Perioperative care in elderly cardiac surgery patients.

INTRODUCTION: Surgery is an extreme physiological stress for the elderly. Aging is inevitably associated with irreversible and progressive cellular degeneration. Patients above 75 years of age are characterized by impaired responses to operative stress and a very narrow safety margin. AIM: To evaluate perioperative complications in patients aged ≥ 75 years who underwent cardiac surgery in comparison to outcomes in younger patients.
MATERIAL AND METHODS: The study was conducted at the Silesian Centre for Heart Diseases in Zabrze in 2009-2014 after a standard of perioperative care in seniors was implemented to reduce complications, in particular to decrease the duration of mechanical ventilation and reduce postoperative delirium. The study group included 1446 patients.
RESULTS: The mean duration of mechanical ventilation was 13.8 h in patients aged ≥ 75 years and did not differ significantly compared to younger patients. In-hospital mortality among seniors was 3.8%, a value significantly higher than that observed among patients younger than 75 years of age. Patients aged ≥ 75 years undergoing cardiac surgery have significantly more concomitant conditions involving other organs, which affects treatment outcomes (duration of hospital stay, mortality).
CONCLUSIONS: The implementation of a standard of perioperative care in this age group reduced the duration of mechanical ventilation and lowered the rate of postoperative delirium.

Introduction

Chronological and biological age differ significantly from each other. Chronological age is the parameter used in clinical practice. It has been widely accepted to divide the elderly subjects into four age categories: young-old age (65–74 years), middle-old age (75–84 years), oldest-old age (≥ 85 years), and longevity (≥ 90 years). No age is a contraindication for surgery if the clinical aim is not only to prolong life. The most important motivation for treatment is the preservation or improvement of the quality of life.

The philosophy of surgical treatment in seniors focuses on the rapid restoration of all physiological functions immediately after the procedure, which reduces the complication rates and the duration of hospital stay [1].

The occurrence of postoperative complications leads to a cascade of events resulting in disability, dependence, reduced quality of life, increased mortality, and a dramatic increase in treatment costs [2].

Seniors are characterized by different physiology, which justifies the individualization of the treatment process. One typical feature is reduced – hepatic perfusion, which leads to impaired drug metabolism. Furthermore, the number of drug-binding receptors is reduced, and the response to medications is thus delayed. Renal dysfunction extends the duration of drug elimination. Due to low albumin levels, the free drug fraction is increased, resulting in increased drug transport to the central nervous system (CNS). The volume of distribution is reduced due to the reduction of circulating blood volume by 20%, which is associated with higher drug levels. Seniors are also more susceptible to medications as a result of increased concentrations of water-soluble drugs (due to lower water content) and reduced concentrations of fat-soluble drugs (due to increased sequestration of anesthetics in the adipose tissue and later their extended release) [3]. Diastolic dysfunction is a typical feature of the aging heart [4].

The effect of general anesthesia on the brain is an increasingly debated issue. The hypothesis that general anesthesia contributes to the progression of neurodegeneration has not been confirmed. However, some studies indicate that excessively deep anesthesia may affect the rate of cognitive dysfunction and postoperative delirium. In addition, increasing evidence suggests that cognitive dysfunction is aggravated by the inflammatory process induced by anesthesia and surgery [5-7].

There is no single ideal method of providing anesthesia to this patient group. An optimal approach seems to consist in multimodal anesthesia with short-acting opioids, propofol, an inhaled anesthetic, and a myorelaxant that does not release histamine [8].

Aim

The aim of the study was to evaluate perioperative complications in patients aged ≥ 75 years who underwent cardiac surgery in comparison to outcomes in younger patients operated on during the same period.

Material and methods

We evaluated patients ≥ 75 years of age who underwent cardiac surgery at the Silesian Centre for Heart Diseases in Zabrze from the implementation of a standard of perioperative care in this age group in October 2009 until October 2014. The study group included 1446 patients. We excluded patients with acute aortic dissection who were subjected to emergency surgery. Patients ≥ 75 years of age were compared with patients below 75 years of age operated on during the same period (6913 patients).

The risk of renal, respiratory, and cardiac failure [9, 10] was evaluated during preoperative anesthesiology consultations.

