Impact of postanesthesia care unit delirium on self-reported cognitive function and perceived health status: a prospective observational cohort study

The study protocol was approved by the local ethics committee at the Hamburg Chamber of Physicians (protocol No. 4782/Prof. Dr. M. Carstensen/02.09.2014). Written informed consent was obtained from all patients prior to participation. The study was registered at ClinicalTrials.gov (Identifier: NCT04168268).

This prospective, observational study was performed at a high-volume prostate cancer center in Northern Germany between November 2017 and October 2018. Male patients aged over 60 years who were admitted to the PACU after elective open radical retropubic prostatectomy (ORP) or robot-assisted radical prostatectomy (RARP) were eligible for study participation. Patients were required to be fluent in German to answer all questionnaires. The exclusion criteria were pre-existing neurological or cerebrovascular conditions, a history of dementia, or mild cognitive impairment. We chose a convenience sample of patients, who were admitted between Mondays and Thursdays on the day before surgery. Patients, who declined to participate, who were absent from the ward or who were prescheduled for postoperative transfer to the intensive care unit (ICU) were not eligible.

Preoperative baseline assessments were performed by one trained anesthesiologist following a standardized protocol. We used the Mini-Mental Status Examination (MMSE) to screen for dementia or mild cognitive impairment [16] and the Patient Health Questionnaire-9 (PHQ-9) to assess depressive symptoms [17]. We conducted baseline PHQ-9 screening, as symptoms of depression should be considered when interpreting delirium, subjective cognitive complaints, and quality of life. Depression has previously been proven to be a risk factor for POD, especially in older patients [18, 19], and it is associated with patient-reported cognitive failures [20, 21] and quality of life [22, 23]. The Cognitive Failures Questionnaire (CFQ) was used to evaluate the type and frequency of self-reported cognitive failures in everyday life [24]. It assesses errors in three areas: life perception, memory, and motor function. The CFQ consists of 25 items. The ratings (0–4) of the individual items are summed to obtain a total score between 0 and 100, with a higher score indicating more self-reported cognitive failures [24]. The Cognitive Reserve Index (CRI) questionnaire, which addresses education, profession, and leisure time activities as the three main sources of cognitive reserve, was administered to measure the quantity of cognitive reserve accumulated through a lifespan [25]. A complete list of assessments performed throughout the perioperative period is presented in Online Resource 1.

To screen for signs of delirium, a trained examiner interviewed all participants 15, 30, 45, and 60 min after arrival at the PACU. At each of these time points, the patients’ level of arousal was assessed using the Richmond Agitation and Sedation Scale (RASS) [26]. Patients with an RASS score of −4 (no response to voice but movement or eye opening in response to physical stimulation) or −5 (no response to voice or physical stimulation) were ineligible for delirium assessment and were re-evaluated 15 min later. If the patients showed a reasonable reaction (i.e., an RASS score between −3 and + 4), the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU) was performed. The CAM-ICU allows for delirium screening in less than 3 min [27, 28]. It comprises four domains that assess the following delirium criteria: (1) an acute onset of mental status change or a fluctuating course; (2) inattention; (3) disorganized thinking, and (4) an altered level of consciousness. A positive CAM-ICU score indicates signs of an acute change in mental status or fluctuating course (feature 1) accompanied by inattention (feature 2), and either disorganized thinking or an altered level of consciousness (feature 3 or 4) [27]. The CAM-ICU is an effective, validated and highly specific tool for delirium diagnosis and can be performed easily at the bedside [27]. Importantly, the CAM-ICU has a higher specificity than sensitivity for the detection of delirium during the recovery period [29]. Postanesthesia care unit delirium was defined as a positive CAM-ICU score together with an RASS score greater than or equal to −3 (any response to verbal stimulation). Participants were also asked to rate their pain perception on a numerical scale between 0 and 10 at the time of delirium screening. All delirium assessments in the PACU were performed by two trained examiners, who were both members of the research team (E.K., M.F.). Repeated assessments in one study participant were performed by one examiner.

