents; OSA,
obstructive sleep apnoea. *Values are no. of patients (%) or median (inter-quartile range). yObesity is defined as BMI ≥30 kg m—2.
The mean (standard deviation) BMI was 30.0 (7.1) kg m —2 for the cases and 27.2 (5.1) kg m—2 for the controls. zPsychiatric history was
defined as a previous note in the medical record indicating psychiatric problems (98 cases; 50 controls), use of antidepressant
medication (103
cases; 69 controls), or use of antipsychotic medication (38 cases; 14 controls). ¶Odds ratio presented by units of 10

Rick factor Univariate analysis* Multivariable analysis

Cases: agitation Controls: no P-value Odds ratio (95% CI) P-value

(n¼510) agitation (n¼510)

Co-morbid conditions
Obesityy 218 (42.7) 132 (25.9) <0.001 2.49 (1.66e3.73) <0.001
Cardiovascular disease 87 (17.1) 76 (14.9) 0.32 1.48 (0.84e2.61) 0.17
Cognitive impairment 44 (8.6) 10 (2.0) <0.001 4.66 (1.79e12.1) 0.002
Pulmonary disease 89 (17.5) 82 (16.1) 0.54 0.88 (0.52e1.49) 0.63
Diabetes mellitus 114 (22.4) 83 (16.3) 0.01 1.27 (0.78e2.06) 0.33
OSA 116 (22.7) 95 (18.6) 0.11 1.26 (0.80e1.99) 0.31
Psychiatric historyz 172 (33.7) 105 (20.6) <0.001 2.05 (1.32e3.19) 0.002
Other risk factors
Fall risk 145 (28.4) 93 (18.2) <0.001 1.66 (1.02e2.70) 0.04
Tobacco use 57 (11.2) 44 (8.6) 0.14 1.77 (0.84e3.72) 0.13
Alcohol use 108 (21.2) 139 (27.3) 0.02 0.66 (0.40e1.07) 0.09
Substance misuse 13 (2.5) 2 (0.4) 0.01 6.77 (1.23e37.2) 0.03
Home medications
Benzodiazepines 86 (16.9) 57 (11.2) 0.01 1.12 (0.64e1.97) 0.69
Gabapentinoids 19 (3.7) 14 (2.7) 0.39 0.59 (0.20e1.70) 0.33
Opioids 191 (37.5) 182 (35.7) 0.56 0.77 (0.51e1.16) 0.22
Intraoperative factors
Duration of surgery (min) 129 (75.0e209.3) 123 (77e216.0) 0.38 0.98 (0.96e1.00)¶ 0.09
Neuraxial opioids 44 (8.6) 46 (9.0) 0.75 2.00 (0.57e6.98) 0.28
Blood product transfusion 67 (13.1) 47 (9.2) 0.04 1.34 (0.65e2.77) 0.43
Crystalloids (L) 1.7 (1.1e2.8) 1.5 (0.9e2.5) 0.004 1.04 (0.83e1.31) 0.73
Delayed extubation 112 (22.0) 18 (3.5) <0.001 16.6 (7.25e38.2) <0.001
Nasogastric tube 61 (12.0) 23 (4.5) <0.001 4.06 (1.51e10.9) 0.006
Chest tube 60 (11.8) 37 (7.3) <0.001 3.46 (1.07e11.2) 0.04
Surgical drain 74 (14.5) 112 (22.0) 0.002 0.44 (0.26e0.76) 0.003
Urine catheter 333 (65.3) 183 (35.9) <0.001 7.25 (4.31e12.2) <0.001
Intraoperative medications
Opioids (IVME mg) 31 (23e42) 25 (15e35) <0.001 1.19 (1.01e1.39)¶ 0.03
Desflurane volatile 245 (48.0) 205 (40.2) 0.007 2.02 (1.32e3.10) 0.001
Ketamine 76 (14.9) 117 (22.9) 0.001 0.63 (0.37e1.05) 0.08
Midazolam 230 (45.1) 219 (42.9) 0.42 1.26 (0.80e1.98) 0.33
Droperidol 94 (18.4) 127 (24.9) 0.006 0.41 (0.23e0.73) 0.002
Neostigmine 296 (58.0) 269 (52.7) 0.32 1.04 (0.64e1.70) 0.88

