Perioperative Risk Factors for Postoperative Delirium in Non-dementia Older Patients after Non-cardiac Surgery and Anesthesia: A Prospective Study
Article information
Abstract
Background
To investigate the incidence and perioperative risk factors for postoperative delirium (POD) in non-dementia older patients who underwent anesthesia for non-cardiac surgery.
Methods
This prospective cohort study was conducted on 195 non-dementia older patients, aged 60 years or older, who were hospitalized after non-cardiac surgery and anesthesia. The Confusion Assessment Method for the Intensive Care Unit was used to evaluate the occurrence of POD. Incidence of POD was reported. We conducted univariate and multivariate logistic regression to identify the risk factors associated with POD.
Results
A total of 195 patients were enrolled; 172 completed the study. POD occurred in seven patients within three days after anesthesia, which is a 4.1% incidence of POD. Multivariate logistic analysis showed arrhythmia, coagulopathy, urinary comorbidity, minimum intraoperative heart rate, and minimum post-anesthesia care unit (PACU) pain score as independent risk factors for POD. A minimum PACU pain score >1 is the optimum cutoff pain score for developing POD, with a sensitivity of 85.7% and a specificity of 69.1%. The postoperative complication rate and in-hospital mortality were significantly higher for patients with POD compared to those without POD.
Conclusion
The incidence of POD in the study population is 4.1%. Arrhythmia, coagulopathy, urinary comorbidity, minimum intraoperative heart rate, and minimum PACU pain score were independent risk factors for POD. The minimum PACU pain score is the strongest independent risk factor of POD. POD is associated with increased postoperative complications and in-hospital mortality rates in non-dementia older patients.
INTRODUCTION
The older population is increasing on a global scale in the present day. Thailand was on track to be classified as a fully aged society by 2023.1) In 2035, the nation is set to transition into a super-aged society, with over 30% of the population being over 60 years old.2) The number and percentage of people aged 60 and above are rising in both Thailand and worldwide. It is estimated to reach 1.4 billion by 2030 and 2.1 billion by 2050.3) Delirium and dementia are two of the most common causes of cognitive impairment in older people. Delirium is an acute confusion characterized by a disruption in attention, awareness, and cognitive function. This disruption occurs relatively quickly, typically within a few hours to a few days. The disturbances do not occur in severely reduced arousal, such as coma, and cannot be better explained by another neurocognitive disorder.4) Any stress such as illness, hospitalization, or operation can precipitate delirium in older adults.5) Delirium that occurs within one week after an operation is known as postoperative delirium (POD). POD is a common complication in older patients after surgery.5) It most frequently occurs within the first 3 days following surgery with an incidence of 3%–50%.6-9) Variation of the incidence depends on the type of surgery, study population, and tools for diagnosis.
Dementia is a chronic and progressive condition that results in a loss of cognitive skills. It is characterized by an impaired ability to remember, think, or make decisions, which can interfere with daily activities.10) Many studies found a relationship between delirium and dementia. The older adults with previous dementia comorbidity have a higher risk of developing POD whereas POD in non-dementia older adults may precipitate subsequent dementia in the future.11-14) Accordingly, the non-dementia older people, most of the geriatric population, are at risk. Information about POD in non-dementia older adults is still limited. Therefore, investigating the incidence and risk factors of POD in non-dementia patients is necessary. The exact mechanism of POD remains unclear. However, much literature has identified several preoperative, intraoperative, and postoperative risk factors that contribute to POD.15,16) Various anesthesia-related factors have been linked to POD, including medication, anesthesia type, drugs used, anesthesia depth monitoring, pain management, hemodynamic control, fasting time, and transfusion.17) However, there is no consensus on the best anesthetic management to reduce the risk of POD. As the older population increases, the number of older patients requiring anesthesia and surgery also increases. Prevention of postoperative complications in older patients is still challenging for anesthesiologists. Therefore, understanding perioperative risk factors may be beneficial in finding appropriate anesthetic care to reduce the likelihood of POD in non-dementia older patients. The purpose of this study was to investigate the incidence and perioperative risk factors for POD in non-dementia older patients who underwent anesthesia for non-cardiac surgery.
MATERIALS AND METHODS
Study Design
The prospective cohort study was approved by the Institutional Review Board of Panyananthaphikkhu Chonprathan Medical Center, a tertiary hospital in Thailand, from September 2022 to June 2023 (Approval No. EC 014/65). The protocol followed the Declaration of Helsinki. Written informed consent was obtained from all study participants.
