Does Apolipoprotein E Genotype Increase Risk of Postoperative Delirium?

Abstract

Objectives

To determine whether Apolipoprotein E (ApoE) is associated with postoperative delirium incidence, severity, and duration in older patients free of dementia at baseline.

Design, Setting, Participants

We examined 557 non-demented patients age ≥70 undergoing major non-cardiac surgery enrolled in the Successful Aging after Elective Surgery (SAGES) Study.

Measurements

We considered three ApoE measures: ε2, ε4 carriers vs. non-carriers, and a three-category ApoE measure. Delirium was determined using the Confusion Assessment Method (CAM) and chart review. We used generalized linear models to estimate the association between ApoE and delirium incidence, severity (peak CAM Severity [CAM-S] score), and days.

Results

ApoE ε2 and ε4 was present in 15% and 19% respectively, and postoperative delirium occurred in 24%. Among patients with delirium, the mean peak CAM-S score was 8.0 (standard deviation 4), with most patients experiencing one or two delirium days (51% or 28%, respectively). After adjusting for age, sex, surgical procedure, and preoperative cognitive function, ApoE ε4 and ε2 carrier status were not associated with postoperative delirium: RR for ε4=1.0, 95% confidence interval (CI) 0.7-1.5 and RR for ε2=0.9, 95% CI 0.6-1.4. No association between ApoE and delirium severity or number of delirium days was observed.

Conclusions

In older surgery patients free of dementia, our findings do not support the hypothesis that the ApoE genotype does not confer either risk or protection in postoperative delirium incidence, severity, or duration. Thus, an important genetic risk factor for Alzheimer's Disease does not affect risk of delirium.

Introduction

Delirium is an acute alteration in cognition characterized by fluctuations in attention, disorganized thinking, and level of consciousness.^1,2 Delirium after major surgery is associated with longer length of hospitalization,³ greater postoperative complications,⁴ and higher rates of discharge to nursing homes.⁵ Additionally, patients with post-surgical delirium have high inhospital mortality (4-17%),^2,3 and this mortality risk remains elevated over a 12 month period.⁵ Despite growing epidemiological evidence of its associated poor outcomes, our understanding of the pathophysiology of delirium remains superficial.² Understanding mechanisms is important for identifying patients at highest risk of delirium and designing better prevention and treatment strategies.^6,7

One major hypothesis that has emerged from the epidemiological and basic research characterizes delirium as an unmasking of an underlying but as yet asymptomatic cognitive decline.⁸ Furthermore, an association between delirium and dementia is well described.⁹ Apolipoprotein E (ApoE) ε4 is considered a risk factor for Alzheimer's Disease (AD) and has been hypothesized as a genetic risk factor for delirium. The literature examining the relationship of ApoE with delirium, however, has reported mixed results. Some studies report an association between ApoE ε4 and increased delirium incidence,¹⁰ others report an association with delirium days,¹¹ while still others observe no relationship at all.^12,13 The majority of this work has been conducted in surgical patients,^10,12-15 with some in ICU¹¹ and general medical populations.^16,17 If the ApoE genotype is in-fact associated with delirium, then preventive efforts could target those patients at highest-risk for delirium.

Dementia may confound the relationship between ApoE and delirium, potentially resulting in the inconsistent findings in previous studies due to differential inclusion of patients with dementia. Our study aims to clarify the relationship between ApoE and postoperative delirium incidence, severity, and duration in a large sample of older adults undergoing major non-cardiac surgery who are free of any clinical evidence of preoperative dementia. Based on their effects in AD, we hypothesize that the ApoE ε4 allele will be associated with increased delirium incidence, higher severity, and longer duration; and the ApoE ε2 allele will be associated with decreased delirium incidence, severity, and duration.

Methods

Study Sample

The Successful AGing after Elective Surgery (SAGES) Study is a prospective observational study focused on furthering our understanding of the multifactorial delirium syndrome. Its primary goal is to investigate the contribution of delirium to long-term cognitive and functional decline. The SAGES study design and methodology has been previously published.¹⁸ Briefly, participants are enrolled from elective surgery rolls, and are assessed at baseline prior to surgery, and every day during their hospital stay.

