Abstract

Background. Emergence delirium in the post-anaesthesia care unit (PACU) is poorly understood. The goal of this prospective study was to determine frequency and risk factors of emergence delirium in adults after general anaesthesia.

Methods. In this prospective study, 1359 consecutive patients were included. Contextual risk factors and occurrence of delirium according to the Riker sedation scale were documented. Groups were defined for the analysis according to the occurrence or not of agitation, then after exclusion of patients with preoperative anxiety and neuroleptics, or both, and antidepressants or benzodiazepines treatments.

Results. Sixty-four (4.7%) patients developed delirium in the PACU, which can go from thrashing to violent behaviour and removal of tubes and catheters. Preoperative anxiety was not found to be a risk factor. Preoperative medication by benzodiazepines (OR=1.910, 95% CI=1.101–3.315, P=0.021), breast surgery (OR=5.190, 95% CI=1.422–18.947, P=0.013), abdominal surgery (OR=3.206, 95% CI=1.262–8.143, P=0.014), and long duration of surgery increased the risk of delirium (OR=1.005, 95% CI=1.002–1.008, P=0.001), while a previous history of illness and long-term treatment by antidepressants decreased the risk (respectively, OR=0.544, 95% CI=0.315–0.939, P=0.029 and OR=0.245, 95% CI=0.084–0.710, P=0.010).

Conclusions. Preoperative benzodiazepines, breast and abdominal surgery and surgery of long duration are risk factors for emergence delirium.

Emergence delirium in the postanaesthesia care unit (PACU) is poorly investigated. Such a delirium can suddenly become dangerous and have serious consequences for the patient such as injury, increased pain, haemorrhage, self-extubation and removal of catheters requiring physical or chemical restraint. The incidence of this type of delirium is around 3%,1 but has rarely been studied with the exception of paediatric anaesthesia.2,3 Furthermore, both the definition of emergence delirium and the phrasing (delirium, confusion, excitement, agitation) vary, as does its onset interval (upon awakening, within the first few hours, days or weeks). Another difficulty is the quantification of this agitation, as there is no specific scale.

We studied only acute, short-lived (that is, resolved at time of PACU discharge) postoperative delirium in the PACU after general anaesthesia. The goals of this prospective study are to define incidence and risk factors of emergence delirium in adults.

Methods

The study protocol was approved by our local ethics committee and informed consent was obtained from all patients. Every patient admitted to the PACU after general anaesthesia over a 3 month period, from November 2003 to February 2004, was included in our study. This 12-bed PACU is open day and night and is located next to the operating rooms (abdominal surgery, urology, ophthalmology, ENT, vascular surgery and abdominal endoscopies). Each PACU nurse is responsible for 2–3 patients and one nursing auxiliary is on-site for the entire unit. An anaesthetist is responsible for the discharge of patients. Every patient admitted to PACU during the study period was included in the study, excluding those under 15 yr of age, those for whom dementia had been diagnosed and those under loco-regional anaesthesia.

After a preliminary feasibility study performed in the PACU with a small cohort of consecutive patients (n=150), we used the Riker sedation–agitation scale (Table 1), which was developed to assess the level of agitation and sedation in intensive-care patients, to evaluate delirium in the PACU.4 We divided the patients into two categories using the Riker sedation–agitation scale: non-agitated patients (levels 1–4) and agitated patients (levels 5–7). We then subdivided those patients in the agitated group as agitated, very agitated and dangerously agitated. This scale is easy to use, even in emergency situations. Two nurses with a long PACU experience were responsible for evaluating the occurrence of agitation. These nurses had no responsibility concerning the care of the patients.

Table 1

Riker sedation-agitated scale4

 Agitated 
    Dangerous agitation Pulling at ET tube, trying to remove catheters, climbing over bed rail, striking at staff, thrashing side-to-side 
    Very agitated Does not calm, despite frequent verbal reminding of limits; requires physical restraints, biting ET tube 
    Agitated Anxious or mildly agitated, attempting to sit up, calms down to verbal instructions 
 Non-agitated 
    Calm and cooperative Calm, awakens easily, follows commands 
    Sedated Difficult to arouse, awakens to verbal stimuli or gentle shaking but drifts off again, follows simple commands 
    Very sedated Arouses to physical stimuli but does not communicate or follow commands may move spontaneously 
    Unarousable Minimal or no response to noxious stimuli, does not communicate or follow commands 
 Agitated 
    Dangerous agitation Pulling at ET tube, trying to remove catheters, climbing over bed rail, striking at staff, thrashing side-to-side 
    Very agitated Does not calm, despite frequent verbal reminding of limits; requires physical restraints, biting ET tube 
    Agitated Anxious or mildly agitated, attempting to sit up, calms down to verbal instructions 
 Non-agitated 
    Calm and cooperative Calm, awakens easily, follows commands 
    Sedated Difficult to arouse, awakens to verbal stimuli or gentle shaking but drifts off again, follows simple commands 
    Very sedated Arouses to physical stimuli but does not communicate or follow commands may move spontaneously 
    Unarousable Minimal or no response to noxious stimuli, does not communicate or follow commands 

