Postoperative delirium: A tragedy for elderly cancer patients

Abstract

In this editorial, we comment on the article by Hu et al entitled “Predictive modeling for postoperative delirium in elderly patients with abdominal malignancies using synthetic minority oversampling technique”. We wanted to draw attention to the general features of postoperative delirium (POD) as well as the areas where there are uncertainties and contradictions. POD can be defined as acute neurocognitive dysfunction that occurs in the first week after surgery. It is a severe postoperative complication, especially for elderly oncology patients. Although the underlying pathophysiological mechanism is not fully understood, various neuroinflammatory mechanisms and neurotransmitters are thought to be involved. Various assessment scales and diagnostic methods have been proposed for the early diagnosis of POD. As delirium is considered a preventable clinical entity in about half of the cases, various early prediction models developed with the support of machine learning have recently become a hot scientific topic. Unfortunately, a model with high sensitivity and specificity for the prediction of POD has not yet been reported. This situation reveals that all health personnel who provide health care services to elderly patients should approach patients with a high level of awareness in the perioperative period regarding POD.

Key Words: Delirium; Anesthesia; Neurocognitive dysfunction; Postoperative cognitive dysfunction; Prevention; Risk management; Synthetic minority oversampling technique; Postoperative delirium; Elderly patients; Abdominal cancer

Core Tip: Postoperative delirium (POD) is a clinical complication with severe adverse consequences that can lead to death, especially in elderly patients. POD can occur at any time after surgery until hospital discharge. Predicting and preventing the disease among the most important clinical goals as the pathophysiology is not fully understood and effective treatment is not available. With this objective, many tools for assessment of delirium have been validated and various models have recently been developed with the help of machine learning using known POD risk factors.

INTRODUCTION

Delirium or acute confusional syndrome is an acute neurocognitive dysfunction characterized by impaired mental abilities such as attention, cognition, and awareness. The diagnostic criteria for delirium are given in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), which states explicitly that the disease occurs acutely and progresses with fluctuations (Table 1)[1]. In the DSM-5, delirium is categorized into different types by etiologic differences, such as substance intoxication and withdrawal delirium and drug-induced delirium. Delirium is further subdivided into hypoactive, hyperactive, and mixed forms by the type of psychomotor disturbances. As the names suggest, there is decreased activity in cognitive functioning in patients with the hypoactive form and the opposite is the case in those with the hyperactive form. In mixed delirium, hypoactive and hyperactive clinical features coexist for short periods[2,3]. Although it is not included in the diagnostic criteria, perceptual disturbances (hallucination and delusion), emotional disturbances, and an impaired sleep-wake cycle are among the clinical features of patients with delirium[4].

Table 1 Definition of delirium by the Diagnostic and Statistical Manual of Mental Disorders, 5^th edition[1].

Diagnostic criteria
A disturbance in attention (i.e. reduced ability to direct, focus, sustain, and shift attention) and awareness (reduced orientation to the environment)
The disturbance develops over a short period of time (usually hours to a few days), represents a change from baseline attention and awareness, and tends to fluctuate in severity during the course of a day
An additional disturbance in cognition (e.g., memory deficit, disorientation, language, visuospatial ability, or perception)
The disturbances in criteria A and C are not better explained by another preexisting, established, or evolving neurocognitive disorder and do not occur in the context of a severely reduced level of arousal, such as coma
There is evidence from the history, physical examination, or laboratory findings that the disturbance is a direct physiological consequence of another medical condition, substance intoxication or withdrawal (i.e. due to a drug of abuse or to a medication), or exposure to a toxin, or is due to multiple etiologies