There were 560 patients aged ≥70 years without recognized dementia or history of delirium scheduled for elective major surgery who were enrolled in SAGES. The major surgeries included: cervical, lumbar, or thoracic laminectomy; total hip or knee replacement; open abdominal aortic aneurysm repair; lower extremity vascular bypass; or colectomy. Exclusion criteria included: 1) active delirium on initial cognitive testing (using the Confusion Assessment Method [CAM]);¹ 2) hospitalization 3 months prior to study enrollment; 3) terminal conditions (e.g., advanced cancer, or receiving palliative care); 4) heavy alcohol consumption (to avoid confounding by delirium tremens); 5) legal blindness or severe deafness; 6) clinically evident dementia; 7) non-English speaking; and 8) history of schizophrenia or psychosis.

Dementia screening was conducted sequentially using three methods: initial review of the medical record, direct questions during telephone recruitment, and assessment of capacity and cognitive testing during the baseline enrollment interview. Patients were excluded from study participation if any of the following criteria for dementia were met: 1) diagnosis of dementia documented in the medical record, 2) diagnosis of dementia reported by the patient during telephone recruitment, 3) capacity assessment failed during the informed consent process, or 4) a score <69 or its education-adjusted equivalent on the Modified Mini-Mental (3MS)¹⁹ test at the baseline interview. Subsequently, all enrolled patients had their baseline cognitive status adjudicated by a clinical consensus panel using results from detailed baseline assessments including neuropsychological testing to classify individuals as: (1) no evidence of cognitive impairment or dementia; (2) cognitive impairment, no dementia (CIND); or (3) dementia. Patients who were enrolled but subsequently adjudicated by the consensus panel to have dementia at baseline were excluded from our analysis (n=6; see Figure 1 for details).

Figure 1.

Number of participants excluded based on sequential dementia screening steps in the Successful Aging after Elective Surgery (SAGES) Study

Among the 560 individuals enrolled without baseline dementia, 557 (99%) participants had information on ApoE genotype and at least two assessments to define delirium from postoperative day one until hospital discharge. Compared with the 557 participants included in the analysis, the three excluded individuals were not statistically different (p>.05) in age, sex, educational attainment, race, type of surgery, and preoperative cognitive function.

Study Measures

Apolipoprotein E. Phlebotomy was performed on the entire cohort at baseline. To determine ApoE genotype, DNA was extracted from whole blood using a previously described technique,²⁰ which yields high quantities of purified DNA of relatively high molecular weight that can be amplified using polymerase chain reaction (PCR) and restriction enzyme digestion. DNA was extracted, allele specific PCR assays were conducted in the Brigham Research Assay Core, and ApoE genotyping was determined from the Partners Center for Personalized Medicine. The genotype frequencies were in Hardy-Weinberg equilibrium (χ² = 2.44, df = 3, p ≈ 0.49). To consider both the potential effects of ApoE ε4 and ε2, we constructed several alternative formulations of ApoE genotype: ε4 allele carriers (in the absence of an ε2 allele); ε2 allele carriers (in the absence of an ε4 allele); and a three-category variable (ε2ε2 or ε3ε2; ε3ε3; ε4ε3 or ε4ε4 -- ε4ε2 was excluded because this genotype includes one risk allele [ε4] and one protective allele [ε2] and their combined effect is unclear) based on previous work.²¹

Delirium

Postoperative delirium was assessed through a structured assessment²² performed daily throughout hospitalization for the scheduled surgery. Delirium was determined using a combination of results from the daily interviews using the CAM and a validated chart review method, described below.

CAM delirium

CAM, the primary method for determining delirium, was rated by trained SAGES field staff. CAM delirium required evidence of an acute change in mental status or fluctuating course, inattention, and either disorganized thinking or an altered level of consciousness.¹ The CAM has high specificity (89%) and sensitivity (94%) compared with delirium ratings by geriatric psychiatrists.²³

Chart review delirium

The secondary method for determining delirium was a validated chart review method^24,25, which was used to detect delirium during times when the CAM could not be administered (e.g., the middle of the night). All chart review diagnoses were adjudicated by at least two experts, and any disagreements were resolved through consensus.