For each patient, the following variables were noted: age, gender, ASA score, pre-existing medical or surgical condition (cardiac, pulmonary, neurological, endocrinological, abdominal, psychiatric or kidney disorder, cancer, and drug or alcohol abuse), long-term (>1 month) treatment by benzodiazepines or antidepressants or neuroleptics, severe preoperative anxiety (as judged by a physician by a binary (yes/no) evaluation), recent surgery (either within the previous 2 weeks or the previous 3 months), length of stay before anaesthesia, emergency or non-emergency surgery, preoperative medication (benzodiazepines or hydroxyzine), type of surgery and its length, anaesthesia protocol (induction and maintenance of anaesthesia by propofol (Diprifusor TCI system), maintenance by inhaled anaesthetics after i.v. induction, or induction and maintenance by inhaled anaesthetics, type of inhaled anaesthetics, use of neuromuscular blockers, length of stay in the PACU and mortality at 3 months. For patients with emergence delirium, we also noted agitation score, its duration and consequences, the number of people needed to control the situation, and, where possible, the most probable cause. An appropriate treatment was then performed with analgetics and anxiolytics. For all patients, pain was assessed and treated so as to obtain a pain level <3 on the visual analogical scale. Aetiologies were assessed on patient records by the anaesthetist responsible for the PACU. This anaesthetist was blinded to the study goals.

This study was conducted as a prospective study. Two groups were defined by the occurrence of agitation: one of agitated patients, and one of non-agitated patients. First, data of the two groups were pooled to analyse all patients as one group, to present an epidemiological knowledge, then each group was classically analysed. In the univariate analysis, tests used to compare groups were the Student t-test, the Kruskal–Wallis test and χ2-test as required. A multivariate analysis was performed using a backward binary stepwise logistic regression (SYSTAT 8.0, SPSS) to examine and determine risk factors of delirium. All variables statistically linked to occurrence of delirium were included in the regression model with delirium as the dependent variable and the independent variables were age, previous history of illness, type of surgery, long-term treatment by antidepressants or neuroleptics, preoperative anxiety, preoperative medication by benzodiazepines or hydroxyzine, length of surgery, type of anaesthesia and ASA score. A second statistical analysis (both univariate and multivariate) was then conducted on a subgroup of patients. This subgroup of patients was defined by patients who had undergone general anaesthesia, and excluding those deemed anxious, or treated by neuroleptics, antidepressants and benzodiazepines, so as to avoid those patients with the probable highest risk of delirium. Data are shown as mean (sd). P <0.050 was considered significant.

Results

Description of the studied population

In the study, 1359 patients were included. Mean age was 51 (range 15–99) yr; 45% of patients were women (n=622). Emergence delirium in the PACU occurred in 64 (4.7%) patients.

The average length of delirium was 15 (20) min. All patients regained a normal cognitive status before their discharge from the PACU. In the agitated patients, the level of excitement according to the sedation–agitation scale was: 26% of dangerously agitated patients (n=17), 32% of very agitated patients (n=20) and 42% of agitated patients (n=27). The consequences were three attempts at self-extubation (4.7%), four catheter removals (both bladder and i.v. catheters) (6%), two injured patients (skin bruises) (3%) and three injuries to staff (4.5%). These cases of emergence delirium generally required 2–3 staff members to restrain and calm the patient (varying from 1–6). The suspected aetiologies were the existence of a tracheal tube (51.5%, n=34), pain (19.6%, n=13), anxiety (15%, n=10), overwhelming need to urinate despite the presence of an indwelling urinary catheter (4.5%, n=3), residual neuromuscular block (3%, n=2), acute urinary retention (3%, n=2) and intolerance to the oxygen catheter (3%, n=2).

Univariate analysis between agitated and non-agitated patients who had undergone general anaesthesia

Mean length of stay in the PACU was significantly longer for agitated patients [207 (171) min (range 50–960)] than in non-agitated patients [156 (146) min (range 10–1380)] (P=0.012). Average age was not significantly different in agitated patients compared with non-agitated patients [48 (range 15–78) yr vs 52 (range 15–99) yr]. The percentage of women in each group was similar (45% in agitated and 44% in non-agitated) but the ASA class distribution was different in agitated than in non-agitated patients (Table 2).

Table 2

Logistic regression: risk factors of emergence delirium in the PACU in general anaesthesia surgery patients. *Pearson chi-square; two-sample t-test; §mean (range or SD)