Emergence agitation differs from emergence delirium.

risk was associated with an increased risk of agitation. Cogni-
Delirium is an acute confusional state and can present with
tive impairment is a debilitating condition known to be
hypoactive or hyperactive signs. Delirium is difficult to assess
asso- ciated with various co-morbid conditions, frailty,21
during recovery from anaesthesia, because the residual effects
and increased fall risk. 22 Therefore, the associations between
of anaesthesia can mimic delirium.7 It is likely that a
these factors and agitation are most likely interrelated.
propor- tion of our cases had the hyperactive variant of
Substance misuse disorders were strongly associated
delirium. Because we did not formally test for delirium, we
with emergence agitation. Such disorders are known to
do not know the incidence of emergence delirium. In the
increase the risk of postoperative delirium, which may be
series by Card and colleagues,7 60% of their agitated
secondary to acute abstinence.23,24 However, we do not
patients were found to be delirious by the CAM-ICU
believe that absti- nence during the immediate perioperative
criteria.
period is of suffi- cient duration to trigger withdrawal
Several clinical characteristics were associated with emer-
symptoms during the period of post-anaesthetic recovery.
gence agitation. The prominent patient features were those
Similarly, a psychiatric history was associated with an
suggestive of mental health impairment, as evidenced by
increased emergence agitation. A meta-analysis found that
strong associations with cognitive impairment, psychiatric
such disorders increase post- operative delirium risk,25 but
history, and substance misuse. In a prospective
this has not been specifically studied in the PACU setting.
observational study of 91 older adults, cognitive impairment
It is not clear why obesity was
was associated with increased risk of delirium in the PACU.19
associated with agitation. When multiple hypotheses are
Similarly, Sprung and colleagues20 showed that preoperative
explored, it is possible that spurious ‘positive’ or ‘negative’
cognitive impair- ment was associated with postoperative
results will be found. This could apply to some of our
delirium, and post- operative delirium in patients without findings.
known pre-existing cognitive impairment was associated with For example, the association between desflurane and agitation
future development of cognitive impairment. In the current and the lack of association between ketamine and agitation
study, an increased fall were both surprising results.
 1058 ■ Fields et al.