POSTOPERATIVE DELIRIUM

Definition

Although delirium can occur before surgery, it often occurs within 7 d after surgery; therefore, the term postoperative delirium (POD) is more commonly used. POD is included in a group of perioperative neurocognitive disorders along with delayed neurocognitive recovery, postoperative neurocognitive disorders, and mild/major neurocognitive decline[5]. Delayed neurocognitive recovery can be defined as a decline in cognitive performance within 30 d after surgery and has been reported to occur in 15%-50% of surgery patients, especially the elderly. As most patients survive and recover fully, the name “delayed neurocognitive recovery” has replaced “early postoperative cognitive dysfunction.” Postoperative neurocognitive disorders refers to cognitive impairment that can occur between the first month and the first year after surgery. Cognitive impairments that occur 1 year after surgery are called mild/major neurocognitive impairments. The description no longer included “postoperative” as the impairment cannot be directly related to either anesthesia or the surgical procedure after 1 year[6]. To avoid this complex confusion of concepts, the International Perioperative Cognition Nomenclature Working Group proposed using “perioperative neurocognitive disorders” for all cognitive changes associated with surgery and anesthesia[7].

Incidence of POD

The increase in life expectancy and the concomitant growth in the elderly population has led to an increasing number of elderly patients undergoing major abdominal surgery for many indications, including cancer. Elderly patients are often frail and have comorbidities, which may expose them to a greater risk than younger patients of unfavorable outcomes such as POD in the perioperative period. Studies of POD in elderly patients have reported conflicting incidence rates. Generally, the overall incidence of POD in adults is 2.5%-4.2%[8]. In elderly patients the incidence varies widely between 4%-53% in different surgery types in different studies depending mainly on patient age, the differences in diagnostic and screening tools, and the training of the medical staff performing the assessment[9]. Regardless of age, POD may occur in 13.7% of cardiothoracic, 13.0% of orthopedic, 13.0% of general, 11.4% of vascular, 8.0% of central nervous system, 7.1% of plastic and otolaryngological, 6.6% of urological, and 4.7% of gynecological surgery patients[10]. Unfortunately, there is no consensus on this issue and different incidence rates have been given for POD after different types of surgery such as 15.3%-23.4% for cardiovascular surgery, 16.9% for hip fracture surgery, and 22.7%-26% for emergency surgery[11-13]. If patients are transferred to the intensive care unit (ICU) after surgery, the risk of delirium is higher than the incidence rate given for that surgery. Furthermore, emergency surgery increases the incidence of POD by 1.5 to 3 times compared with elective surgery[14].

Risk factors

Multiple risk factors for the development of POD have been identified (Table 2). Older cancer patients are expected to be at higher risk of delirium owing to additional risk factors such as drug toxicity and side effects, possible brain metastases, accelerated cognitive decline related to chemotherapy, and immune system impairment that result from the cancer disease process[15,16]. Delirium develops in up to 57% of elderly patients with oncological disease, depending on tumor stage, age, analyzed risk factors, diagnostic method, and treatment environment[17]. In a recent study examining the risk factors of delirium in older adults after elective abdominal cancer surgery and its effects on adverse surgical outcomes, the incidence of POD was reported to be 5.5%, which was lower than the 10%-15% reported in similar studies. However, the study reported different incidence rates observed in different cancer surgeries, including 2.2% for colorectal, 8% for stomach, 8.6% for liver, and 11.4% for pancreatic or non-hepatic biliary tract cancer[18].

Table 2 Factors associated with the risk of postoperative delirium[4].

Predisposing factors	Precipitating factors
Advanced age	Preoperative preparation: (1) Long-duration of fluid fasting; and (2) Preoperative pain
Lower educational level	Perioperative medication: (1) Anticholinergic drugs; (2) Benzodiazepines; (3) Opioid use; and (4) Use of pethidine
Functional status: (1) Cognitive impairment; (2) Hearing impairment; and (3) Frailty	Intraoperative factors/management: (1) Deep anesthesia; (2) Intraoperative blood loss/ blood transfusion; (3) Cerebral oxygen desaturation; (4) Hypotension; and (5) Hypothermia
Comorbidities: (1) Depression; (2) Psychiatric illness; (3) Cerebrovascular disease; (4) Parkinsonism; (5) Heart failure; (6) Hypertension; (7) Mitral valve disease; (8) Diabetes; (9) Obstructive sleep; (10) Pulmonary disease; (11) Kidney disease; and (12) Number of medications	Surgical management: (1) Abdominal/orthopedic surgery; (2) Higher surgical Apgar score; and (3) Long-duration surgery
Comorbidity scores: (1) Higher ASA grade; (2) NYHA functional class III/IV; (3) Higher EuroSCORE; and (4) Higher Charlson comorbidity index	Postoperative management: (1) Severe pain; (2) Long-duration mechanical ventilation; and (3) Prolonged stay in intensive care unit
Alcohol abuse
Nutritional status: (1) Malnutritional/low albumin; (2) Low preoperative hematocrit; and (3) Vitamin D deficiency