 Total (n=1359) Non-agitated (n=1295) (95.3%) Agitated (n=64) (4.7%) P-value 
Female 622 (45%) 593 (45%) 29 (45%) 0.940* 
Age 51 (15–99) yr§ 52 (15–99) yr§ 48 (15–78) yr§ 0.077 
Pre-existing illness 929 (68%) 897 (69%) 32 (50%) 0.001* 
Antidepressant drugs (>1 month) 315 (23%) 311 (24%) 4 (6%) 0.001* 
Neuroleptics (>1 month) 122 (9%) 122 (9%) 0 (0%) 0.010* 
Benzodiazepines (>1 month) 109 (8%) 107 (8%) 2 (3%) 0.139* 
Preoperative anxiety 84 (6%) 76 (5%) 8 (12%) 0.032* 
Recent surgery (<3 months) 90 (6%) 85 (5%) 5 (7%) 0.454* 
Recent surgery (<2 weeks) 76 (4%) 75 (4%) 1 (1%) 0.235* 
Length of stay before anaesthesia 3.2 (6.7) days§ 3.2 (6.8) days§ 2.2 (4.2) days§ 0.257 
Emergency surgery 140 (10%) 131 (10%) 9 (14%) 0.314* 
Preoperative benzodiazepine 333 (24%) 307 (23%) 26 (40%) 0.002* 
Preoperative hydroxyzine 1050 (77%) 1008 (77%) 42 (65%) 0.076* 
General anaesthesia procedure    0.006* 
    Propofol 490 (36%) 479 (37%) 11 (17%)  
    I.V. + halogenated 855 (63%) 803 (62%) 52 (81%)  
    Halogenated 14 (1%) 13 (1%) 1 (1%)  
    Desflurane 6 (0.7%) 6 (0.7%) 0 (0%)  
    Sevoflurane 521 (63%) 486 (63%) 35 (68%) 0.645* 
    Isoflurane 291 (35%) 275 (35%) 16 (31%)  
    Neuromuscular blockers 560 (41%) 521 (40%) 39 (61%) 0.001* 
ASA 
    I 425 (31%) 396 (30%) 29 (46%)  
    II 614 (45%) 589 (45%) 25 (40%)  
    III 293 (21%) 285 (22%) 8 (13%) 0.034* 
    IV 23 (1%) 23 (1%) 0 (0%)  
Length of surgery 76 (67) min§ 75 (67) min§ 104 (68) min§ 0.001 
Vascular surgery 159 (11%) 153 (11%) 6 (9%) 0.552* 
ENT 226 (16%) 218 (16%) 8 (12%) 0.362* 
Ophthalmology 142 (10%) 138 (10%) 4 (6%) 0.260* 
Abdominal 280 (20%) 253 (19%) 27 (42%) <0.001* 
Endoscopies 159 (11%) 158 (12%) 1 (1%) 0.010* 
Urology 169 (12%) 162 (12%) 7 (11%) 0.707* 
Breast 29 (2%) 24 (1%) 5 (7%) 0.001* 
External genitalia 23 (1%) 23 (1%) 0 (0%) 0.282* 
Thyroidectomy 79 (5%) 75 (5%) 4 (6%) 0.881* 
Length of stay in the PACU 158 (147) min§ 156 (146) min§ 207 (171) min§ 0.012 
Death 21 (1.5%) 19 (1.4%) 2 (3.1%) 0.293* 
 Total (n=1359) Non-agitated (n=1295) (95.3%) Agitated (n=64) (4.7%) P-value 
Female 622 (45%) 593 (45%) 29 (45%) 0.940* 
Age 51 (15–99) yr§ 52 (15–99) yr§ 48 (15–78) yr§ 0.077 
Pre-existing illness 929 (68%) 897 (69%) 32 (50%) 0.001* 
Antidepressant drugs (>1 month) 315 (23%) 311 (24%) 4 (6%) 0.001* 
Neuroleptics (>1 month) 122 (9%) 122 (9%) 0 (0%) 0.010* 
Benzodiazepines (>1 month) 109 (8%) 107 (8%) 2 (3%) 0.139* 
Preoperative anxiety 84 (6%) 76 (5%) 8 (12%) 0.032* 
Recent surgery (<3 months) 90 (6%) 85 (5%) 5 (7%) 0.454* 
Recent surgery (<2 weeks) 76 (4%) 75 (4%) 1 (1%) 0.235* 
Length of stay before anaesthesia 3.2 (6.7) days§ 3.2 (6.8) days§ 2.2 (4.2) days§ 0.257 
Emergency surgery 140 (10%) 131 (10%) 9 (14%) 0.314* 
Preoperative benzodiazepine 333 (24%) 307 (23%) 26 (40%) 0.002* 
Preoperative hydroxyzine 1050 (77%) 1008 (77%) 42 (65%) 0.076* 
General anaesthesia procedure    0.006* 
    Propofol 490 (36%) 479 (37%) 11 (17%)  
    I.V. + halogenated 855 (63%) 803 (62%) 52 (81%)  
    Halogenated 14 (1%) 13 (1%) 1 (1%)  
    Desflurane 6 (0.7%) 6 (0.7%) 0 (0%)  
    Sevoflurane 521 (63%) 486 (63%) 35 (68%) 0.645* 
    Isoflurane 291 (35%) 275 (35%) 16 (31%)  
    Neuromuscular blockers 560 (41%) 521 (40%) 39 (61%) 0.001* 
ASA 
    I 425 (31%) 396 (30%) 29 (46%)  
    II 614 (45%) 589 (45%) 25 (40%)  
    III 293 (21%) 285 (22%) 8 (13%) 0.034* 
    IV 23 (1%) 23 (1%) 0 (0%)  
Length of surgery 76 (67) min§ 75 (67) min§ 104 (68) min§ 0.001 
Vascular surgery 159 (11%) 153 (11%) 6 (9%) 0.552* 
ENT 226 (16%) 218 (16%) 8 (12%) 0.362* 
Ophthalmology 142 (10%) 138 (10%) 4 (6%) 0.260* 
Abdominal 280 (20%) 253 (19%) 27 (42%) <0.001* 
Endoscopies 159 (11%) 158 (12%) 1 (1%) 0.010* 
Urology 169 (12%) 162 (12%) 7 (11%) 0.707* 
Breast 29 (2%) 24 (1%) 5 (7%) 0.001* 
External genitalia 23 (1%) 23 (1%) 0 (0%) 0.282* 
Thyroidectomy 79 (5%) 75 (5%) 4 (6%) 0.881* 
Length of stay in the PACU 158 (147) min§ 156 (146) min§ 207 (171) min§ 0.012 
Death 21 (1.5%) 19 (1.4%) 2 (3.1%) 0.293* 