Table 3 Phase 1 anaesthesia recovery course and treatments Table 4 Outcomes: cases vs matched controls. Values are
in PACU: cases vs matched controls. Values are no. of no. of patients (%) or median (inter-quartile range)
patients (%), mean (standard deviation), or median (inter-
quartile range). IVME, i.v. morphine equivalents; NRS,
Outcome Cases: Controls: no P-
numeric pain score. *Urinary symptoms included report of agitation agitation value
urgency or (n¼510) (n¼510)
bladder scan identifying urinary retention
50 (9.8) 24 (4.7) 0.002
Pulmonary
Outcome Cases: Controls: no P- complications
agitation agitation) value 36 (7.1) 19 (3.7) 0.03
Pneumonia 19 (3.7) 5 (1.0) 0.006
(n¼510) (n¼510) Acute respiratory
distress syndrome 17 (3.3) 3 (0.6) 0.002
Report of severe 108 (21.2) 106 (20.8) 0.94 Reintubation (≤48 h
pain (NRS ≥7) after operation) 82 (16.1) 32 (6.3) <0.001
Respiratory 80 (15.7) 24 (4.7) <0.001 Delirium 7 (1.4) 4 (0.8) 0.55
depression Myocardial infarction 11 (2.2) 6 (1.2) 0.33
episodes Thromboembolic
Opioid analgesic 359 (70.4) 246 (48.2) <0.001 complication 11 (2.2) 2 (0.4) 0.02
requirement Behavioural emergency
Opioid dose 9.7 (11.6) 5.3 (8.0) <0.001 response team
(IVME mg) 13 (2.5) 7 (1.4) 0.26
activation
Midazolam 148 (29.0) 8 (1.6) <0.001 Rapid response 18 (3.5) 11 (2.2) 0.26
administration team activation 3 (0.6) 1 (0.2) 0.62
Ketamine 65 (12.7) 36 (7.1) 0.003 ICU admission 2 (1e6) 2 (1e5) 0.04
administration In-hospital death
Postoperative 59 (11.6) 28 (5.5) <0.001 Hospital length of stay 11 (2.2) 4 (0.8) 0.12
nausea or (day)
vomiting 30-day death
Charted nausea 25 (4.9) 7 (1.4) 0.002
Charted emesis 2 (0.4) 2 (0.4) >0.99
Rescue anti- 40 (7.8) 22 (4.3) 0.03
emetic
administration
emergency departments or psychiatric facilities, violence also
Urinary 112 (22.0) 18 (3.5) <0.001
symptoms* occurs in PACUs. 4 Many agitated patients required physical
Hypertension 47 (9.2) 13 (2.5) <0.001 restraints and more staff members to provide safe care. The
requiring higher rates of midazolam administration probably reflect
treatment attempts to treat agitation and may reflect another form of
PACU stay 108 71.4 <0.001
‘chemical restraint’ besides haloperidol. After PACU
duration (min) (77.4e151.0) (46.0e110.0)
discharge, agitated patients were more likely to have delirium
Disposition 0.28
Outpatient 196 (38.4) 172 (33.7) or require behavioural evaluation response team in-
Standard ward 188 (36.9) 207 (40.6) terventions. Studies of emergence delirium have also found
ICU 126 (24.7) 131 (25.7) that these patients are at an increased risk for postoperative
delirium.7,19,29,30 Agitated patients may, therefore, benefit
Several features of Phase 1 recovery (e.g. PONV and need for from anti-delirium measures.31 The higher incidence of pul-
further analgesics) could contribute to agitation. monary complications in patients who were agitated might
Hypercarbia secondary to respiratory depression can lead to reflect that agitated patients more commonly experienced
restlessness and agitation.26 Patient discomfort can also respiratory depression in the PACU. Respiratory depression
contribute to agitation.5,17 Prolonged mechanical ventilation during Phase 1 recovery predicts subsequent postoperative
and presence of indwelling catheters are well-known risk pulmonary complications.3,12
factors for post- operative delirium.27,28 Theoretically, This study has the inherent limitations of a retrospective
discomfort from these devices could trigger agitation during design. Even after accounting for the more stringent criteria of
anaesthesia recovery. This could also explain why agitated emergence agitation used in this study, the incidence of severe
patients were more likely to receive larger doses of opioids; agitation was lower than expected. 5,17 This probably reflects
however, the incidence of severe pain did not differ the deficiency of relying on retrospective data, and our inci-
between groups. In contrast, Card and colleagues7 found dence of emergence agitation is probably underestimated.
that delirium risk was greater with increasing doses of Similarly, several risk factors, such as substance misuse and
perioperative opioids. The use of ketamine in the PACU was psychiatric history, may not be accurately documented.
also higher in agitated patients, but its administration in Furthermore, substance misuse screening was potentially
relation to agitation onset could not be established. biased to reflect clinicians’ impressions regarding which pa-
However, it was presumably administered before agitation tients might be misusing substances. The matched study
onset, because ketamine administration for anal- gesia in design precluded an assessment of the effect of age, sex,
agitated patients is discouraged. Thus, ketamine use in the and type of procedure, known risk factors for postoperative
PACU may have contributed to agitation. delirium; in turn, postoperative delirium is known to be highly
Not surprisingly, postoperative recovery for agitated pa- prevalent in agitated postoperative patients. The assessment
tients was more complicated. Several patients inflicted self- of postoperative delirium was based on the CAM-ICU criteria.
harm or committed violence. Although violence against
healthcare staff is widely believed to primarily occur in The CAM-ICU has lower sensitivity among patients on