Participants
Patients aged 60 years and older who underwent non-cardiac surgery under anesthesia and are expected to be admitted for more than 3 days of postoperative care are included. Participants were excluded if they refused to participate or met any of the following criteria: previous delirium or dementia, communication impairments, neurological surgery, or critical condition (e.g., shock, respiratory failure, or on ventilatory support). Patients scheduled for surgery and anesthesiology services undergo pre-anesthetic evaluation. Participant data was collected. During surgery, the patients were given anesthesia by a supervising anesthesiologist in the operating room. The anesthesiologist selected the most suitable method according to the established treatment standards. Data during surgery was recorded in the anesthetic record form. In the postoperative phase, patients were assessed for postoperative pain using a numeric rating scale at the post-anesthesia care unit (PACU). The initial pain assessment takes place when patients arrive at the PACU, and this is referred to as the maximum PACU pain score. If the pain score is greater than 3, a nurse anesthetist will administer medication to alleviate the pain. Following that, a second pain assessment is performed and recorded as the minimum PACU pain score. The patients are discharged from the PACU when they meet the criteria of the Modified Aldrete Score with a numeric rating lower than 4. Screening for POD using the Thai Confusion Assessment Method of the ICU (Thai CAM-ICU) was performed by an anesthesiologist or a nurse anesthetist.18) Screening for POD was conducted three times by the same research assistant, at 0–24 hours, 24–48 hours, and 48–72 hours after surgery. A CAM ICU-positive diagnosis indicates POD, while a CAM ICU negative indicates no POD. The attending physician was notified if delirium was detected for appropriate treatment. The results of the CAM-ICU assessment, postoperative complications, and the number of hospitalization days were recorded on the case report form.
Statistical Analyses
A sample size calculation was done to specify the number of individuals that should be included in the study. Based on a previous study, the incidence of POD was found to be 11.6% among older patients undergoing non-cardiac surgery.12) The sample size was calculated as:
where n is the sample size;
The required sample size is 151 individuals with an additional 10% to account for potential dropouts during data collection. Therefore, this study used a minimum of 167 participants. Descriptive statistics were computed for all study variables using SPSS statistics version 18.0 (SPSS Inc., Chicago, IL, USA). Continuous variables are typically expressed as their mean values and standard deviations for normally distributed data, while median and interquartile range are used for non-normally distributed data. Categorical data is represented using frequency and percentage. Univariate logistic regression analyses were conducted to compare the differences between patients with and without POD. Statistical tests including Pearson chi-square test, likelihood chi-square test, independent t-tests, Wilcoxon rank-sum test, and Mann-Whitney two-sample test were used. A p-value of <0.05 was considered statistically significant. To conduct univariate analysis, we used the variance inflation factor (VIF) to check for multicollinearity in logistic regression analysis. A higher VIF indicates a stronger correlation between independent variables, suggesting multicollinearity. Any independent variable with a VIF >10 should be excluded. Univariate and multivariate logistic regression analyses were used to explore risk factors for POD. The risk factors with a univariate p-value <0.05 are included in a multivariable regression model to identify independent variables for POD. The study reported relative risks and adjusted relative risks along with their respective 95% confidence intervals. All p-values <0.05 were considered statistically significant. The researchers used receiver operating characteristic (ROC) curve analysis to identify the optimal cutoff PACU pain score for predicting the development of POD. The analysis results are presented as sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, Yoden’s index, and area under the curve (AUC).
RESULTS
During the study period, 359 patients underwent assessment for eligibility, and 164 were excluded. The reasons for their exclusion are presented in Fig. 1. The remaining 195 patients were included in the study. Fifteen patients needed ventilatory support at the end of surgery. Eight patients were discharged from the hospital before completion of the assessment period. Thus, the data of 23 patients were withdrawn. The remaining 172 patients were included at the end of the analysis. Seven of 172 patients were diagnosed with delirium within 3 days postoperatively by CAM-ICU screening, yielding a 4.1% incidence of POD. Most patients who developed delirium did so on postoperative day 1 (57.1%), followed by 28.6% on day 3, and 14.3% on day 2. Patient demographics and perioperative characteristics are presented in Table 1. In this study, the most common types of operation procedures were orthopedic (50.6%) followed by general surgery (43.6%), and others (5.8%). Thirteen variables showed statistical differences between POD and without POD group including the American Society of Anesthesiologists (ASA) physical status (p=0.032), arrhythmia (p=0.009), coagulopathy (p<0.001), thrombocytopenia (p<0.001), diabetes mellitus (p=0.047), gastrointestinal comorbidity (p=0.008), urinary comorbidity (p=0.022), and infection (p=0.006), a minimum intraoperative heart rate (p=0.028), a maximum intraoperative heart rate (p=0.040), a minimum PACU pain score (p=0.029), postoperative complications(p=0.005), and discharge status (p<0.001) (Table 1). Diabetes mellitus and maximum intraoperative heart rate had VIF >10. As a result, we excluded diabetes mellitus and maximum intraoperative heart rate from the univariate analysis. There are eight risk factors with a univariate p-value <0.05, including ASA physical status greater than IV, arrhythmia, coagulopathy, gastrointestinal comorbidity, urinary comorbidity, infection, minimum intraoperative heart rate, and minimum PACU pain score were associated with POD (p<0.05) (Table 2). Multivariate logistic analysis showed arrhythmia (adjusted relative risk [aRR]=29.71; 95% confidence interval [CI] 3.61–244.75; p=0.002), coagulopathy (aRR=197.26; 95% CI 5.15–7,562.99; p=0.005), urinary comorbidity (aRR=9.92; 95% CI 1.47–66.69; p=0.018), minimum intraoperative heart rate (aRR=1.06; 95% CI 1.01–1.11; p=0.011), and minimum PACU pain score (aRR=1.47; 95% CI 1.17–1.84; p=0.001) as independent risk factors for POD in older patients without dementia (Table 2).