A pre-existing illness was significantly associated with a lower incidence of agitation (P=0.001). Thus, within agitated patients, 50% of patients had a known history of illness against 69% in non-agitated patients. Long-term medication with antidepressants and antipsychotic drugs was most often noted in non-agitated patients compared with agitated patients (24 and 9% against 6 and 0%, P=0.001 and P=0.010). No patient on neuroleptics developed an emergence delirium. No difference in the occurrence of emergent delirium was found in patients on long-term treatment with benzodiazepines (8% in non-agitated patients vs 3% in agitated patients, P=0.139).

Severe preoperative anxiety, noted as such in the patients' medical records, was positively correlated to the occurrence of agitation (12% in agitated patients vs 5% in non-agitated patients, P=0.032).

The interval between ward admission and surgery day was not significantly different in agitated than in non-agitated patients [2.2 (4.2) days vs 3.2 (6.8) days, P= 0.257]. Neither, a previous surgical procedure within the past 15 days or 3 months, nor emergency surgery had an impact on the occurrence of postoperative agitation.

In agitated patients, 40% of patients had received premedication by benzodiazepines, while only 23% did, in non-agitated patients (P=0.002). The proportion of patients who had been premedicated with hydroxyzine was comparable in the two groups.

Abdominal and breast surgery were significantly more frequent in agitated patients (42 and 7%) than in non-agitated patients (19 and 1%) (P<0.001 and P=0.001). Conversely, endoscopies were significantly less frequent in agitated than non-agitated patients (1% vs 12%, P=0.010). Surgery was longer in agitated patients [104 (68) min] than in non-agitated patients [75 (67) min] (P=0.001).

Patients in which anaesthesia was maintained by inhaled anaesthetics were more often agitated than those in which anaesthesia was maintained by propofol (P=0.006). There was no significant difference in agitation depending on the type of inhaled anaesthetics used. The use of neuromuscular blockers during surgery was significantly more frequent in agitated patients (61% vs 40%, P=0.001).

Postoperative mortality rate at 90 days, for all patients, was 1.4% (20 patients), with a mean age of 62 (range 42–79) yr. Mortality rate in agitated patients was 3%, 2 men of 52 and 69 yr of age, who died at Days 4 and 26 postop, respectively. The first died of a massive stroke after an emergency laparotomy for sigmoid perforation. The other, a 69-yr-old, died of multi-organ failure secondary to septic shock after postoperative peritonitis. Mortality rate in non-agitated patients was 1.3% (18 patients) with a mean age of 62 (range 42–79) yr. Death occurred on average around Day 13 (range 1–41 days). Univariate analysis of incidence of death did not show a significant difference of mortality rate at 90 days between the two groups (P=0.293). Moreover, length of stay in the PACU is significantly longer in agitated patients (P= 0.012).

Multivariate analysis of emergence agitation probability in general anaesthesia surgery patients

The univariate analysis between both groups showed that the incidence of emergence delirium was significantly related to 12 variables, of which 8 were more frequent in agitated patients (preoperative anxiety, preoperative medication by benzodiazepines, general anaesthesia procedure, use of neuromuscular blockers, ASA score, length of surgery, abdominal and breast surgery), and four were more often noted in non-agitated patients (a known history of illness, long-term use of antidepressants or neuroleptics, endoscopies) (Table 2).

Multivariate analysis by backward binary stepwise logistic regression found that of the 12 variables used in our model, six were significant in the onset of emergence delirium in the PACU (Table 3). Breast and abdominal surgery were the highest risk factors, with the risk of developing emergence agitation increased by 5 and 3 times (respectively, OR=5.190, 95% CI=1.422–18.947, P=0.013 and OR=3.206, 95% CI=1.262–8.143, P=0.014). Multivariate analysis confirmed that preoperative medication by benzodiazepines is a significant risk factor of emergence delirium in the PACU. Use of benzodiazepines before surgery nearly doubled the risk of emergence delirium in our study (OR=1.910, 95% CI=1.101–3.315, P=0.021). The length of surgery was a slight risk factor (OR=1.005, 95% CI=1.002–1.008, P=0.001) but the OR increased with the length of surgery. There was an exponential relationship between emergence delirium and duration surgery. A previous history of illness and long-term use of antidepressant drugs decrease the risk of emergence delirium by 0.5 and 0.2 times (respectively, OR=0.544, 95% CI=0.315–0.939, P=0.029 and OR=0.245, 95% CI=0.084–0.710, P=0.010) (Table 3).