Final delirium determination

Patients were considered to be delirious if delirium was present by either the CAM or the chart review method on one or more postoperative days; otherwise, patients were considered non-delirious. The combination of interview and chart methods is considered optimal for identification of delirium.²⁵

Delirium Severity and Delirium Days

To further explore the relationship between ApoE and postoperative delirium, we considered delirium severity and number of delirium days among the patients with postoperative delirium. Delirium severity was quantified using the CAM Severity (CAM-S) score, as previously described.²⁶ The CAM-S long form is the sum of ten CAM items (range 0-19, 19 most severe), which includes fluctuating course, inattention, disorganized thinking, altered level of consciousness, disorientation, memory impairment, perceptual disturbances, psychomotor agitation, psychomotor retardation, and sleep-wake disturbance, each scored 0-absent, 1-present, mild, 2-present, severe, except for fluctuating course, which is scored 0-absent, 1-present. CAM-S scores were obtained for each post-surgical day. For our analysis, we used the peak CAM-S score from all available CAM-S scores among the delirious patients. Number of delirium days was determined by summing the number of postoperative days the patient was considered delirious by either the CAM or chart review methods.

Covariates

We examined covariates that have been associated with ApoE and postoperative delirium, including age, sex, surgical procedure, and preoperative cognitive functioning. Surgical procedures were categorized into three types: (1) orthopedic (total hip replacement or revision, single or bilateral; total knee replacement or revision, single, or bilateral; lumbar laminectomy; cervical laminectomy); (2) vascular: lower extremity bypass surgery; abdominal aortic aneurysm repair [open procedure, not endovascular]; thoracoabdominal aortic aneurysm repair), and (3) gastrointestinal (open or laparoscopic colectomy). Preoperative cognitive functioning was measured by a composite variable, the General Cognitive Performance (GCP), described previously.^27,28 The GCP was created from the SAGES battery of commonly used neuropsychological measures,²⁷ which were selected to reflect domains considered vulnerable to the impact of delirium. GCP demonstrates excellent psychometric properties with little floor and ceiling effects spanning a broad range of values²⁷ and was scaled to the US older adult population using the Aging, Demographics, and Memory Study (ADAMS).²⁸ Our analysis considered the GCP as a continuous variable.

Statistical Analysis

We estimated generalized linear models (GLM) with a log link and binomial error term to assess the association (unadjusted and adjusted relative risks [RR]) between ApoE and postoperative delirium incidence. For severity and days of delirium, we used generalized linear identity-link models. We conducted our analysis for three outcomes: postoperative delirium incidence, severity, and days. The latter two outcomes were examined only among patients who developed postoperative delirium. For all three outcomes, separate unadjusted models examined postoperative delirium risk by three parameterizations of ApoE: ApoE ε2 carrier status, ApoE ε4 carrier status, and a three-ApoE genotype categorization (described above). We examined these relationships further by adjusting for age, sex, type of surgery, and preoperative GCP (adjusted models). Poisson models were used for delirium-days analyses. Prior to analyzing the full SAGES study sample, power analysis was conducted to determine the sample size required to observe a 12% effect of ApoE ε4 on delirium (effect size: 12% absolute difference in risk of incident delirium between ε4 carriers and non-ε4 carriers, assuming type-I error 0.05). All analyses were conducted in SAS Version 9.3, Cary NC. Informed consent for study participation was obtained from all subjects according to procedures approved by the institutional review boards of Beth Israel Deaconess Medical Center and Brigham and Women's Hospital, the two surgical sites, and Hebrew SeniorLife, the study coordinating center, all located in Boston, MA.

Sensitivity Analysis

To test the possibility that either exclusion of patients with dementia or inclusion of patients with CIND may have influenced the lack of association between ApoE and delirium, we conducted two sets of post-hoc sensitivity analyses. The first sensitivity analysis, which considered the impact of exclusion of patients with dementia, considered two scenarios which re-introduced dementia patients into the cohort: 1) Add back the 51 patients excluded from our sample (including those not enrolled in SAGES) because of dementia (see Figure 1) and assume a maximally plausible prevalence of 40% ApoE ε4 carriers and 70% incidence of postoperative delirium based on the literature^29,30, and 2) Increase the prevalence of dementia to that observed in ADAMS³¹ by adding 145 patients with dementia (20% prevalence) to the SAGES sample and again assume the maximally plausible prevalence of 40% ApoE ε4 carriers and 70% incidence of delirium. The second sensitivity analysis excluded SAGES participants with evidence of CIND (n=38). CIND diagnosis, made by expert panel, was based on a detailed neurocognitive battery assessed at baseline. Unadjusted GLM models were rerun using the simulated data (for sensitivity analysis 1) and the available data (for sensitivity analysis 2) to obtain the alternative RRs. See eAppendix for details.

