Impact of Pre-admission Central Nervous System Medication Use on Delirium in Hospitalized Patients with Advanced Cancer: A Multicenter, Prospective, Observational Study in Korea
Article information
Abstract
Purpose
This study investigated the impact of pre-admission central nervous system (CNS) medication use on delirium incidence, duration, and survival in hospitalized patients with advanced cancer.
Materials and Methods
In this multicenter prospective study across four tertiary hospitals in South Korea, adults with advanced cancer were enrolled and categorized based on their use of CNS medications within 90 days preceding admission. Associations between pre-admission CNS medication use and outcomes (delirium incidence, delirium duration, and overall survival) were assessed using multivariable regression and Cox proportional hazards models.
Results
Of the 190 patients enrolled, 140 had used CNS medications prior to admission. Delirium occurred in 22.1% of the patients with CNS medication use versus 14.0% of those without (adjusted odds ratio [aOR], 2.53; 95% confidence interval [CI], 0.95 to 7.60; not significant). Opioid (aOR, 2.48; 95% CI, 1.01 to 6.61) and antidepressant (aOR, 5.58; 95% CI, 1.22 to 27.35) use were significantly associated with increased delirium risk. Use of three or more CNS medication classes was associated with a markedly high risk (aOR, 11.15; 95% CI, 2.13 to 64.17). Delirium duration did not differ significantly between groups. Patients with pre-admission CNS medication exposure exhibited shorter overall survival (adjusted hazard ratio [aHR], 1.45; 95% CI, 1.01 to 2.09). Prior opioid use was also associated with increased mortality (aHR, 1.45; 95% CI, 1.03 to 2.05).
Conclusion
Pre-admission exposure to CNS medication, particularly opioids and antidepressants, was associated with an increased risk of delirium in patients with advanced cancer. A thorough medication history review upon admission is crucial to identifying high-risk patients and implementing early preventive interventions.
Introduction
Delirium is a frequent and serious neuropsychiatric syndrome in patients with advanced cancer, especially among those receiving palliative care [1]. It is characterized by an acute disturbance in attention, awareness, and cognition that typically fluctuates over time [2]. Delirium can manifest as hyperactive, hypoactive, or mixed subtypes. Notably, hypoactive delirium is often challenging to detect in routine clinical settings [3,4], necessitating careful observation to ensure timely diagnosis and appropriate management. As a result, identifying risk factors and predictive markers for delirium has been a major focus of clinical research.
The onset and development of delirium have been associated with various factors, including metabolic imbalances, infections, and medication effects [5]. Among them, medication use has been identified as a major modifiable contributor [6]. Given that delirium pathogenesis involves neurotransmitter dysregulation, particular attention has been directed toward drugs that act on the central nervous system (CNS). Common CNS-active medications—such as opioids, benzodiazepines, antipsychotics, and antidepressants—have been implicated in the development of delirium [7]. These medications may influence the incidence, severity, and duration of delirium through their influence on neurotransmitter pathways, such as those involving dopamine and acetylcholine. Previous studies have reported a correlation between CNS-active medication use and delirium incidence [8-10].
However, in patients with advanced cancer, opioid use is unavoidable for managing cancer-related pain, and anxiolytics or antidepressants are frequently used to address psychological distress. Medications administered during hospitalization can be adjusted under close supervision to mitigate delirium risk; however, any CNS-active medications used prior to admission represent an established exposure that cannot be modified upon hospitalization. Therefore, reviewing pre-admission medication histories is essential to identify patients who may be at high risk of developing delirium upon admission. Early identification of these high-risk patients is fundamental to delirium prevention and prompt management [11,12]. Understanding the influence of historical CNS medication use on delirium risk is crucial for developing tailored interventions and improving outcomes of patients with advanced cancer receiving palliative care. Despite its importance, research on the association between pre-admission medication use and delirium in patients with advanced cancer remains limited. A previous study about delirium in patients with advanced cancer examined medication effects during hospitalization, rather than pre-admission exposure [9].
Therefore, this study aims to investigate the impact of pre-admission CNS medication use on both the incidence and duration of delirium in hospitalized patients with advanced cancer receiving palliative care. By examining this association through a multicenter, prospective, observational study in Korea, we seek to elucidate the role of pre-existing CNS medication exposure in the development of delirium, ultimately guiding clinical strategies for risk assessment and early intervention.