Table 3

Logistic regression: risk factors of emergence delirium in the PACU in 1359 general anaesthesia surgery patients. P*, backward binary stepwise logistic regression, significant to include or exclude equal to 0.15. 95% CI, confidence interval of 95%

Factor Odds ratio 95% CI P*-value 
Breast surgery 5.190 1.422–18.947 0.013 
Abdominal surgery 3.206 1.262–8.143 0.014 
Preoperative medication with benzodiazepine 1.910 1.101–3.315 0.021 
Length of surgery 1.005 1.002–1.008 0.001 
Previous history of illness 0.544 0.315–0.939 0.029 
Long-term antidepressant treatment 0.245 0.084–0.710 0.010 
Factor Odds ratio 95% CI P*-value 
Breast surgery 5.190 1.422–18.947 0.013 
Abdominal surgery 3.206 1.262–8.143 0.014 
Preoperative medication with benzodiazepine 1.910 1.101–3.315 0.021 
Length of surgery 1.005 1.002–1.008 0.001 
Previous history of illness 0.544 0.315–0.939 0.029 
Long-term antidepressant treatment 0.245 0.084–0.710 0.010 

Univariate analysis of emergence delirium probability in general anaesthesia surgery patients excluding those with preoperative anxiety and long-term use of antidepressants, neuroleptics and benzodiazepines (Table 4)

Emergence delirium was significantly related to seven variables, of which six were more frequent in agitated patients (preoperative medication by benzodiazepines, anaesthesia procedure, use of neuromuscular blockers, surgery duration, abdominal and breast surgery), and one was more often noted in non-agitated patients (endoscopies).

Table 4

Logistic regression: risk factors of emergence delirium in the PACU in general anaesthesia surgery patients excluding preoperative anxiety and long-term treatment by benzodiazepines or antidepressants or neuroleptics. *Pearson chi-square; two-sample t-test; §mean (range or SD)

 Total (n=850) Non-agitated (n=799) (94%) Agitated (n=51) (6%) P-value 
Female 348 (41%) 326 (41%) 22 (43%) 0.742* 
Age 46 (00–00) yr§ 46 (00–00) yr§ 47 (00–00) yr§ 0.773 
Pre-existing illness 460 (54%) 435 (54%) 25 (49%) 0.457* 
Recent surgery (<3 months) 46 (5%) 42 (5%) 4 (7%) 0.430* 
Recent surgery (<2 weeks) 31 (3%) 30 (3%) 1 (2%) 0.507* 
Length of stay before anaesthesia 2.7 (6.3) days§ 2.7 (6.3) days§ 2.2 (4.3) days§ 0.495 
Emergency surgery 100 (11%) 92 (11%) 8 (15%) 0.374* 
Preoperative benzodiazepine 215 (25%) 195 (24%) 20 (39%) 0.018* 
Preoperative hydroxyzine 641 (75%) 609 (76%) 32 (62%) 0.092* 
GA 
    Propofol 299 (35%) 289 (36%) 10 (19%) 0.047* 
    I.V. + halogenated 544 (64%) 504 (63%) 40 (78%)  
    Halogenated 8 (1%) 7 (0.8%) 1 (2%)  
    Desflurane 5 (1%) 5 (1%) 0 (0%)  
    Sevoflurane 325 (63%) 298 (63%) 27 (67%) 0.718* 
    Isoflurane 183 (35%) 170 (36%) 13 (32%)  
    Neuromuscular blockers 330 (38%) 298 (37%) 32 (62%) <0.001* 
ASA    0.881* 
    I 386 (45%) 362 (45%) 24 (47%)  
    II 336 (39%) 316 (39%) 20 (39%)  
    III 119 (14%) 113 (14%) 6 (11%)  
    IV 7 (1%) 7 (1%) 0 (0%)  
Length of surgery 72 (67) min§ 72 (67) min§ 108 (68) min§ 0.001 
Vascular surgery 77 (9%) 72 (9%) 5 (9%) 0.848* 
ENT 175 (20%) 168 (21%) 7 (13%) 0.211* 
Ophthalmology 69 (8%) 68 (8%) 1 (2%) 0.096* 
Abdominal 185 (21%) 161 (20%) 24 (47%) <0.0001* 
Endoscopies 101 (11%) 100 (12%) 1 (2%) 0.024* 
Urology 100 (11%) 96 (12%) 4 (8%) 0.369* 
Breast 13 (1%) 10 (1%) 3 (5%) 0.009* 
External genitalia 15 (1%) 15 (1%) 0 (0%) 0.323* 
Thyroidectomy 57 (6%) 53 (6%) 4 (7%) 0.740* 
Length of stay in the PACU 146 (128) min§ 146 (128) min§ 205 (142) min§ 0.011 
Death 10 (1%) 8 (1%) 2 (4%) 0.060* 
 Total (n=850) Non-agitated (n=799) (94%) Agitated (n=51) (6%) P-value 
Female 348 (41%) 326 (41%) 22 (43%) 0.742* 
Age 46 (00–00) yr§ 46 (00–00) yr§ 47 (00–00) yr§ 0.773 
Pre-existing illness 460 (54%) 435 (54%) 25 (49%) 0.457* 
Recent surgery (<3 months) 46 (5%) 42 (5%) 4 (7%) 0.430* 
Recent surgery (<2 weeks) 31 (3%) 30 (3%) 1 (2%) 0.507* 
Length of stay before anaesthesia 2.7 (6.3) days§ 2.7 (6.3) days§ 2.2 (4.3) days§ 0.495 
Emergency surgery 100 (11%) 92 (11%) 8 (15%) 0.374* 
Preoperative benzodiazepine 215 (25%) 195 (24%) 20 (39%) 0.018* 
Preoperative hydroxyzine 641 (75%) 609 (76%) 32 (62%) 0.092* 
GA 
    Propofol 299 (35%) 289 (36%) 10 (19%) 0.047* 
    I.V. + halogenated 544 (64%) 504 (63%) 40 (78%)  
    Halogenated 8 (1%) 7 (0.8%) 1 (2%)  
    Desflurane 5 (1%) 5 (1%) 0 (0%)  
    Sevoflurane 325 (63%) 298 (63%) 27 (67%) 0.718* 
    Isoflurane 183 (35%) 170 (36%) 13 (32%)  
    Neuromuscular blockers 330 (38%) 298 (37%) 32 (62%) <0.001* 
ASA    0.881* 
    I 386 (45%) 362 (45%) 24 (47%)  
    II 336 (39%) 316 (39%) 20 (39%)  
    III 119 (14%) 113 (14%) 6 (11%)  
    IV 7 (1%) 7 (1%) 0 (0%)  
Length of surgery 72 (67) min§ 72 (67) min§ 108 (68) min§ 0.001 
Vascular surgery 77 (9%) 72 (9%) 5 (9%) 0.848* 
ENT 175 (20%) 168 (21%) 7 (13%) 0.211* 
Ophthalmology 69 (8%) 68 (8%) 1 (2%) 0.096* 
Abdominal 185 (21%) 161 (20%) 24 (47%) <0.0001* 
Endoscopies 101 (11%) 100 (12%) 1 (2%) 0.024* 
Urology 100 (11%) 96 (12%) 4 (8%) 0.369* 
Breast 13 (1%) 10 (1%) 3 (5%) 0.009* 
External genitalia 15 (1%) 15 (1%) 0 (0%) 0.323* 
Thyroidectomy 57 (6%) 53 (6%) 4 (7%) 0.740* 
Length of stay in the PACU 146 (128) min§ 146 (128) min§ 205 (142) min§ 0.011 
Death 10 (1%) 8 (1%) 2 (4%) 0.060* 