For premedication, the seniors received melatonin (5 mg orally) on the evening before the surgery and another dose 1 h before being transferred to the operating theatre. In the study group, the induction of anesthesia was obtained with intravenous etomidate (0.2–0.3 mg/kg), sufentanil 50–100 μg (0.5–1 μg/kg), and rocuronium (1 mg/kg). In patients with hypotension, ketamine (1–2 mg/kg) was used instead of propofol. Anesthesia was maintained using continuous infusion of propofol (100–200 mg/h) and sufentanil (0.3–1 μg/kg/h) along with fractionated doses of rocuronium (0.1 mg/kg). Inhaled sevoflurane was also used at the same time, and was also administered during cardiopulmonary bypass (CPB).

Hemodynamic monitoring included ECG as well as direct blood pressure and central venous pressure measurements. Swan-Ganz catheters were placed in patients with low left ventricular ejection fraction (LVEF < 30%) or pulmonary hypertension and in patients undergoing complex procedures. We recorded deep body temperature and monitored urine output. We continuously monitored oxygen saturation and frequently measured arterial blood gases with lactate levels during the anesthesia. Mixed venous blood gases were measured if hemodynamic instability occurred. In patients with generalized atherosclerosis, carotid artery stenosis, or a history of stroke, we monitored regional CNS oxygen saturation with near-infrared spectroscopy (NIRS), which is the current standard [11]. An important aspect of management was to maintain normothermia both during the surgery and in the postoperative period. We also established the acceptable level ranges for blood glucose (6.5–8.5 mmol/l) and hemoglobin (≥ 5 mmol/l). Fluid balance was strictly monitored, and blood loss was replaced with colloid fluids. We maximally restricted the filling of the CPB circuit. During the initial phase of CPB, the patients received 250 mg of methylprednisolone. During CPB, the mean arterial pressure was kept at 70 mm Hg, and cardiac output was maintained at ≥ 2.4 l/m2. A lung-sparing ventilation strategy was used (tidal volume 6 ml/kg, positive end-expiratory pressure ≥ 5 cm H2O, FiO2 50%). Pharmacological and mechanical support of cardiovascular function was provided in accordance with the principles of goal-oriented therapy. The goal was defined as a normal lactate level and normal venous blood oxygen saturation, and, in patients monitored using a Swan-Ganz catheter, the goal was to achieve a cardiac index of at least 2.4 l/m2.

Analgesia was used from admission to the stay in the postoperative unit (morphine in fractionated doses, fixed doses of paracetamol). Antiemetic therapy included dexamethasone and ondansetron. The oxygen level in inspired air was adjusted based on transcutaneous oxygen saturation monitoring. For mechanical ventilation, the SIMV mode was initially used, followed by the bilevel or a proportional mode and, finally, continuous positive airway pressure (CPAP) with pressure support. The intubation tube was removed when the extubation criteria were met. Respiratory rehabilitation and patient mobilization strategies were initiated immediately after extubation. Patients received oral fluids two hours after extubation, and enteral feeding was initiated several hours later. Fluid balance was planned individually.

In younger patients, the approach to anesthesia and early postoperative treatment was decided by the anesthesiologist in charge of anesthesia and the physician on duty in the postoperative unit.

Patients aged ≥ 75 years were characterized by lower body mass index values, higher risk scores (EuroSCORE I), higher Canadian Cardiovascular Society (CCS) and New York Heart Association (NYHA) classes, and longer durations of hospital stay in comparison to younger patients (Tab. I). Among the seniors, there were more women, patients in NYHA class 3 or 4, patients with an unstable course of ischemic heart disease, left main coronary artery disease, diabetes, hypertension, preoperative renal failure, pulmonary disease, and carotid artery disease (Tab. II).

Tab. I

Statistical analysis of intergroup differences in selected parameters: age, body mass index (BMI), risk evaluated using EuroSCORE I, severity of angina according to the Canadian Cardiovascular Society (CCS) classification, severity of heart failure according to the New York Heart Association (NYHA) classification, preoperative left ventricular ejection fraction (LVEF), duration of cardiopulmonary bypass, duration of mechanical ventilation, and duration of hospital stay

Parameter	≥ 75 years		< 75 years		P-value
	Mean	SD	Mean	SD
Age	78.2	3.0	61.7	9.4	–
BMI	27.5	4.1	28.0	4.4	< 0.001
LVEF	49.2	9.8	49.8	10.2	0.057
EuroSCORE I	7.66	2.39	4.28	2.60	< 0.001
CCS	2.30	1.02	2.21	1.03	0.005
NYHA	2.08	0.86	1.78	0.95	< 0.001
Duration of cardiopulmonary bypass	115.7	51.4	114.3	50.1	0.493
Duration of mechanical ventilation	13.8	37.2	13.4	48.6	0.782
Duration of hospital stay	9.36	6.81	8.16	7.34	< 0.001

SD – standard deviation.