Flow diagram of patient enrollment.
Patient demographics and perioperative characteristics
Univariate and multivariate models of binary logistics regression analysis for independent risk factors for delirium at a hospital (n=172)
Patients with POD had significantly higher postoperative complications, 85.7% versus 33.3% for patients without POD (p=0.005). The in-hospital mortality rate was 14.3% for patients with POD and 0.6% for patients without POD (p<0.001) (Table 1). The ROC curve with AUC for minimum PACU pain score classification of delirium patients was 0.74 (95% CI 0.59–0.89) (Fig. 2). The optimum cutoff PACU pain score for developing delirium was >1. The sensitivity, specificity, positive likelihood ratio (LR+), negative likelihood ratio (LR-), and Youden’s index for optimum cutoff PACU pain scores were 85.7%, 69.1%, 2.77, 0.21, and 0.55, respectively (Table 3).
Receiver operating characteristic curve of delirium patients’ minimum PACU pain score classification. PACU, post-anesthesia care unit; AUC, area under the curve; CI, confidence interval.
Receiver operating characteristic curve analysis for the optimal cutoff minimum PACU pain score for developing delirium
DISCUSSION
In CAM-ICU screening, the incidence of POD in non-dementia older patients detected in this study was 4.1% within 3 days postoperatively after non-cardiac surgery and anesthesia. Most episodes of POD were registered on the first postoperative day, which agrees with previous studies.12) Multivariate logistic analysis showed arrhythmia, coagulopathy, urinary comorbidity, minimum intraoperative heart rate, and minimum PACU pain score, as independent risk factors for POD.
In Thailand, the incidence of POD in older patients who underwent non-cardiac surgery was found to be 4.8%–45%.11,12,19) The variation of the incidences depends on the type of surgery, study population, and tool for diagnosis. An incidence of 11.6% was reported in a previous research study with a similar population.12) Thus, the incidence of 4.1% in this study is low. An explanation for this variance may include the following. First, pre-existing dementia has been reported to be a strong risk factor for POD in many studies.12,19) After excluding older patients with pre-existing dementia, the occurrence of POD may be lower. Second, we focused on measuring POD within the first 72 hours after surgery, mainly related to anesthesia. We conducted POD screening for 3 days after the surgery, which differed from a previous study that screened for 7 days.12) As a result, incidents of later onset POD may have been overlooked. Third, we chose CAM-ICU for delirium screening instead of the Diagnostic and Statistical Manual of Mental Disorders Fifth Edition (DSM-5) from the American Psychiatric Association, the gold standard diagnostic criteria for delirium that psychiatrists perform.4) When utilized by clinicians and trained researchers, CAM-ICU screening in older patients has moderate sensitivity (68%–72%), but excellent specificity (98.6%). For this reason, the variation in incidence might depend on the sensitivity of the diagnostic tool used by non-psychiatrists. However, there are many postoperative patients in our hospital. Clinicians and nurses routinely perform POD screening. Psychiatric consultation is initiated when the patient presents delirium-like symptoms, and DSM-5 criteria are utilized for the definitive diagnosis. The advantages of the CAM-ICU include being accessible by non-medical staff and requiring less time for evaluation.20) This screening tool detected POD in 41% of cases, which is better than the 22.2% detection rate by physicians with no screening tool.21) Therefore, it would be suitable to utilize the CAM-ICU to screen for delirium in postoperative patients at our hospital.