Materials and Methods
1. Study design and population
We conducted a multicenter, prospective cohort study across four tertiary medical institutions in South Korea to evaluate the relationship between pre-admission use of CNS-active medications and the risk of developing delirium among patients with advanced cancer. The participating institutions were CHA University Bundang Medical Center, Seoul National University Bundang Hospital, Seoul National University Hospital, and Yonsei University Severance Hospital. Patient enrollment occurred from May 1, 2021, through May 31, 2023, across participating hospitals.
Patients eligible for the study were adults (≥ 20 years of age) admitted to the Acute Palliative Care Units (APCUs) at the participating hospitals during the study period. The inclusion criteria required that participants have an advanced-stage solid malignancy and be admitted specifically for acute palliative care management. To ensure a homogenous study cohort and accurate assessment of the primary outcome, we excluded patients with any of the following conditions: hospital stays exceeding three months (n=4), transfer from the APCU to other departments prior to completion of APCU care (n=7), terminal delirium, defined as delirium occurring within 2 weeks prior to death (n=20) and significant pre-existing cognitive impairment such as dementia (n=0). After applying these exclusion criteria, a total of 190 patients remained eligible and were included in the final analysis. The use of CNS-active medications within 90 days prior to admission was assessed for enrolled patients, but post-admission use was not considered in the eligibility criteria.
At study entry, participants underwent a comprehensive assessment that included the collection of demographic and disease characteristics, family and social histories, medical history, CNS medication usage and classification, and functional measures. All patients were followed from their index date, defined as the date of hospital admission, until death from any cause or the end of the study period (November 30, 2024), whichever occurred first.
2. Exposure
To evaluate the association between pre-admission CNS medication use and clinical outcomes, we defined the exposure time window as the 90-day period prior to hospital admission [13]. Exposure was considered present if the drug was prescribed and administered systemically at least once during this period. Post-admission changes in medication use were not incorporated into the exposure definition, as the study aimed to assess the impact of pre-existing CNS medication use prior to hospitalization. After admission, CNS-active medications were managed by specialized palliative care teams based on clinical judgment and patient needs. CNS medications were defined to include opioid analgesics (identified by Anatomical Therapeutic Chemical (ATC) code N02A), sedatives (N01AX, N03AE, N05BA, N05CC, N05CD, N05CF, N05CM), antipsychotics (N05A), antidepressants (N06A), antiepileptics (N03A, N03BF), cholinergics (N06DA, N07A), and anticholinergics [14] (A03AA, A03AB, A03B, A03CA, A03DA, A04AD01; scopolamine, N04A, N05BB01; hydroxyzine, R03BB, R06AA02; diphenhydramine, S01FA). A detailed list of CNS medications actually used by the study participants, categorized by drug class and ATC code, is provided in S1 Table.
3. Outcomes
The primary endpoint was the incidence of delirium, identified via medical records based on in-person assessments conducted by physicians using validated tools, including the Confusion Assessment Method (CAM), the Memorial Delirium Assessment Scale (MDAS), the 4 ‘A’s Test (4AT) and the Diagnostic and Statistical Manual of Mental Disorders, fifth edition (DSM-5). Secondary endpoints included the duration of delirium and survival. This study aimed to determine whether the current or recent use of CNS medications before hospitalization in patients with advanced cancer affects delirium incidence, duration, and survival. Delirium-related outcomes were analyzed by stratifying them into delirium subtypes, and the influence of each drug class, as well as the number of drug classes used on all outcomes, were also analyzed.
4. Covariates
Based on a comprehensive review of the literature, we identified a range of covariates that could potentially confound the relationship between CNS medication use and the onset of delirium. The study incorporated various patient-related factors, including demographic characteristics (such as age and sex), living arrangements (cohabitation with family members), marital status, primary caregiver, socioeconomic indicators (economic status and educational attainment), and the use of assistive medical devices (e.g., glasses or hearing aids). Nutritional status was evaluated using body mass index, categorized as underweight (< 18.5 kg/m2), normal (18.5-25 kg/m2), overweight (25-30 kg/m2), or obese (≥ 30 kg/m2). Physiological parameters included body temperature (classified as normal [< 38°C] or hyperthermic [≥ 38°C]) and blood pressure (with hypertension defined as systolic blood pressure ≥ 140 mmHg or diastolic blood pressure ≥ 90 mmHg). A prior history of delirium was also recorded. Lifestyle factors encompassed alcohol consumption (categorized as non-drinker, former drinker, less than once per month, 2-4 times per month, 2-3 times per week, and 4 or more times per week) and smoking status (categorized as never, former, or current smoker). Comorbid medical conditions included cardiovascular and cerebrovascular diseases, diabetes mellitus, respiratory illnesses, psychiatric disorders, and a history of head injury. Cancer-specific variables included cancer type and whether chemotherapy was administered during hospitalization. All variables were extracted from medical records and classified based on the International Classification of Diseases, 10th revision (ICD-10).