EuroSCORE: European system for cardiac operative risk evaluation; NYHA: New York Heart Association.

Pathophysiologic mechanisms of delirium

The precise pathophysiologic mechanisms underlying postoperative cognitive impairment and delirium are not fully understood. Some evidence suggests that various neuroinflammatory mechanisms and neurotransmitters may be involved. Perioperative increases in the concentration of C-reactive protein and interleukin 6 were found to increase the risk of POD[19]. Structural and functional damage caused by systemic inflammation at the blood-brain barrier may result in the accumulation of various inflammatory factors in the central nervous system, leading to loss of synaptic plasticity, neuro-apoptosis, and impaired neurogenesis[20]. Acetylcholine elevation caused by acetylcholine esterase deficiency is considered an independent risk factor for the development of POD, especially after cardiac surgery[21]. Other pathological processes that are thought to be involved in developing POD are mitochondrial dysfunction, oxidative stress, neurotrophic support impairment, and synaptic damage[22].

Adverse outcomes

Many adverse outcomes make POD an undesirable complication for elderly surgical patients. Delirium increases 1-mo, 6-mo, and 1-year postoperative mortality rates after elective surgery. POD was found to be associated with postoperative cardiac, respiratory, neurological, and renal complications, unplanned ICU admissions, prolonged hospital and ICU stays, and readmission to hospital. Sleep disturbances, feeding problems and aspiration pneumonia, malnutrition and fluid and electrolyte abnormalities, constipation, fever, infections, decubital ulcers, hip luxation, and urinary incontinence are more likely to develop in older patients undergoing emergency surgery[23]. Therefore, understanding how to identify delirium in recognizing acute illness is crucial, especially in elderly patients.

Diagnosis of POD

Several diagnostic tools can be used for early recognition of POD such as the Confusion Assessment Method, the Richmond Agitation-Sedation Scale, the Memorial Delirium Assessment Scale, the Delirium Rating Scale-Revised-98, the 4 'A's Test, Confusion Assessment Method and Single Question to identify Delirium, PINCHME (a mnemonic for the review of possible causes for delirium), the Observational Scale of Level of Arousal, the National Early Warning Score, and Recognizing Acute Delirium As Part of Your Routine. The acceptance and routine use of these tests vary from country to country and from region to region within the same country. Some of the tests require training of health personnel to use them easily and without problems.

In a recent issue of the World Journal of Gastrointestinal Oncology, Hu et al[24] published an interesting retrospective analysis entitled “Predictive modeling for POD in elderly patients with abdominal malignancies using synthetic minority oversampling technique” that addresses the importance of early and accurate recognition of POD. They retrospectively analyzed 611 elderly patients who had undergone surgical treatment of malignant abdominal tumors. The Ambiguity Assessment Scale was used, and POD was diagnosed in 140 patients (22.91%). After the identification of independent risk factors for POD including the Charlson comorbidity index, American Society of Anesthesiologists (ASA) classification, history of cerebrovascular diseases, surgical duration, perioperative blood transfusion, and postoperative pain score, the investigators developed a novel predictive model called P1 and an early warning model called P2 using a synthetic minority oversampling technique (SMOTE) algorithm. After comparing their effectiveness, the authors found no significant difference between these two predictive models. They concluded that the SMOTE-based predictive model had a predictive accuracy comparable to that of traditional methods. SMOTE is an oversampling technique for classification problems in which synthetic samples are generated for the minority class. The algorithm helps to overcome the overfitting problem posed by random oversampling[25]. As the authors pointed out, the SMOTE algorithm balanced the data and improved the predictive ability of the P2 model that they tested.