postoperative wards than when used in the ICU,7 so our results
may reflect an underestimate of this complication.
In conclusion, agitation in the PACU in adult patients is
uncommon but can result in patient self-harm or violence
against healthcare staff, requires added resources, and is
associated with an increased risk of postoperative delirium.
We identified that patients requiring tracheal tubes and
indwelling catheters after operation, and those with features
suggestive of mental health impairment, were at the greatest
risk for emergence delirium. A better understanding of po-
tential risks and precipitating factors for emergence
agitation in adults may be useful for developing appropriate
staffing models to provide safe care, institute prevention
strategies, understand the risk of subsequent postoperative
complica- tions, and guide appropriate disposition
protocols.
Authors’ contributions
Study conception/design: T.N.W.
Data acquisition and interpretation: A.R.F.,
J.H. Data analysis: D.R.S., T.N.W.
Drafting of paper: T.N.W.
Study supervision: T.N.W., J.S.
Critical revisions for important intellectual content: all
authors.
Final approval of manuscript: all authors.
Acknowledgements
The authors thank the Anesthesia Clinical Research Unit study
coordinator, Ms Danette L. Bruns, for her help with data
extraction.
Declarations of interest
T.N.W. currently serves as a consultant to Medtronic in the
role as Chairman of the Clinical Endpoint Committee for
the Prodigy Trial; has received research support from
Respiratory Motion (study equipment), and unrestricted
investigator- initiated grants from Merck (active) and
Baxter (completed). The other authors have no conflicts of
interest to report.
Appendix A. Supplementary data
Supplementary data related to this article can be found at
https://doi.org/10.1016/j.bja.2018.07.017.
References
1. Weingarten TN, Bergan TS, Narr BJ, Schroeder DR,
Sprung J. Effects of changes in intraoperative manage-
ment on recovery from anesthesia: a review of practice
improvement initiative. BMC Anesthesiol 2015; 15: 54
2. Weingarten TN, Chong EY, Schroeder DR, Sprung J. Pre-
dictors and outcomes following naloxone administration
during Phase I anesthesia recovery. J Anesth 2016; 30:
116e22
3. Weingarten TN, Herasevich V, McGlinch MC, et al. Pre-
dictors of delayed postoperative respiratory depression
assessed from naloxone administration. Anesth Analg
2015; 121: 422e9
4. Hahn S, Muller M, Hantikainen V, Kok G, Dassen T,
Halfens RJ. Risk factors associated with patient and visitor
Postoperative agitation in adult patients ■ 1059
violence in general hospitals: results of a multiple
regression analysis. Int J Nurs Stud 2013; 50: 374e85
5. Lepouse C, Lautner CA, Liu L, Gomis P, Leon A.
Emergence delirium in adults in the post-anaesthesia care
unit. Br J Anaesth 2006; 96: 747e53
6. Marcantonio ER. Delirium in hospitalized older adults.
N Engl J Med 2017; 377: 1456e66
7. Card E, Pandharipande P, Tomes C, et al. Emergence from
general anaesthesia and evolution of delirium signs in the
post-anaesthesia care unit. Br J Anaesth 2015; 115: 411e7
8. Sessler CN, Gosnell MS, Grap MJ, et al. The Richmond
Agitation-Sedation Scale: validity and reliability in adult
intensive care unit patients. Am J Respir Crit Care Med
2002; 166: 1338e44
9. Hendrich AL, Bender PS, Nyhuis A. Validation of the
Hen- drich II Fall Risk Model: a large concurrent
case/control study of hospitalized patients. Appl Nurs Res
2003; 16: 9e21
10. Flemons WW. Clinical practice. Obstructive sleep apnea.
N Engl J Med 2002; 347: 498e504
11. Aldrete JA. The post-anesthesia recovery score revisited.
J Clin Anesth 1995; 7: 89e91
12. Gali B, Whalen FX, Schroeder DR, Gay PC, Plevak DJ.
Identification of patients at risk for postoperative respi-
ratory complications using a preoperative obstructive
sleep apnea screening tool and postanesthesia care
assessment. Anesthesiology 2009; 110: 869e77
13. Herasevich V, Kor DJ, Li M, Pickering BW. ICU data mart:
a non-iT approach. A team of clinicians, researchers and
informatics personnel at the Mayo Clinic have taken a
homegrown approach to building an ICU data mart.
Healthc Inform 2011; 28: 44e5
14. Ely EW, Inouye SK, Bernard GR, et al. Delirium in me-
chanically ventilated patients: validity and reliability of
the confusion assessment method for the intensive care
unit (CAM-ICU). JAMA 2001; 286: 2703e10
15. Loucks J, Rutledge DN, Hatch B, Morrison V. Rapid
response team for behavioral emergencies. J Am Psychiatr
Nurses Assoc 2010; 16: 93e100
16. Weingarten TN, Venus SJ, Whalen FX, et al. Postoperative
emergency response team activation at a large tertiary
medical center. Mayo Clin Proc 2012; 87: 41e9
17. Kim HC, Kim E, Jeon YT, et al. Postanaesthetic emergence
agitation in adult patients after general anaesthesia for
urological surgery. J Int Med Res 2015; 43: 226e35
18. Riker RR, Picard JT, Fraser GL. Prospective evaluation of
the Sedation-Agitation Scale for adult critically ill pa-
tients. Crit Care Med 1999; 27: 1325e9
19. Neufeld KJ, Leoutsakos JM, Sieber FE, et al. Outcomes of
early delirium diagnosis after general anesthesia in the
elderly. Anesth Analg 2013; 117: 471e8
20. Sprung J, Roberts RO, Weingarten TN, et al. Postoperative
delirium in elderly patients is associated with subsequent
cognitive impairment. Br J Anaesth 2017; 119: 316e23
21. Searle SD, Rockwood K. Frailty and the risk of cognitive
impairment. Alzheimers Res Ther 2015; 7: 54
22. Tinetti ME, Speechley M, Ginter SF. Risk factors for falls
among elderly persons living in the community. N Engl J
Med 1988; 319: 1701e7
23. Fineberg SJ, Nandyala SV, Marquez-Lara A, Oglesby M,
Patel AA, Singh K. Incidence and risk factors for post-
operative delirium after lumbar spine surgery. Spine (Phila
Pa 1976) 2013; 38: 1790e6
24. Yu YH, Chen AC, Hu CC, Hsieh PH, Ueng SW, Lee MS.
Acute delirium and poor compliance in total hip
 1060 ■ Fields et al.

