The incidence of postoperative delirium following general surgery ranges from 5% to 10% in the general population and 10% to 15% in the elderly [1]. The incidence of postoperative delirium depends upon the type of surgery. In those undergoing elective orthopedic surgery, the incidence varies from 9% to 15% [2]. Other studies have reported an incidence of postoperative delirium from 7% to 75% in patients who underwent total joint arthroplasty [3,4].

Total knee arthroplasty is an effective and safe procedure in patients with osteoarthritis [5]. Together with total hip replacement, these are the most common orthopedic procedures [6]. About 61,439 Koreans (9,029 men and 52,410 women) underwent total knee arthroplasty in 2016, and of these, 48,244 (88.5%) were aged 65 years or above¹⁾ (7,301 men and 44,241 women). The current population aged 65 years and above in Korea is around 6.57 million (13.2% of the total population); this number is 2.25 times more than it was in 2010 (11% of the total population). The ratio between the elderly and the young (0–14 years old) populations has also risen from 0.68 in 2010 to 0.95 in 2016 [7].

The pathophysiology of delirium is not well-understood, and preventive and therapeutic measures are lacking [8,9]. Assessment of preexisting risk factors of postoperative delirium is an important part of the perioperative care in elderly patients. Postoperative delirium is associated with longer hospital stays, delayed rehabilitation, cognitive dysfunction, higher health care costs, and increased mortality [10,11]. The aim of this study was to determine the prevalence of postoperative delirium after total knee arthroplasty in elderly patients and identify which factors might contribute to its development.

This study was conducted with the approval of the Institutional Review Board of our institute (SGPAIK2016-08-015). The subjects were patients aged 65 years and above who underwent total knee arthroplasty under either general or regional anesthesia between March 2009 and June 2016. We excluded patients with multiple trauma, metastatic malignant tumors, and those undergoing reoperations. Finally, a total of 318 out of 430 eligible patients were included in the study. The collected data were limited to the patients’ medical records from their hospitalization for surgery. The incidence of postoperative delirium was used as the primary outcome variable. The incidence of postoperative delirium was limited to cases in which the patient’s symptoms met the criteria of the Diagnostic and Statistical Manual of Mental Disorders, fifth edition (DSM-V) in consultation with a psychiatric specialist. The demographics (sex, age, height, and weight), American Society of Anesthesiologists physical status, alcohol history (number of drinks per week), smoking history (cigarettes per day, duration of smoking), other comorbidities (pre-existing dementia, delirium, hypertension, pathological arrhythmia, myocardial infarction, cardiac valvular disease, cardiomyopathy, chronic obstructive pulmonary disease, asthma, diabetes, acute or chronic renal failure, liver disease, cerebrovascular disease, and neurological disease) were evaluated through a retrospective review of the medical records. We also evaluated postoperative mortality, length of hospitalization, admission to an intensive care unit, type and duration of anesthesia, type of surgery, use and type of sedation during regional anesthesia, blood transfusion, frequency of hypotension, and frequency of hypothermia. Dexmedetomidine or midazolam was used for sedation during regional anesthesia. Hypotension was defined as an absolute systolic blood pressure below 90 mmHg occurs 3 times. Body temperature was measured using an esophageal temperature stethoscope or tympanic thermometer, and hypothermia was defined as below 36.0°C. All patients in this institute underwent routine postoperative analgesia using the following protocols: after admission to the post-anesthesia care unit, pain intensity was evaluated by a numeric rating scale (NRS, 0 to 10). If the NRS was above 5, intravenous fentanyl 1 μg/kg was repeatedly administrated until the patient’s NRS dropped below 4. Oxycodone- and nefopam hydrochloride-based patient-controlled intravenous analgesia were also provided continuously for all patients.

This study showed that the prevalence of postoperative delirium after total knee arthroplasty in elderly patients aged 65 years and above was 6.0% (n = 19/total 318 patients). The incidence of postoperative delirium can vary from study to study, and these differences seem to be based on the diagnostic methods used. Chung et al. [12] used trained research members to screen for delirium symptoms and confirmed the diagnosis of postoperative delirium with a psychiatrist; the prevalence of delirium in that study was 3.1%. Radcliff et al. [13] used trained research members to confirm the diagnosis of postoperative delirium and the prevalence was higher (10.4%). In this study, the diagnosis was limited to cases in which the patient’s symptoms met the DSM-V criteria in consultation with a psychiatric specialist.

Previously known risk factors for postoperative delirium include age, male sex, cognitive impairment, type and duration of surgery, general anesthesia, anesthetic time, postoperative pain, hypothermia, and blood transfusion [10]. Preoperative and postoperative hemoglobin, serum sodium levels, and postoperative PaO₂ are reported risk factors of postoperative delirium [14,15]. Postoperative pain is also known to be an important risk factor for postoperative delirium [16]. In this study, we found that preexisting dementia, the frequency of hypotension, and preoperative anemia were significantly associated with postoperative delirium. Although the pathologic relationship between dementia and delirium is not yet understood, many studies have investigated this significant correlation [17,18].

In this study, we also confirmed that the frequency of hypotension was associated with a higher incidence of postoperative delirium; however, the role of intraoperative hypotension in postoperative delirium remains controversial. Hirsch et al. [19] did not find a significant relationship between intraoperative hypotension or duration of hypotension and postoperative delirium in patients aged 65 years and above undergoing major non-cardiac surgery. However, meta-analysis of risk factors for postoperative delirium in gastrointestinal surgery identified that perioperative blood transfusion (OR = 3.2) and intraoperative hypotension (OR = 2.7) were both highly associated with postoperative delirium [20].

It is reported that anemia in elderly patients increases mortality and worsens their functional status [14]. Maldonado [21] have suggested that inadequate cerebral oxygenation may be the cause of delirium in severely ill patients who have an imbalance between oxygen supply and demand. Another recent study reported focal decrease of cerebral oxygenation could also be associated with cognitive dysfunction in elderly patients [22].

Decreased cerebral blood flow is influenced by hypotension and low hemoglobin levels, both of which impair oxygen delivery. A study that used transcranial Doppler found that blood flow velocity in the middle cerebral artery is lesser in the delirium group than in the non-delirium group in patients with dementia [23]. This could, in part, account for the correlation between postoperative delirium and hypotension or anemia.

Several limitations of our study are due to its retrospective design. First, a relatively small number of patients was included in this study. Although 10 years’ worth of medical records were used, this study was only conducted in one healthcare institution. Second, hypoactive delirium, wherein patients demonstrate reduced motor activity and sedation, is frequently less recognized than hyperactive, wherein delirium is often associated with psychotic features. Some studies have found that only 1.6% of intensive care unit patients experience pure hyperactive delirium, and hypoactive delirium is more common in patients aged 65 years and above [24]. This indicates that hypoactive delirium could be neglected and underestimated without careful monitoring using standardized scales to detect hypoactive or mixed-type delirium. Third, the results of this study are derived from only one type of surgery, making it difficult to generalize the findings to patients undergoing other surgeries. Forth, we could not collect specific data on the duration and severity of preexisting comorbidities, as this information was missing from the electronic medical records. Last is that postoperative delirium was only assessed during the hospitalization period; therefore, we might have missed delirium that occurred after patients were discharged.

In summary, we found that preoperative dementia is the most important risk factor of postoperative delirium. High-risk patients undergoing total knee arthroplasty should be fully evaluated, and when possible, dementia should be managed preoperatively. Adequate management of preoperative anemia and intraoperative hypotension might also be helpful in reducing the incidence of postoperative delirium in this population.