Results

Table 1 reports the clinical characteristics of our study sample. On average, our sample was older but not exceptionally old, was highly educated, and had a higher than U.S. average preoperative GCP score (Table 1). Most individuals had a score of 3 (indicating severe systemic disease) for the American Society of Anesthesiology (ASA) class, a classification system assessing the patient's physical state prior to surgery. Slightly more than half of the study sample was women and most underwent orthopedic surgery with fewer vascular and gastrointestinal surgeries. Among the participants who developed postoperative delirium (24%), the peak CAM-S score indicated moderate to high delirium severity, with most patients experiencing delirium on one or two postoperative days.

Table 1.

Clinical Characteristics of the Study Sample

Characteristic	Mean ± SD or N (%)
Age (mean ± SD)	76.6 ± 5.2
Female, N (%)	324 (58)
Educational attainment (years)	15.0 ± 3.0
Non-White or Hispanic, N (%)	42 (8)
Type of Surgery, N (%)
Orthopedic	451 (81)
Vascular	35 (6)
Gastrointestinal	71 (13)
General cognitive performance (externally scaled, mean ± SD)a	57.6 ± 7.3
American Society of Anesthesiology class, N (%)b
1	2 (0)
2	203 (37)
3	347 (62)
4	5 (1)
Any postoperative delirium, N (%)	132 (24)
Confusion Assessment Method-Severity score (long form, mean ± SD)c,d	8.0 ± 3.6
Total days with delirium, N (%)c
1	67 (51)
2	37 (28)
3+	28 (21)

Notes: Participant N=557. SD: standard deviation; ApoE: Apolipoprotein E.

^aGeneral cognitive performance, a composite measure of neuropsychological measures reflecting cognitive domains vulnerable to delirium.²¹

^bAmerican Society of Anesthesiology class, a classification system assessing the patient's physical state prior to surgery.

^cAmong participants with postoperative delirium (n=132).

^dConfusion Assessment Method-Severity score range 0-19, 19 most severe.

Table 2 reports the ApoE parameter characteristics of the study sample. There were slightly fewer ApoE ε2 carriers than ApoE ε4 carriers. Among our three ApoE genotype categories, about two-thirds had an ε3ε3 genotype, one-fifth had an ε3ε4 or ε4ε4 genotype, and the fewest proportion had ε2ε2 or ε3ε2 genotype (Table 2).

Table 2.

Apolipoprotein E Parameters of the Study Sample

ApoE parameter	N (%)
ApoE genotypes
ε2ε2	4 (1)
ε3ε3	358 (64)
ε3ε2	78 (14)
ε4ε2	9 (2)
ε4ε3	104 (18)
ε4ε4	4 (1)
ApoE ε2 carriera (vs. non-ε2 carrier)	82 (15)
ApoE ε4 carrierb (vs. non-ε4 carrier)	108 (19)
ApoE categoryc
ε2ε2 or ε3ε2	82 (15)
ε4ε3 or ε4ε4	108 (20)
ε3ε3	358 (65)

Notes: Participant N = 557. ApoE: Apolipoprotein E.

^aIn the absence of an ε4 allele

^bIn the absence of an ε2 allele

^cExcludes ε4ε2 (n=9)

Table 3 shows the incidence and RR of postoperative delirium among ApoE ε2 carriers (vs. non-carriers), ApoE ε4 carriers (vs. non-carriers), and the three ApoE genotype categories. Unadjusted and adjusted models for each of the three ApoE parameterizations indicated no association between ApoE and postoperative delirium. ApoE ε2 carriers had a non-significant lower risk of postoperative delirium compared to non-ε2 carriers, and ApoE ε4 carriers had a non-significant higher risk of incident postoperative delirium relative to their non-ε4 counterparts. Among the 132 patients who developed postoperative delirium, non-significant associations between the three ApoE parameterizations and CAM-S scores were observed (Table 4). Among participants with postoperative delirium, all models yielded non-significant associations between ApoE and number of delirium days (Table 4). Altogether, our findings suggest an absence of an association between ApoE and postoperative delirium incidence, severity, and days.

Table 3.