5. Statistical analysis
Differences between groups for categorical variables were analyzed using the chi-square test or Fisher’s exact test, depending on the distribution. Continuous variables were compared using Student’s t test or the Mann-Whitney U test. Logistic regression analysis was performed to estimate the odds ratio (OR) and 95% confidence interval (CI) for the association between pre-admission CNS medication exposure and delirium incidence. Analyses were stratified by CNS medication classification and the number of CNS medication classes combined. For classifications with sparse data in which delirium incidence was zero, Firth’s penalized logistic regression was applied to obtain more reliable estimates. To assess the association between CNS medication use and delirium duration, linear regression analysis was conducted, with results reported as standardized coefficients (beta) and corresponding 95% CIs. Two extreme outliers in delirium duration—defined as values more than three times the interquartile range below the first quartile or above the third quartile—were excluded from the regression analysis. Survival analysis was conducted using the Kaplan-Meier method, with group differences assessed using the log-rank test. Multivariate analysis was performed using the Cox proportional hazards regression model, and results were presented as hazard ratios with 95% CIs. Subgroup analyses were also conducted by age group (< 65 years vs. ≥ 65 years) for each clinical outcome. A p-value of less than 0.05 was considered statistically significant. All statistical analyses were performed using R ver. 4.4.2 (R Foundation for Statistical Computing).
Results
1. Patient characteristics
A total of 190 patients were enrolled in the study, with 140 patients classified into the CNS medication group and 50 patients into the non-CNS medication group based on their pre-admission exposure to CNS medications. The baseline characteristics of the study population are presented in Table 1. The mean age was lower in the CNS medication group compared to the non-CNS medication group (66.1 vs. 70.9 years, p < 0.01). Otherwise, the groups were generally comparable in terms of demographic, socioeconomic, and clinical characteristics.
Baseline characteristics of study subjects
Among patients in the CNS medication group, opioids were the most commonly used (91.4%), followed by sedatives (17.1%) and antipsychotics (10.0%) (S2 Table). Most patients (71.4%) were exposed to a single class of CNS medication; however, 29 (20.7%) and 11 patients (7.9%) were exposed to two and three or more classes, respectively. The detailed distribution of CNS medication co-administration is provided in S3 Table.
2. Outcomes of delirium and survival
Delirium was diagnosed in a total of 38 patients (20.0%) during hospitalization. The relative risk of delirium incidence associated with pre-admission CNS medication use was 1.6. Although the difference was not statistically significant, the delirium incidence rate was higher in the CNS medication group compared to the non-CNS medication group (22.1% vs. 14.0%). Hypoactive delirium was the most prevalent subtype in both groups. Notably, hypoactive and mixed-type delirium occurred at higher rates in the CNS medication group than in the non-CNS medication group (Table 2). The median duration of delirium was slightly longer in the CNS medication group, with a mean of 6.5 days compared to 5.0 days in the non-CNS medication group (Table 2). The overall mortality rate was 95.0% in the CNS medication group and 86.0% in the non-CNS medication group; however, no statistically significant difference was observed between the groups (Table 2).
Delirium and survival outcomes
3. Impact of pre-admission CNS medication on delirium and survival
In the multivariable analysis, pre-admission CNS medication use revealed a trend toward increased risk of delirium incidence, although it did not reach statistical significance (adjusted OR [aOR], 2.53; 95% CI, 0.95 to 7.60) (Fig. 1). However, when examining specific medication classes, prior opioid use (aOR, 2.48; 95% CI, 1.01 to 6.61) and antidepressants (aOR, 5.58; 95% CI, 1.22 to 27.35) were each significantly associated with an increased risk of delirium occurrence (Fig. 1). Additionally, antipsychotics (aOR, 3.74; 95% CI, 1.00 to 13.71) exhibited a trend toward increased delirium risk. Furthermore, the use of three or more CNS medication classes was significantly associated with a higher risk of delirium (aOR, 11.15; 95% CI, 2.13 to 64.17) (Fig. 1).