Three months after surgery, we assessed physical and mental health status using the 36-item Short-Form Health Survey (SF-36). Self-reported cognitive failures in everyday life were evaluated with the CFQ. All study participants received questionnaires by mail. The SF-36 is a validated questionnaire that assesses eight health concepts. The eight domains can be categorized into two summary measures: physical and mental health component scores [30]. Higher scores indicate a better self-reported health status [31]. Raw SF-36 data from the completed questionnaires were transferred to an online analysis tool (Hogrefe Testsystem 5, Testzentrale, Göttingen, Germany) for further processing according to a predefined algorithm.

The choice of surgical technique (ORP or RARP) was based on individual risk factors, surgical considerations, oncological factors, and patient preference. General anesthesia was administered according to our institutional standard operating procedures. Sufentanil (0.3–0.7 μg/kg) and propofol (2–3 mg/kg) were used for induction of general anesthesia, followed by neuromuscular blockade with rocuronium (0.6 mg/kg) to facilitate endotracheal intubation. A gastric tube was inserted in all patients and prophylactic antiemetic medication was administered preoperatively (dexamethasone 4 mg). Intraoperatively, rocuronium was titrated under the guidance of neuromuscular blockade monitoring (train-of-four, TOF-Watch Organon; IntelliVue NMT module, Philips GmbH, Hamburg, Germany), and anesthesia depth was monitored using a bispectral index monitor (BIS™, Medtronic GmbH, Meerbusch, Germany). Sevoflurane-sufentanil was used for anesthesia maintenance to achieve an end-tidal sevoflurane concentration of 2.0 vol% (MAC 0.8–1.2). Normothermia was maintained using a forced-air warming system throughout the entire procedure. Patients who underwent RARP received peritoneal insufflation with carbon dioxide and were positioned at a 45-degree head-down tilt. Postoperative pain management included non-opioid medication (metamizole 1000 mg/100 ml) 30 min before emergence and every 4–6 h thereafter. During the PACU stay, piritramide 3.75–7.5 mg was administered intravenously when pain scores exceeded 3. Subsequent PACU management in all patients included frequent control of wound drains, postoperative urine output, blood gas analyses, and pain management as described above.

Details on data collection are presented in Online Resource 2. Continuous data are presented as medians with interquartile ranges (IQR), and categorical data are presented as frequencies with percentages. For group comparisons (no PACU delirium vs. PACU delirium), the Mann–Whitney U-test (continuous variables), Chi-square test, or Fisher’s exact test (categorical variables) were used as appropriate.

The association between PACU delirium and our outcome measures was assessed with three separate path diagrams, one for each outcome. For path analyses, we selected clinically relevant variables that did not fulfill the criteria for collinearity (Online Resource 2). The path diagrams were formed by clinical considerations about the dependencies of the variables. An important advantage of path diagrams over conventional regression methods is that mediating effects can be evaluated. Of main interest were the paths from PACU delirium to the outcomes, and the paths that lead into PACU delirium. Thereby, the effects of variables on either PACU delirium or the outcome measure can be distinguished in one model.

Additional mediation analyses were performed for associations of PACU delirium duration on each of our outcome measures (CFQ, SF-36 physical, and SF-36 mental component score). The following subgroups were formed to define and differentiate PACU delirium duration: (1) patients who were delirious at least at one of assessment points 2–4 (30, 45 or 60 min after arriving in the PACU; n=35), and (2) patients with delirium 15 mins after arriving in the PACU only (n=37).

For path analyses, a p-value of less than 0.017 was considered statistically significant (Bonferroni-corrected for multiplicity). We used IBM SPSS Statistics 24 (IBM Corporation, Armonk, New York), and STATA 16 (STATA Corporation, College Station, Texas) for statistical analyses. This manuscript adheres to the STROBE reporting guidelines for observational studies.