Generalized Linear Log-Link Models Predicting Postoperative Delirium by Apolipoprotein E

Parameters	Delirium Incidence (%)	Unadjusted Model		Adjusted Modela
		RR	(95% CI)	RR	(95% CI)
ApoE ε2 carrier status
ε2 carrier (vs. non-ε2 carrier)	20	0.8	(0.5-1.3)	0.9	(0.6-1.4)
ApoE ε4 carrier status
ε4 carrier (vs. non-ε4 carrier)	24	1.0	(0.7-1.5)	1.0	(0.7-1.5)
ApoE categories
ε2ε2 or ε3ε2	20	0.8	(0.5-1.3)	0.9	(0.5-1.4)
ε4ε3 or ε4ε4	24	1.0	(0.7-1.4)	1.0	(0.7-1.5)
ε3ε3	25	Reference		Reference

Notes: Participant N = 557. RR: relative risk; CI: confidence interval; ApoE: Apolipoprotein E.

Maximum-likelihood-estimation derived Wald chi-square with 1 df for the ApoE carrier status and categorical variables in all models.

^aAdjusted Model: Adjusted for age, sex, type of surgery, and preoperative general cognitive performance using generalized linear log-link regression.

Table 4.

Generalized Linear Identity-Link Models Predicting Postoperative Delirium Severity and Delirium Days by Apolipoprotein E Parameterizations among Patients with Postoperative Delirium

	CAM-S Score				# Days with Delirium
	Unadjusted Model		Adjusted Modela		Unadjusted Model		Adjusted Modela
Parameters	β	p-value	β	p-value	β	p-value	β	p-value
ApoE ε2 carrier status
ε2 carrier (vs. non-ε2 carrier)	1.7	0.08	1.6	0.10	0.2	0.63	0.1	0.69
ApoE ε4 carrier status
ε4 carrier (vs. non-ε4 carrier)	−0.7	0.38	−0.6	0.41	0.3	0.30	0.2	0.55
ApoE categories
ε2ε2 or ε3ε2	1.5	0.11	1.4	0.14	0.3	0.49	0.2	0.62
ε4ε3 or ε4ε4	−0.5	0.57	−0.4	0.58	0.4	0.25	0.2	0.50
ε3ε3	Reference		Reference		Reference		Reference

Notes: Participant N = 132. ApoE: Apolipoprotein E; CAM-S: Confusion Assessment Method Severity; B: Beta estimate (defined as points on the CAM-S score or number of days with delirium)

Maximum-likelihood-estimation derived Wald chi-square with 1 df for the ApoE carrier status and categorical variables in all models.

^aAdjusted Model: Adjusted for age, sex, type of surgery, and preoperative general cognitive performance using generalized linear identity-link models; since these are linear models, the beta estimates represent CAM-S points (for the first two columns) and delirium days (for the last two columns).

Sensitivity Analyses

For sensitivity analysis 1, scenario 1 (including patients with dementia at 8% prevalence, as observed in SAGES) resulted in a non-significant association between ApoE ε4 and postoperative delirium, while scenario 2 (including patients with dementia at 20% prevalence) yielded a significant association between ApoE ε4 and delirium. Sensitivity analysis 2, which excluded enrolled individuals with CIND, had little impact on our results, and did not substantively alter either the magnitude or statistical significance of our results. See eAppendix for additional details.

Discussion

This study examined the association between ApoE polymorphisms and postoperative delirium incidence, severity, and days in non-demented older adults undergoing major non-cardiac surgery, and we observed no significant associations. Contrary to our hypotheses, the ApoE ε4 allele was not a risk factor for postoperative delirium, and the ApoE ε2 allele was not protective against postoperative delirium. Our results demonstrate that an important risk factor for AD did not affect risk of delirium in our non-demented sample. A significant corollary of this conclusion is that the cognitive decline and dementia that have been reported following delirium in non-demented individuals may not represent an “unmasking” of preclinical AD.