Multivariate analysis of emergence agitation probability in general anaesthesia surgery patients excluding those with preoperative anxiety and long-term use of antidepressants, neuroleptics and benzodiazepines (Table 5)

Multivariate analysis by backward binary stepwise logistic regression found that of the seven variables used in our model, four were significant in the onset of emergence delirium in the PACU (Table 5). Breast and abdominal surgery were the highest risk factors, with the risk of developing emergence agitation increasing respectively by 7 and 3.3 times (OR=7.024, 95% CI=1.692–29.156, P=0.007 and OR=3.376, 95% CI=1.775–6.418, P<0.0001). Multivariate analysis confirmed that preoperative medication by benzodiazepines is a significant risk factor of emergence delirium in the PACU. Use of benzodiazepines before surgery nearly doubled the risk of emergence delirium in our study (OR=1.838, 95% CI=1.002–3.519, P=0.050). The length of surgery was a slight risk factor (OR=1.004, 95% CI=1.000–1.008, P=0.035).

Table 5

Logistic regression: risk factors of emergence delirium in the PACU in general anaesthesia surgery patients excepting preoperative anxiety and long-term treatment by benzodiazepines or antidepressants or neuroleptics. P*, backward binary stepwise logistic regression, significant to include or exclude equal to 0.15. 95% CI, confidence interval of 95%

Factor Odds ratio 95% CI P*-value 
Breast surgery 7.024 1.692–29.156 0.007 
Abdominal surgery 3.376 1.775–6.418 <0.0001 
Preoperative medication with benzodiazepine 1.838 1.002–3.519 0.050 
Length of surgery 1.004 1.000–1.008 0.035 
Factor Odds ratio 95% CI P*-value 
Breast surgery 7.024 1.692–29.156 0.007 
Abdominal surgery 3.376 1.775–6.418 <0.0001 
Preoperative medication with benzodiazepine 1.838 1.002–3.519 0.050 
Length of surgery 1.004 1.000–1.008 0.035 

Discussion

Emergence delirium is an everyday problem in the PACU. In our population, the incidence of delirium was overall 4.7%. In a review of literature, we found an average of 3%.1 Of all agitated patients in our study, more than 50% were graded very agitated or dangerously agitated.

Different scales of agitation exist, such as the Riker agitation–sedation scale, the Richmond sedation–agitation scale, the motor activity assessment scale and the New Sheffield sedation scale. Studies have demonstrated that both the Riker and the Richmond scales have an excellent inter-rater reliability.5–8 The scales' advantages are their ease of use, especially in emergency situations, and their more precise evaluation of agitation compared with the Ramsay sedation scale, which rates essentially sedation. All of these scales were created for adults in intensive care. None has so far been established in the PACU. However, we feel that patients admitted to the PACU are comparable with patients admitted to the ICU.