A total of 222 patients were enrolled and assessed for the development of PACU delirium. One patient died during the early postoperative period. The remaining patients were discharged to home after a median length of hospital stay of 6 days (IQR: 6–7). Two hundred and twenty-one subjects were available for follow-up after 3 months. Questionnaires for both endpoints, CFQ and SF-36, were returned by 178 patients (80.5% of survivors) and were included in the statistical analyses (Fig. 1). Detailed descriptive data stratified by response status are presented in Online Resource 3.

Patient’s characteristics divided by signs of PACU delirium are shown in Table 1. The median age of patients with PACU delirium was higher than that of patients without PACU delirium. The majority of participants had low perioperative risk and were classified as category I or II in the American Society of Anesthesiologists physical status classification system (89.6%, n = 199/222). There was no relevant imbalance of baseline psychometric properties between groups. Detailed information on psychometric scores at baseline is provided in Table 1.

Delirium signs were present in 32.4% of patients (n = 72/222). The incidence was highest 15 min after extubation and decreased over time, with the lowest delirium rates occurring 60 min after extubation. Details of the incidence of PACU delirium during the first hour after extubation are presented in Fig. 2. Hypoactive delirium was the predominant type of delirium in our study population. The majority of patients diagnosed with PACU delirium showed hypoactive features, whereas only one patient was in a hyperactive state 15 min upon arrival at the PACU.

Robot-assisted radical prostatectomy was performed in 51.8% (n = 115/222) and ORP in 48.2% (n = 107/222) of study participants. The surgical approach, RARP vs. ORP, did not differ significantly between patients with and without delirium (p = 0.840). The durations of both surgery and anesthesia were significantly longer in patients with PACU delirium (Table 2). There was no significant difference in the estimated blood loss or the amount of fluids administered between patients with and without delirium. The length of PACU stay and the dosage of analgesic and antiemetic drugs administered during the recovery period did not differ significantly between patients with and without PACU delirium. Table 2 shows the details of anesthesiologic and surgical management stratified by delirium status. The laboratory parameters are listed in Online Resource 4.

Cognitive failures at the 3-month follow-up were not significantly different between patients with (18 [IQR 8–24]) and without delirium (14 [IQR 6–22], p = 0.217). Patients diagnosed with PACU delirium scored lower in the physical component of the SF-36 after 3 months compared to patients without PACU delirium (50.6 [IQR 42.3–54.9] vs. 52.6 [IQR 47.3–55.5], p = 0.04). The mental health component scores did not differ between both groups (55.1 [IQR 49–57.7] vs. 55.6 [IQR 49.9–57.8], p = 0.927). Details of patient-reported outcomes at 3 months are presented in Table 1.

There was no significant association between PACU delirium and self-reported cognitive failures after controlling for age, PHQ-9, the CFQ baseline score, and additive therapy in our study population (Fig. 3a, Table 3). There was a statistically significant association between a higher preoperative CFQ sum score and more self-reported cognitive failures 3 months after surgery (p < 0.01). The effect of the PHQ-9 score on cognitive failures at 3 months was fully mediated by the number of preoperative cognitive failures (p < 0.01).

No significant association was found between PACU delirium and the SF-36 physical component score after controlling for age, nerve-sparing resection, type of surgery, and additive therapy. The direct effect of age on the physical component score was partially mediated by PACU delirium, although not reaching statistical significance (Fig. 3b, Table 3).

There was no significant association between PACU delirium and the SF-36 mental component score after controlling for age, nerve-sparing resection, additive therapy, and PHQ-9 score (Fig. 3c, Table 3). The PHQ-9 score at baseline was significantly associated with the mental component score at 3 months (p < 0.01). There was a statistical trend for the effect of additive therapy on the mental component score (p = 0.02).

In subgroup mediation analyses there was no statistically significant association of PACU delirium duration on cognitive failures and SF-36 physical and mental component scores. Details on the subgroup analysis are presented in Online Resource 5.