Forest plot of delirium incidence. CI, confidence interval; CNS, central nervous system; OR, odds ratio. a)The model was adjusted for age, sex, living with family, medical aid recipients, body mass index, alcohol consumption, smoking, chemotherapy during hospitalization, and previous history of delirium, b)Since the absence of events, Firth’s penalized logistic regression was applied. Numbers in bold indicate statistical significance.
Neither pre-admission CNS medication use nor multiple CNS medication combinations had a statistically significant impact on delirium duration (standardized beta, 0.07; 95% CI, –0.30 to 0.44) (S4 Fig.).
The survival difference between the two groups approached borderline statistical significance (p=0.05) (Fig. 2). In the multivariable analysis, pre-admission CNS medication use was associated with an increased risk of death (aHR, 1.45; 95% CI, 1.01 to 2.09) (S5 Fig.). Among individual drug classes, only opioid use was independently linked with significantly reduced survival (aHR, 1.45; 95% CI, 1.03 to 2.05) (S5 Fig.). Mortality was not dependent on the number of concomitant CNS medication classes (S5 Fig.).
Kaplan-Meier curves for overall survival. CNS, central nervous system.
In the age-stratified subgroup analysis, no statistically significant interaction was observed between age group (< 65 vs. ≥ 65 years) and pre-admission CNS medication use for any of the outcomes (S6 Table). However, among patients aged ≥ 65 years, the incidence of delirium was higher in those who used CNS medications prior to admission (30.1%) compared to those who did not (15.4%) (S6 Table). In this subgroup, pre-admission CNS medication use was significantly associated with delirium incidence, with an aOR of 3.17 (95% CI, 1.07 to 11.16) (S6 Table).
Discussion
This multicenter, prospective, observational study demonstrated that pre-admission exposure to CNS medications was associated with an increased risk of delirium incidence in hospitalized patients with advanced cancer. Specifically, the use of opioids and antidepressants, as well as polypharmacy involving three or more CNS medication classes, were identified as significant risk factors. Additionally, the use of antipsychotics demonstrated a potential association with increased delirium risk, although statistical significance was not attained. However, pre-admission CNS medication use did not have a statistically significant impact on the duration of delirium. Furthermore, overall survival was lower among patients with pre-admission CNS medication exposure, particularly those who had received opioids.
Delirium pathogenesis is known to involve disruptions in neurotransmitters such as dopamine, acetylcholine, and serotonin [15,16]. CNS-active medications typically act on these mechanisms, supporting prior research findings that highlight their role in increasing delirium risk [10,17]. Among these medications, opioid use is not only associated with increased delirium risk but also exhibits a dose-dependent relationship with delirium incidence [18]. Opioid dose titration for cancer-related pain was common among study participants, which may explain the increased risk of delirium. Additionally, the use of antidepressants was also associated with delirium occurrence, with the most commonly used antidepressants in this study being trazodone, mirtazapine, and nortriptyline. This finding aligns with prior pharmacovigilance studies indicating a significant association between these antidepressants and delirium risk [19]. Although our data suggested a possible association between antipsychotic use and delirium, this relationship did not achieve statistical significance. The primary antipsychotics used by patients in our cohort were quetiapine and olanzapine, which are frequently used to manage established delirium but are also known for their anticholinergic properties that could contribute to delirium onset [20]. Moreover, the concurrent use of three or more CNS medication classes was significantly associated with an elevated risk of delirium. These findings underscore the importance of thoroughly reviewing the pre-admission medication history of patients, with particular attention to the use of CNS medications or polypharmacy, to identify patients at high risk of delirium and facilitate early prevention and management strategies. Notably, in our subgroup analysis, the association between pre-admission CNS medication use and delirium incidence appeared more pronounced among patients aged ≥ 65 years, supporting prior evidence that older adults are particularly vulnerable to the neuropsychiatric effects of CNS-active drugs [21,22]. This age-related susceptibility highlights the need for age-specific delirium risk assessments and careful medication reconciliation upon hospital admission.