Tab. II

Statistical analysis of differences in demographic parameters (gender) and preoperative characteristics – severity of heart failure according to the New York Heart Association (NYHA) classification, severity of ischemic heart disease according to the Canadian Cardiovascular Society (CCS) classification, unstable disease course, left main coronary artery disease, previous myocardial infarction, diabetes, hypercholesterolemia, hypertension, renal dysfunction/failure, chronic pulmonary disease, previous stroke, history of transient ischemic attack (TIA), concomitant carotid artery disease (critical stenoses), and peripheral arterial disease (critical stenoses/occlusion) between seniors and younger patients

Parameter	≥ 75 years (N = 1446)		< 75 years (N = 6909)		P-value
	n	%	n	%
Female gender	681	47.1	2097	30.4	< 0.001
Male gender	765	52.9	4812	69.7	< 0.001
CCS class 4	146	10.1	633	9.2	0.288
NYHA class 3 or 4	426	29.5	1390	20.1	< 0.001
Unstable disease course	889	61.5	3773	54.6	< 0.001
Left main coronary artery disease	341	23.6	1472	21.3	0.061
Previous Q-wave infarction	602	41.6	2780	40.2	0.341
Diabetes	507	35.1	2046	29.6	< 0.001
Glucose intolerance	59	4.1	233	3.4	0.210
Diabetes (on oral therapy)	251	17.4	918	13.3	< 0.001
Diabetes (on insulin therapy)	197	13.6	886	12.8	0.435
Hypercholesterolemia	813	56.2	4110	59.5	0.024
Hypertension	1189	82.2	5323	77.0	< 0.001
Renal failure	330	22.8	704	10.2	< 0.001
Chronic pulmonary disease	183	12.7	571	8.3	< 0.001
Previous stroke	85	5.9	354	5.1	0.269
History of TIA	40	2.8	126	1.8	0.026
Carotid artery disease	166	11.5	530	7.7	< 0.001
Peripheral arterial disease	249	17.2	1163	16.8	0.750

A comparison of procedures performed in both patient groups is presented in Table III.

Tab. III

Statistical analysis of intergroup differences with regard to the operations performed (OPCAB – off-pump coronary artery bypass surgery, CABG – coronary artery bypass surgery, TAVI – transcatheter aortic valve implantation, complex surgery – valve replacement or repair with concomitant coronary artery bypass surgery, urgent surgery – hospitalization required until the time of the surgery, emergency surgery – surgery required within 24 h from admission)

Parameter	≥ 75 years (N = 1446)		< 75 years (N = 6909)		P-value
	n	%	n	%
OPCAB	666	46.1	3271	47.3	0.389
CABG	148	10.2	1152	16.7	< 0.001
Single valve surgery	291	20.1	1262	18.3	0.106
TAVI	74	5.1	35	0.5	< 0.001
Surgery involving two or more valves	59	4.1	377	5.5	0.038
Complex surgery	194	13.4	652	9.4	< 0.001
Ascending aortic aneurysm surgery	14	1.0	164	2.4	0.001
Urgent surgery	847	58.6	3549	51.3	< 0.001
Emergency surgery	62	4.3	210	3.0	0.019
Use of cardiopulmonary bypass	706	48.8	3607	52.2	0.021

Statistical analysis

Data were collected and initially evaluated in a spreadsheet. Verified data were transferred to the statistical software which was used for the proper analysis. Normal distribution of quantitative variables was checked using the Shapiro-Wilk test, while the non-parametric Mann-Whitney U test was used for non-normal variable distribution. For qualitative variables, a two-sided χ2 test with the Yates correction was used. The results are presented in tables showing mean values and standard deviations for quantitative variables and numbers with percentages for qualitative variables. P < 0.05 was considered statistically significant. Statistical analysis was performed using the Statistica data analysis software system, version 10.1 (StatSoft, Inc.).