arthroplasty patients with substance abuse disorders. de Castro SM, Unlu C, Tuynman JB, et al. Incidence and risk
J Arthroplasty 2012; 27: 1526e9 factors of delirium in the elderly general surgical patient. Am J
25. Copeland LA, Zeber JE, Pugh MJ, Mortensen EM, Surg 2014; 208: 26e32
28. Radtke FM, Franck M, MacGuill M, et al. Duration of fluid
Restrepo MI, Lawrence VA. Postoperative complications in
fasting and choice of analgesic are modifiable factors for
the seriously mentally ill: a systematic review of the
early postoperative delirium. Eur J Anaesthesiol 2010; 27:
literature. Ann Surg 2008; 248: 31e8
411e6
26. Kaye J, Buchanan F, Kendrick A, et al. Acute carbon
29. Sharma PT, Sieber FE, Zakriya KJ, et al. Recovery room
dioxide exposure in healthy adults: evaluation of a novel
delirium predicts postoperative delirium after hip-fracture
means of investigating the stress response. J
repair. Anesth Analg 2005; 101: 1215e20. table of contents
Neuroendocrinol 2004; 16: 256e64
30. Barr J, Fraser GL, Puntillo K, et al. Clinical practice
27. Mufti HN, Hirsch GM. Perioperative prediction of
guide- lines for the management of pain, agitation, and
agitated (hyperactive) delirium after cardiac surgery in
delirium in adult patients in the intensive care unit. Crit
adultsdthe development of a practical scorecard. J Crit
Care Med 2013; 41: 263e306
Care 2017; 42: 192e9

Handling editor: M. Avidan

Postoperative agitation in adult patients ■ 1061