The absence of an association between ApoE and delirium onset is consistent with some studies but incongruent with others. Our findings differ from those of three reported studies: one conducted in a non-cardiac surgery population,¹⁰ one in a medical and hip fracture population,¹⁶ and one in a medical and surgical population.¹⁴ Subsequently a meta-analysis, which included many of the studies above, also reported an association between ApoE and delirium.¹⁶ Additionally, there was one study conducted in ICU patients,¹¹ that found an association between ApoE ε4 carrier status and delirium days but not delirium incidence. Conversely, our findings agree with the lack of an association between ApoE and delirium reported in 5 studies, two conducted in medical populations,^12,13 one in a cardiac surgery population,¹⁴ one in a stroke population,¹⁷ and one in patients undergoing major non-cardiac surgery.¹⁵

One possible explanation for the widely divergent findings of the relationship between ApoE genotype and delirium is variation in study sample characteristics, specifically the inclusion or exclusion of patients with dementia. Notably, most of the above cited studies did not conduct detailed cognitive evaluations of the participants, and thus, we are not able to make accurate determinations of the prevalence of dementia in the study samples. Many of the studies enrolled medical, hip fracture, or ICU patients, who were already sick at the time of presentation, precluding assessment of baseline cognitive function. One study that enrolled elective surgical patients¹⁰ completed only a brief cognitive screening test (the Telephone Interview for Cognitive Status). A unique strength of our study was the careful assessment and exclusion of patients with dementia through a multi-stage process, culminating with a full neurocognitive assessment with adjudication by an expert panel. Moreover, our sensitivity analyses suggest that the exclusion of patients with dementia is unlikely to explain the absence of an association between ApoE and delirium in the SAGES sample. However, the inclusion of patients with dementia could explain the previously reported associations between ApoE and delirium observed in the literature,^10,14,16 particularly if the prevalence of dementia is 20% or higher in these study samples.

Dementia is an established risk factor for delirium.³² There is also a growing literature that among patients without dementia, those who develop delirium may be at increased risk of cognitive decline or dementia in the future. This suggests that the development of delirium may be an indicator of a dysfunctional response in the brain, which may represent an underlying vulnerability to subsequent dementia. Most studies that have examined the relationship between dementia and postoperative delirium have not characterized dementia subtypes; therefore it remains unknown whether delirium is preferentially associated with AD or other forms of dementia. Since ApoE ε4 is the most penetrant genetic risk factor known for AD, the absence of an association between ApoE ε4 and delirium observed in our study suggests, but does not explicitly test, the possibility that delirium and dementia may not originate from the same pathological process. Therefore, the associations observed between delirium and dementia are not attributable to the strongest genetic risk factor for AD. Ultimately, this suggests delirium is unlikely to represent simply an unmasking of pre-clinical dementia. We are confident in our results given that our sample size was substantial and our study sufficiently powered to detect a clinically significant effect attributable to ε4 (See Methods), if it was present.

Our current study has a number of strengths that underscore its innovation. SAGES represents the first study conducted in a large patient sample who were free of dementia at baseline. Enrolling patients scheduled for major non-cardiac surgery enabled us to carefully characterize their baseline cognitive status, exclude individuals with dementia, and ultimately clearly distinguish between risk factors for delirium and dementia. These strengths distinguish our study from previous studies examining the relationship of ApoE and delirium. Another strength of our study is the use of state-of-the art delirium measures, including delirium incidence, severity, and duration. The consistent findings across all three outcomes underscore the robustness of our results and strengthen our conclusion of a lack of an association between ApoE and delirium.

We also note some study limitations. The restricted enrollment of non-demented older adults along with the high education of the SAGES sample reflect findings that may not be generalizable to the entire population of older adults. While acknowledging this threat to generalizability, this restriction enabled us to conduct a more pristine analysis of the relationship between ApoE and postoperative delirium in the absence of dementia. Our sample of non-demented adults aged ≥70 may also represent a population with high reserve, which may protect against the detrimental effects of ApoE ε4 on dementia and delirium. We also enrolled from two academic medical centers in a single geographical region, and had relatively low minority representation. Our findings should be reproduced in other settings with more diverse samples.

In summary, in this study of older adults without dementia undergoing major elective non-cardiac surgery, we found no association between ApoE genotype and delirium incidence, severity, or duration. Our findings do not support the role of ApoE genotype in conferring protection or risk in the development of postoperative delirium. We conclude that in our non-demented older surgical sample, an important genetic risk factor for AD does not affect risk of delirium. Our results are consistent with the possibility that delirium is not merely an unmasking of preclinical AD and that efforts to address delirium may represent a fruitful avenue to prevent the post-operative cognitive decline that is frequently observed in older adults.