Results

Table IV shows early postoperative complications in both study groups. Mediastinal revision was more common in older patients, as was low cardiac output syndrome requiring high catecholamine doses and mechanical support with an intra-aortic balloon. Other complications were also more frequently diagnosed in older patients, including pulmonary complications, acute renal failure (defined as serum creatinine level ≥ 200 μmol/l), type I neurological complications, and abdominal organ dysfunction (intestinal ischemia/necrosis, acute cholecystitis, acute pancreatitis, gastrointestinal bleeding). Atrial fibrillation and delirium were more common in older patients. Transfusion of more than 4 units of packed red blood cells was also more frequent in this group. The overall complication rate was higher in older patients (17.8%), and in-hospital mortality in this group was 3.8%.

Tab. IV

Statistical analysis of intergroup differences with regard to postoperative complications (reoperation – a surgical revision of the mediastinum required due to excessive drainage, cardiac tamponade, graft or valve dysfunction, or exploratory surgery due to unexplained cardiovascular decompensation; low cardiac output syndrome defined as the need to use two catecholamines in doses above 10 μg/kg/min each or epinephrine in a dose above 0.1 μg/kg/min or as a cardiac index below 2 l/min/m2 in two consecutive measurements; IABP – intra-aortic balloon counterpulsation, perioperative infarction – documented by ECG and biochemical testing, AF – atrial fibrillation; pulmonary complications – postoperative occurrence of pneumothorax, hemopericardium requiring drainage, pneumonia, pulmonary edema, ARDS, or hypodynamic respiratory failure; delirium – sudden change of mental status, impaired consciousness or attention, chaotic thinking; type I neurological complications – ischemic or hemorrhagic stroke, TIA, coma, anoxic encephalopathy; renal failure – serum creatinine level above 200 μmol/l or the need for renal replacement therapy; abdominal organ dysfunction – intestinal ischemia/necrosis, acute cholecystitis, acute pancreatitis, gastrointestinal bleeding)

Parameter	≥ 75 years (N = 1446)		< 75 years (N = 6909)		P-value
	n	%	n	%
Mechanical ventilation > 48 h	40	2.8	154	2.2	0.254
Reoperation	137	9.5	518	7.5	0.013
Excessive mediastinal drainage	76	5.3	295	4.3	0.113
Cardiac tamponade	26	1.8	110	1.6	0.654
Low cardiac output syndrome	579	40.0	1932	28.0	< 0.001
Epinephrine dose > 0.1 μg/kg/min	85	5.9	264	3.8	< 0.001
IABP	72	5.0	229	3.3	0.003
Perioperative infarction	12	0.8	53	0.8	0.934
AF	424	29.3	1306	18.9	< 0.001
Pulmonary complications	28	1.9	90	1.3	0.083
Delirium	116	8.0	240	3.5	< 0.001
Type I neurological complications	49	3.4	112	1.6	< 0.001
Renal failure	133	9.2	317	4.6	< 0.001
Abdominal organ dysfunction	36	2.5	64	0.9	< 0.001
Transfusion of > 4 units of packed red blood cells	326	22.5	906	13.1	< 0.001
Complications overall	257	17.8	744	10.8	< 0.001
Postoperative unit mortality	38	2.6	76	1.1	< 0.001
In-hospital mortality	55	3.8	106	1.5	< 0.001

The rate of prolonged mechanical ventilation, defined as patient dependence on the ventilator lasting more than 48 h, was the same in both groups (Tab. V). Gender and preoperative factors had no effect on the duration of mechanical ventilation (Tab. VI).

Tab. V

The effect of preoperative parameters – age, body mass index (BMI), left ventricular ejection fraction (LVEF), risk evaluated using EuroSCORE I, severity of angina according to the Canadian Cardiovascular Society (CCS) classification, severity of heart failure according to the New York Heart Association (NYHA) classification) and duration of cardiopulmonary bypass on the need of prolonged (> 48 h) postoperative mechanical ventilation in patients ≥ 75 years of age

Parameter	Mechanical ventilation > 48 h
	No		Yes		P-value
	Mean	SD	Mean	SD
Age	78.2	3.0	78.7	2.5	0.353
BMI	27.5	4.1	26.5	4.0	0.110
LVEF	49.3	9.7	46.0	11.5	0.046
EuroSCORE I	7.62	2.36	9.03	3.19	< 0.001
CCS	2.30	1.01	2.14	1.22	0.341
NYHA	2.07	0.86	2.35	0.68	0.053
Duration of cardiopulmonary bypass	113.7	49.0	159.4	77.6	< 0.001

SD – standard deviation.