Several caveats should be considered when interpreting the results of this prospective observational study. We used the CAM-ICU, which has a higher specificity (98%) than sensitivity (28%) for delirium diagnosis in the PACU setting compared to a formal psychiatric evaluation based on DSM criteria [29]. Thus, the low sensitivity of the CAM-ICU when conducted in the recovery room may have led to underdiagnosis of PACU delirium. However, in contrast to evaluation based on DSM criteria, the CAM-ICU can be performed easily and rapidly at the bedside and is, therefore, more likely to be adopted into daily routines, making it a suitable tool for PACU delirium screening. At the time of data collection for our study, there was no screening tool available that was sensitive, specific to PACU delirium, and timesaving to use. In the meantime, Hight and colleagues have published a modified and extended version of the CAM-ICU assessment tool for delirium screening in the recovery room (CAM-PACU) [40]. The CAM-PACU score may have higher sensitivity as it includes additional criteria and may, therefore, be considered for PACU delirium screening in future studies [40].

Cognitive assessment of our study population was based on the CFQ, which is the most widely used tool for measuring patient-reported cognitive failures [41]. Importantly, self-reported cognitive function may not necessarily correlate with objective assessment [41]. We may have missed subtle changes in postoperative cognitive performance that might have been detected with an objective neuropsychological assessment. However, the two methods represent different but equally valuable approaches to evaluating cognitive function [41]. Thus, the use of subjective measures in conjunction with an objective neuropsychological assessment should be considered in future studies that investigate the impact of PACU delirium. Importantly, self-reported cognitive function represents one essential patient-centered outcome, which is why we chose this endpoint for our trial.

We did not assess delirium signs beyond 1 h after arrival in the PACU and follow-up was performed only once at 3 months. Therefore, we cannot comment on the continued trajectories of delirium signs and possible subsequent effects on cognitive function and perceived health status during the early postoperative period. Of all study participants, 32.4% screened positive for delirium at one or more time point during the first hour, with the highest delirium rates recorded early after extubation. The fact that the number of patients with delirium features decreased with time suggests that signs of PACU delirium resolve rapidly.

The timing of the follow-up in our study may have biased the results regarding the impact of PACU delirium on cognitive functions. By assessing delirium during the PACU period only, we may have missed single patients with persisting delirium beyond the recovery period. Inouye and colleagues published a model showing a biphasic course of cognitive trajectories after POD, characterized by an acute peak of cognitive dysfunction, subsequent recovery, and a gradual decline throughout 36 months. At 2 months, there was no significant difference between delirious and non-delirious patients [9]. The postulated recovery phase includes our follow-up time point (3 months after surgery). Therefore, it is possible that our finding of no adverse effects of PACU delirium was due to the dynamic course of cognitive impairment.

Our study population was homogeneous with relatively highly educated, high-functioning, elderly male patients without pre-existing cognitive impairment and with low perioperative risk undergoing elective radical prostatectomy. Thus, the generalizability of our findings might be limited and should be confirmed in more diverse samples.

The homogeneity of our study population, all of whom underwent highly standardized surgical interventions, anesthesia, and perioperative care within a single clinical setting, is also a remarkable strength of our study design as it ensures almost identical timing and ambient conditions of PACU delirium screening of all patients.

The analysis of self-reported health status in addition to cognitive outcomes adds to important current knowledge on the intermediate- and long-term impact of PACU delirium.

Furthermore, our data illustrate the reproducibility and feasibility of PACU delirium screening during the first hour after extubation. Since PACU delirium may be a predictor of subsequent POD and its associated adverse outcomes [6], we strongly support systematic delirium screening in the PACU setting to identify patients who might be particulary susceptible to prolonged episodes of delirium. In a previous observational trial, our research group found an incidence of PACU delirium of 48% in patients over 60 years of age. Interestingly, the majority of patients had received midazolam for premedication, which might explain the higher incidence of PACU delirium [42].