Multivariate logistic analysis showed arrhythmia, coagulopathy, urinary comorbidity, minimum intraoperative heart rate, and minimum PACU pain score as independent risk factors for POD. We focus on modifiable factors that could change the outcome. In this study, the minimum PACU pain score is the strongest independent risk factor of POD (p=0.001). We utilized ROC curve analysis to determine the best cutoff PACU pain score for predicting the occurrence of delirium. Ideally, the ROC curve should be positioned as close to the upper left corner as achievable. The AUC is commonly employed to assess the predictive capacity of a model. A value closer to 1 indicates a stronger predictive capability of the model. In the present study, the ROC curve with an AUC for minimum PACU pain score classification of delirium patients was 0.74 (95% CI 0.59–0.89) (Fig. 2). The optimum cutoff pain score for developing POD was ≥1. The sensitivity was 85.7%, while the specificity was 69.1% (Table 3). It is suggested that the minimum PACU pain score may be one of the adjuvant tools for predicting POD in non-dementia older patients. Postoperative pain should be vigilantly monitored.
Pain is the most common complication after surgery. Higher postoperative pain scores are associated with an increased risk of POD because pain can induce an acute stress response that accelerates POD.8,17,22) Experiencing more pain may accelerate the stress response and precipitate POD as a result. Appropriate pain management may reduce the acute stress response and concurrently reduce POD as well. Strong opioids are generally used in postoperative pain management. Unfortunately, the use of opioids has been associated with POD.23) These suggest that the use of multimodal opioid-sparing analgesia should be considered. Once pain occurs, it should be treated as early as possible. Other effective perioperative interventions such as regional anesthesia, paracetamol, and NSAIDs should be evaluated in further study.
A surprising result was made that older age, which is commonly associated with an increased risk of POD in older adults, did not show a significant relationship with a higher incidence of POD in the non-dementia group.12,24) The data showed that after excluding abnormal brain function, age may not be a crucial factor in determining the risk of POD in this population.
The present study showed no substantial disparities between regional and general anesthesia methods. The result is consistent with a systematic review and meta-analysis by Zhuang et al.,25) who suggested that regional anesthesia and general anesthesia did not significantly impact POD. Additionally, there is no evidence that any anesthetic drugs influenced POD in our study. However, another systematic review showed that patients undergoing surgery under general anesthesia were approximately three times more likely to experience POD.16) The reason could be that this is a prospective observational study. The research team did not interrupt anesthetic management. Anesthesiologists adjusted the doses of anesthetic drugs as appropriate for each older patient depending on the type of operation, functional status, and clinical findings. Many confounding factors were established outcomes. Consequently, it is too early to conclude that anesthetic drugs and anesthesia methods do not take part in POD. Further study should be performed to confirm this finding.
The study revealed that patients who developed POD had a significantly greater incidence of complications than those who did not (85.7% vs. 33.3%; p=0.005). Unfortunately, one patient passed away in each group during their hospital stay, which resulted in a higher in-hospital mortality rate among those with POD, compared to those without (14.3% vs. 0.6%; p<0.001). Others also reported that POD has been associated with increased postoperative complications and mortality.7,26) Patients with POD had longer hospital stays, but this increased duration was statistically insignificant compared to those without POD (9 vs. 13 days; p=0.344). This result is congruent with those of previous studies.7,8,26)
There are some limitations in this study. We excluded previously diagnosed dementia patients using medical documents and history taken from the patients and their caregivers. We did not reassess the cognitive function of our patients. Mild cognitive impairment without daily activity disturbance for older patients who can decide to give informed consent independently be included in this study.
In conclusion, the incidence of POD in the study population is 4.1%. Arrhythmia, coagulopathy, urinary comorbidity, minimum intraoperative heart rate, and minimum PACU pain score were independent risk factors for POD. The minimum PACU pain score is the strongest independent risk factor of POD. POD is associated with increased postoperative complications and in-hospital mortality rates in non-dementia older patients.
Notes
CONFLICT OF INTEREST
There is no conflict of interest to declare in this study.
FUNDING
This study was supported by a grant from Panyananthaphikkhu Chonprathan Medical Center, Srinakharinwirot University, Nonthaburi, Thailand (No. 545/2565).
AUTHOR CONTRIBUTIONS
Conceptualization, NW, PP; Data curation, NW, WK, PP; Funding acquisition, NW; Investigation, NW, WK; Methodology, NW, PP; Project administrations, NW; Supervision, NW; Writing-original draft, NW, WK; Writing-review & editing, NW, WK, PP.