Tab. VI

The effect of preoperative parameters (gender, New York Heart Association (NYHA) class 3 or 4 of heart failure, severe ischemic heart disease (Canadian Cardiovascular Society (CCS) class 4), unstable disease course, left main coronary artery disease, previous myocardial infarction, diabetes, hypercholesterolemia, hypertension, renal dysfunction/failure, chronic pulmonary disease, previous stroke, history of transient ischemic attack (TIA), concomitant carotid artery disease (critical stenoses), and peripheral artery disease (critical stenoses/occlusion)) on the need of prolonged (> 48 h) postoperative mechanical ventilation in patients ≥ 75 years of age

Parameter	Mechanical ventilation > 48 h
	No		Yes		P-value
	(N = 1406)		(N = 40)
	n	%	n	%
Female gender	657	46.7	24	60.0	0.134
Male gender	749	53.3	16	40.0	0.134
CCS class 4	140	10.0	6	15.0	0.437
NYHA class 3 or 4	409	29.1	17	42.5	0.097
Unstable disease course	862	61.3	27	67.5	0.530
Left main coronary artery disease	331	23.5	10	25.0	0.980
Previous Q-wave infarction	589	41.9	13	32.5	0.305
Diabetes	497	35.4	10	25.0	0.236
Glucose intolerance	59	4.2	0	0.0	0.359
Diabetes (on oral therapy)	249	17.7	2	5.0	0.060
Diabetes (on insulin therapy)	189	13.4	8	20.0	0.338
Hypercholesterolemia	796	56.6	17	42.5	0.107
Hypertension	1155	82.2	34	85.0	0.798
Renal failure	320	22.8	10	25.0	0.887
Chronic pulmonary disease	180	12.8	3	7.5	0.451
Previous stroke	83	5.9	2	5.0	0.919
History of TIA	39	2.8	1	2.5	0.700
Carotid artery disease	162	11.5	4	10.0	0.963
Peripheral arterial disease	238	16.9	11	27.5	0.125

The mean duration of mechanical ventilation in our patients ≥ 75 years of age was shorter by more than 10 h compared to the mean value in a similar group of patients analyzed in 2003–2008 (Tab. VII).

Tab. VII

The duration of postoperative mechanical ventilation and the rate of postoperative delirium in patients ≥ 75 years of age who underwent cardiac surgery at the Silesian Centre for Heart Disease in Zabrze in the years 2003–2008

Duration of mechanical ventilation (≥ 75 years)
Period [years]	N	Mean	SD
2003–2008	593	24.2	55.3
Delirium (≥ 75 years)
Period [years]	n		%
2003–2008	96		15.96

SD – standard deviation.

The rate of postoperative delirium in patients ≥ 75 years of age operated on due to cardiac disease in 2003–2008 was 15.7%, i.e., nearly twice as high as in the present study group (Tab. VII).

Discussion

The implementation of the Early Recovery After Surgery (ERAS) approach, used in general surgical, orthopedic, and other surgical units, has introduced some unrest into the monotonous standard of perioperative care in patients undergoing cardiac surgery. The idea was born and developed in Scandinavia, from where it later emanated to other parts of the world [12]. It seems that this philosophy of perioperative care may have an effect on the reduction of complication rates in all areas of surgery. This is even more likely due to the current trend to operate on older and older patients who tend to stay for a longer time in the postoperative unit, which results in postoperative delirium and increased severity of dementia, ultimately annihilating much of the initial effect of the surgery.

Years ago, some emotions in cardiac surgery were generated by the fast-track standard, but ultimately it focused on early extubation in low-risk patients.

The aging of society is a problem in developed and developing countries [13]. Patients referred for cardiac surgery also tend to be increasingly old. Frailty is a feature of the multidimensional syndrome associated with aging, characterized by a loss of physiological reserve and deficit accumulation [14]. As a result, these patients are helpless when trauma occurs.