The data were recorded in the PACU by nurses. Nurses were blinded to the study's objectives. Nonetheless, they were free to refer to preoperative evaluations, which may constitute a bias and a confounding factor. However, our study identified the same risk factors of emergence delirium (except for known history of illness), whether anxious patients were included or not in the analysis.

Probable aetiologies of emergence delirium found most often in our study include existence of a tracheal tube, pain, and anxiety. In awake patients, a tracheal tube can be stressing, and therefore should be removed as early as possible. In our study, where this was the cause of delirium, simple removal of the tube was sufficient to calm patients. Pain should be anticipated and treated preventively, as early as during surgery itself. However, in our study all patients had a pre-emptive analgesic protocol. More infrequent causes of delirium are the overwhelming need to urinate despite the presence of an indwelling urinary catheter, bladder distension and urinary retention. Better information should be given to patients before surgery concerning the use of various catheters (gastric tube, oxygen catheter, urinary catheter) to limit their surprise upon emergence. The use of ultrasound measurement at the patients' bedside gives a rapid diagnosis of urinary retention, thereby enabling rapid treatment.9

All PACU patients' oxygen saturation level was monitored according to recommendations. They received oxygen so as to maintain saturation ≥95%. None therefore developed acute and durable desaturation.

Only one study in adults compared the incidence of emergence delirium between patients anaesthetized with inhalational anaesthetics and those anaesthetized with propofol.10 In our study, the multivariate analysis did not suggest any difference between anaesthesia procedures, although the univariate analysis found more agitated patients anaesthetized with inhalational anaesthetics (62%) as compared with those anaesthetized with propofol (37%). No difference was observed between the different inhalational anaesthetics.

The residual neuromuscular block can be frightening. However, proper monitoring of neuromuscular block during surgery and in the PACU should reduce the risk of a residual neuromuscular block. Should a residual neuromuscular block be detected, patients can easily be antagonized.

It is important to define at-risk patients so as to best prevent emergence delirium. Our study has found that breast and abdominal surgery, and preoperative medication by benzodiazepines are risk factors of postoperative delirium. Length of surgery, while statistically significant, had no clinical impact except for long surgical procedure. On the other hand, a known history of illness and long-term treatment by antidepressants were found to be protective factors.

When anxious patients and those with long-term antidepressants, neuroleptics, or benzodiazepines are excluded the same risk factors are identified, with no protective factors.

Antidepressants were a protective factor of delirium, probably because they are anxiolytics and have a long half-life thereby covering the entire length of surgery. Previous history of illness was also a protective factor. This could be explained by the fact that these patients are more accustomed to a medical or hospital environment, and therefore less subject to anxiety.

Benzodiazepines surprisingly increased the odds of emergence delirium, as they were prescribed to alleviate anxiety and would have been expected to protect from delirium. However, they are known to have paradoxical effects such as irritability, aggressiveness or even confusion. Various reports find more frequent postoperative delirium in elderly patients on benzodiazepines, but did not study the incidence of delirium specially at emergence.11–13 A prospective study by Schor and colleagues14 found that benzodiazepines were protective of delirium in elderly hospitalized patients, whether for surgery or other reasons. In paediatrics, preoperative use of midazolam has variable effects on emergence delirium, certain studies having found an increased frequency of delirium,15,16 while others found it to be a protective factor.17 The exact role played by benzodiazepines cannot be inferred from our study, as their use was not randomized. They were more often chosen by anaesthetists as preoperative medication, when patients were considered to be particularly anxious.

Anxiety is a very subjective feeling, and may be perceived differently by various physicians. Preoperative anxiety was not assessed by structured interviews during the preoperative consultation. The evaluation in itself is known to modify the baseline anxiety level.18 Our study found no relationship between preoperative anxiety and postoperative delirium. Conversely, a relationship between anxieties in children who underwent surgery was linked to emergence delirium.19

Breast surgery may be linked to agitation, as the modification of corporal representation may create important anxiety.20 Abdominal surgery is known to be very painful. This may explain the higher incidence of agitation.21 Unfortunately, it was very often difficult to monitor the level of pain accurately during the emergence delirium episode. Moreover, treatment of pain made it impossible to evaluate pain as a risk factor of emergence delirium.

Emergence delirium is costly in several senses: in terms of morbidity, in human resources and on a financial level. Self-extubation and removal of catheters can lead to aspiration pneumonia or emergency surgery. As far as human resources are concerned, it can require up to six staff members to restrain an agitated patient. The precise number of persons necessary has not yet been established, but will be important to determine, especially at night when less people are present. Staff must be present on-site at all times, and while nurses are attending an agitated patient, other patients might be less closely watched, thereby increasing their anxiety. Besides these risks, there is also a risk of injury to staff. Finally, emergence delirium is probably costly but not studied, in that it necessitates an increase in pharmaceuticals and staff, and lengthens the stay in the PACU. In our study, there is a significant increase in length of stay in agitated patients.

In conclusion, based on this descriptive study, we feel that anxiety is at the heart of the problem. Few studies have focused on preoperative risk factors of postoperative delirium episodes in the PACU. Even though preoperative anxiety was not found to be a risk factor, it seems to be closely intricate with other factors. More stressful situations such as stressed patients, and breast or abdominal surgery patients, have a higher rate of delirium, while patients deemed calmer or less impressed by medical procedures are less subject to delirium. To reduce the occurrence and consequences of delirium episodes, one would need a structured preoperative evaluation of anxiety, and to take into account risk factors so as to treat them appropriately.