Early prediction of POD

As delirium is considered a preventable clinical entity in about half of the cases, various early prediction models developed with the support of machine learning have recently become a hot scientific topic. When examining the studies in general, differences and contradictions in the basic approach in model development such as type of surgical operation (cardiac vs non-cardiac), patient group, in-hospital location (ICU vs non-ICU), and risk factors are seen. In an analysis of 14 studies that included predictive models developed for the early diagnosis of POD in elderly patients treated outside of ICU, only two of 14 clinical prediction models were identified as best performing, with fair discrimination and acceptable calibration and were recommended to assist physicians in estimating POD risk and selecting patients for preventive intervention[26]. Zhou et al[27] evaluated 16 studies of risk prediction models for POD in patients with fragility hip fractures. They reported that all models had a high risk of bias, because of inadequate sample size, inappropriate handling of missing data, lack of model performance evaluation, and overfitting of the models. They concluded that all the models were still in the development stage. Therefore, new studies with more accurate planning and methodology are expected in this field.

Surgical factors in the development of POD

A study by Hu et al[24] provided information about the relationship between POD and the intraoperative process including anesthesia. Firstly, when evaluating the association of POD with surgical procedures, cardiac and non-cardiac surgery should be considered separately. Age was an important risk factor, and the risk of developing POD in patients ≥ 85 years of age and was 6.2 times higher than in those ≥ 65 years of age and over. It was stated found each 1 hour increase in the duration of surgery caused a cumulative 11% increase in the risk of POD[28]. Although no difference was found between regional and general anesthesia techniques regarding POD, the choice of anesthetic agent has been reported to be important, with intravenous anesthetics such as propofol and ketamine being more advantageous than inhaled agents. As the risk of POD increases with the depth of anesthesia, it is recommended to monitor the depth by electroencephalogram-guided techniques such as the bispectral index monitoring or auditory-evoked potential monitoring. The European Society of Anaesthesiology and Intensive Care Medicine published an evidence-based and consensus-based guideline on POD in adult patients. It gave recommendations regarding intraoperative risk factors, including those mentioned above (Table 3)[29].

Table 3 European Society of Anesthesiology and Intensive Care recommendations regarding the risk of postoperative delirium in elderly patients[29].

Recommendation	Quality of evidence	Strength of recommend
We recommend evaluating the following preoperative risk factors for POD: Older age, American Society of Anesthesiology Physical status score > 2, Charlson comorbidity index > 2, and Mini-Mental State Examination score lower than 25 points	Moderate	Strong
We do not suggest any specific type of surgery or type of anesthesia to reduce the incidence of POD	Low	Weak
When dexmedetomidine is used intra-operatively or postoperatively with the aim to prevent POD, we recommend balancing the expected benefits against the most important side effects (bradycardia and hypotension)	Moderate	Strong
We recommend that preoperative anesthesia consultation in older adults includes the screening for risk factors for POD and addresses patients’ needs to optimize their preoperative status	Low	Strong
We recommend that the results of the screening for POD risk factors are shared among the care team and the preventive strategies discussed and registered in the medical records	Low	Strong
We suggest Index-based EEG monitoring depth of anesthesia guidance to decrease the risk of POD	Low	Weak
We suggest multiparameter, intraoperative EEG monitoring (burst suppression, density spectral array) during anesthesia to decrease the risk of POD	Low	Weak

EEG: electroencephalogram; POD: Postoperative delirium.

CONCLUSION

The world population is aging, and while surgery is the primary treatment for many cancers, it can cause various postoperative complications, especially in elderly patients. POD is an undesirable complication that may have devastating effects, especially in elderly patients, and may be life threatening. As it is preventable in many patients, efforts have recently been made to develop predictive models that provide early diagnosis in the name of preventive medicine. Indeed, the lack of an effective treatment for delirium necessitates a high level of awareness of POD in the perioperative period when providing healthcare to older patients, until a highly sensitive and specific predictive method is available.