In the present study, patients aged 75 years and above constituted 21% of the overall population undergoing surgery. For premedication, we used melatonin, which, according to various authors, restores the normal sleep and wake cycle, thus contributing to a reduction in the rate of delirium. This therapy was continued in the early postoperative period [15]. We avoided benzodiazepine use throughout the perioperative period. Statin treatment was not discontinued before the surgery, and statins were initiated in patients who had not previously received these drugs [16].

Short-acting drugs were used for multimodal, intravenous, and inhaled anesthesia. We also used sevoflurane, which is well known for its preconditioning properties.

Hemodynamic monitoring was performed according to the accepted standard, which was not modified in the evaluated age group. In the case of low cardiac output syndrome after the surgery, monitoring was supplemented with cardiac output measurements using the thermodilution method. Venous blood oxygen saturation and the lactate level were measured in order to enable early detection of impending hemodynamic destabilization. When treating heart failure, we were fully aware of a high likelihood of concomitant left ventricular diastolic dysfunction, which is characteristic for the evaluated age group. The treatment of choice for this type of heart failure is the administration of milrinone combined with norepinephrine. Fluid therapy in patients with diastolic heart failure is a real challenge. Adequate cardiac filling requires higher pressures, but the stiff myocardium is characterized by low end-diastolic volume. Low cardiac output syndrome is defined as the need to use two catecholamines in doses above 10 μg/kg/min each or epinephrine in a dose above 0.1 μg/kg/min or as a cardiac index below 2 l/min/m2 in two consecutive measurements. Such a condition was observed in 40% of the seniors undergoing surgery. Mechanical cardiac support was used in 5% of patients aged 75 years and above.

Modern sedation facilitates patient cooperation with the medical personnel, as it provides comfort, while allowing patients to communicate their needs. In order to achieve this condition, effective analgesia should be provided, nausea and vomiting should be avoided, and the ventilator’s operation should be synchronized with the patient’s spontaneous breathing. All patients who required deeper sedation and were ventilated for more than 24 h were evaluated daily to determine their readiness to be weaned from mechanical ventilation [17]. Despite many concomitant conditions and risks present in our seniors, the duration of postoperative mechanical ventilation did not differ between the groups. The only modifiable factor that affects the duration of postoperative mechanical ventilation is cardiopulmonary bypass, particularly its duration. Our results are consistent with often presented data indicating benefits of avoiding CPB in high-risk patients [18]. When evaluating the duration of mechanical ventilation in two different periods (2003–2008 versus 2009–2014), one should take into account the advances in surgical techniques, the increasing surgeon experience, and the introduction of protective mechanical ventilation. It seems, however, that the observed significant reduction in the duration of mechanical ventilation is also associated with the implementation of a standard of perioperative care for seniors.

Acute renal failure was treated in accordance with the standard approach introduced many years ago [19]. For more than 5 years, renal replacement therapy in the early postoperative period has been conducted under local anticoagulation using citrate [20].

An issue of ongoing debate in the medical literature is the optimal hemoglobin level during the perioperative period in elderly patients. On the one hand, every transfused unit of packed red blood cells is associated with an increase in mortality risk; on the other hand, anemia can cause acute postoperative renal failure, strokes, and respiratory failure [21]. In our study, indications for packed red blood cell transfusions were individualized, taking into account factors other than the hemoglobin level. The need for transfusion was considered in the context of the lactate level, venous blood oxygen saturation, and, in the perioperative period, also patient participation in rehabilitation. Despite the restrictive indications for transfusion, the need to use more than 4 units of packed red blood cells occurred in 22.55% of seniors and was significantly more frequent than in the case of younger patients.

Early postoperative delirium is associated with an increased risk of complications, prolonged duration of hospital stay, and increased mortality [22]. In our seniors, delirium was more than twice as frequent as in our younger patients. However, we were able to reduce the rate of delirium in comparison to the period before the implementation of the standard of perioperative care in seniors in 2009. In 2003–2008, postoperative delirium was observed in 15.7% of patients aged ≥ 75 years, as compared to 8% of patients at a similar age in the present study group.

Conclusions

Patients aged ≥ 75 years who undergo cardiac surgery have significantly more concomitant conditions involving other organs, which affects treatment outcomes (duration of hospital stay, mortality). The implementation of a standard of perioperative care in this age group reduced the duration of mechanical ventilation and lowered the rate of postoperative delirium.