References

1
Rose
DK
Recovery room problems or problems in the PACU
Can J Anaesth
 , 
1996
, vol. 
43
 (pg. 
R116
-
28
)
2
O'Brien
D
Acute postoperative delirium: definitions, incidence, recognition, and interventions
J Perianesth Nurs
 , 
2002
, vol. 
17
 (pg. 
384
-
92
)
3
Lipov
EG
Emergence delirium in the PACU
Crit Care Nurs Clin North Am
 , 
1991
, vol. 
3
 (pg. 
145
-
9
)
4
Riker
RR
Picard
JT
Fraser
GL
Prospective evaluation of the sedation-agitation scale for adult critically ill patients
Crit Care Med
 , 
1999
, vol. 
27
 (pg. 
1325
-
9
)
5
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
, vol. 
166
 (pg. 
1338
-
44
)
6
Riker
RR
Fraser
GL
Simmons
LE
Wilkins
ML
Validating the sedation-agitation scale with the bispectral index and visual analog scale in adult ICU patients after cardiac surgery
Intensive Care Med
 , 
2001
, vol. 
27
 (pg. 
853
-
8
)
7
Simmons
LE
Riker
RR
Prato
B
Fraser
GL
Assessing sedation during intensive care unit mechanical ventilation with the bispectral index and the sedation agitation scale
Crit Care Med
 , 
1999
, vol. 
27
 (pg. 
1499
-
504
)
8
Brandl
KM
Langley
KA
Riker
RR
Dork
LA
Quails
CR
Levy
H
Confirming the reliability of the sedation-agitation scale administered by ICU nurses without experience in its use
Pharmacotherapy
 , 
2001
, vol. 
21
 (pg. 
431
-
6
)
9
Lamonerie
L
Marret
E
Deleuze
A
Lembert
N
Dupont
M
Bonnet
F
Prevalence of postoperative bladder distension and urinary retention detected by ultrasound measurement
Br J Anaesth
 , 
2004
, vol. 
92
 (pg. 
544
-
6
)
10
Patel
SS
Goa
KL
Desflurane. A review of its pharmacodynamic and pharmacokinetic properties and its efficacy in general anaesthesia
Drugs
 , 
1995
, vol. 
50
 (pg. 
742
-
67
)
11
Kudoh
A
Takase
H
Takahira
Y
Takazawa
T
Postoperative confusion increases in elderly long-term benzodiazepine users
Anesth Analg
 , 
2004
, vol. 
99
 (pg. 
1674
-
8
)
12
Marcantonio
ER
Juarez
G
Goldman
L
, et al.  . 
The relationship of postoperative delirium with psychoactive medications
JAMA
 , 
1994
, vol. 
272
 (pg. 
1518
-
22
)
13
Rogers
MP
Liang
MH
Daltroy
LF
, et al.  . 
Delirium after elective orthopaedic surgery: risk factors and natural history
Int J Psychiatry Med
 , 
1989
, vol. 
19
 (pg. 
102
-
21
)
14
Schor
JD
Levkoff
SE
Lipsitz
LA
, et al.  . 
Risk factors for delirium in hospitalized elderly
JAMA
 , 
1992
, vol. 
267
 (pg. 
827
-
31
)
15
Viitanen
H
Annila
P
Viitanen
M
Yli-Hankala
A
Midazolam premedication delays recovery from propofol-induced sevoflurane anesthesia in children 1-3 yr
Can J Anesth
 , 
1999
, vol. 
46
 (pg. 
766
-
71
)
16
Cohen
IT
Drewsen
S
Hannallah
RS
Propofol or midazolam do not reduce the incidence of emergence agitation associated with desflurane anaesthesia in children undergoing adenotonsillectomy
Paediatr Anaesth
 , 
2002
, vol. 
12
 (pg. 
604
-
9
)
17
Lapin
SL
Auden
SM
Goldsmith
LJ
Reynolds
AM
Effects of sevoflurane anaesthesia on recovery in children: a comparison with halothane
Paediatr Anaesth
 , 
1999
, vol. 
9
 (pg. 
283
-
6
)
18
Maltby
N
Mayers
MF
Allen
GJ
Tolin
DF
Anxiety sensitivity: stability in prospective research
J Anxiety Disord
 , 
2005
, vol. 
19
 (pg. 
708
-
16
)
19
Kain
ZN
Caldwell-Andrews
AA
Maranets
I
, et al.  . 
Preoperative anxiety and emergence delirium and postoperative maladaptive behaviors
Anesth Analg
 , 
2004
, vol. 
99
 (pg. 
1648
-
54
)
20
Burgess
C
Cornelius
V
Love
S
Graham
J
Richards
M
Ramirez
A
Depression and anxiety in women with early breast cancer: five year observational cohort study
Br Med J
 , 
2005
, vol. 
330
 pg. 
702
 
21
Manworren
RC
Paulos
CL
Pop
R
Treating children for acute agitation in the PACU: differentiating pain and emergence delirium
J Perianesth Nurs
 , 
2004
, vol. 
19
 (pg. 
183
-
93
)

Comments

0 Comments