Previous studies have suggested that CNS medication use may influence not only the incidence of delirium but also its severity, duration, and overall survival [23,24]. However, in this study, CNS medication use did not have a statistically significant impact on delirium duration. One possible explanation for this finding is that the exposure time window in our study was limited to pre-admission medication use, rather than medication exposure at the time of delirium onset, which could attenuate the apparent impact on delirium duration. Additionally, given that the majority of patients in our study were admitted to palliative care units for symptom management of advanced cancer, the potential for inpatient medication optimization may have contributed to the prevention of prolonged delirium episodes. Further research incorporating more detailed medication exposure data including in-hospital medication changes and standardized delirium assessment tools is warranted to better understand the relationship between CNS medication use and delirium duration.
In this study, regardless of CNS medication class combinations, patients with pre-admission CNS medication use—particularly those exposed to opioids—exhibited a higher risk of mortality. Opioids may contribute to increased mortality risk through sedation, respiratory depression, and overall physiological decline. However, this association should be interpreted with caution, as opioid use may also be indicative of greater disease severity, poor pain management, or other underlying health conditions. Additionally, opioid therapy is often indispensable in cancer pain management, and its use is likely to continue during hospitalization. Therefore, baseline patient screening that accounts for these factors is critical for optimizing treatment strategies and implementing appropriate palliative interventions aimed at improving survival outcomes.
This study provides valuable insights into the assessment of delirium risk in patients with advanced cancer. The observed association between pre-admission CNS medication use and increased delirium incidence underscores the importance of conducting comprehensive medication history reviews upon hospital admission—an aspect that has received comparatively less attention in earlier studies, which have largely focused on medication use during hospitalization [9,25]. Among the delirium subtypes, hypoactive delirium is particularly prone to misdiagnosis, often mistaken for fatigue or depression in patients with cancer, which can result in delayed or missed diagnoses in the absence of systematic screening [26,27]. In this study, a prospective cohort analysis revealed that more than half of the delirium cases (21 out of 38; 55.3%) were classified as hypoactive delirium, with a higher prevalence in the CNS medication group. These findings underscore the need for early screening and precise diagnosis of hypoactive delirium, reinforcing the importance of structured patient assessment approaches. Identifying patients at high risk of delirium, particularly those with opioid exposure or polypharmacy, may facilitate early interventions and closer monitoring.
This study has several limitations that should be acknowledged. First, as an observational study, causal relationships cannot be definitively established, and residual confounding may remain. However, we mitigated this through adjustment for a comprehensive set of demographic and health-related characteristics. Second, this study did not examine the specific dosage or duration of pre-admission CNS medications, and post-admission use was not assessed, both of which may have influenced the observed effects. Third, the relatively small sample size for certain CNS medication subgroups may have limited the statistical strength to detect significant associations. Future prospective studies incorporating detailed medication exposure data, standardized delirium assessment tools, and larger patient cohorts are warranted to enhance our understanding of the impact of specific CNS medications on delirium outcomes.
Pre-admission CNS medication use is associated with an increased risk of delirium in hospitalized patients with advanced cancer, particularly among those exposed to opioids, antidepressants, or multiple CNS medication classes. Although we did not observe a significant effect of pre-admission CNS medications on the duration of delirium, such exposure was linked to a higher risk of mortality. These findings highlight the need for comprehensive medication history reviews upon hospital admission to identify high-risk patients, enabling targeted early interventions and improved clinical outcomes in palliative care settings.
Electronic Supplementary Material
Supplementary materials are available at Cancer Research and Treatment website (https://www.e-crt.org).
Notes
Ethical Statement
We conducted a multicenter, prospective cohort study across four tertiary medical institutions in South Korea. Institutional Review Boards at each site approved the study protocol prior to study initiation (approval numbers: CHAMC 2021-03-054-002; Seoul National University Hospital H-2103-028-1201; Seoul National University Bundang Hospital B-2104/681-405; Yonsei University Severance Hospital 4-2021-0323). Written informed consent was obtained from all study participants. All methods were performed in accordance with relevant guidelines and regulations.
Author Contributions
Conceived and designed the analysis: Geum MJ, Kang B.
Collected the data: Yoo SH, Lee SW, Hong M, Jung EH, Kim YJ, Kang B.
Contributed data or analysis tools: Geum MJ, Yoo SH, Lee SW, Hong M, Jung EH, Kim YJ, Kang B.
Performed the analysis: Geum MJ, Kang B.
Wrote the paper: Geum MJ, Kang B.
Final approval of manuscript: Geum MJ, Yoo SH, Lee SW, Hong M, Jung EH, Kim YJ, Kang B.
Conflicts of Interest
Conflict of interest relevant to this article was not reported.
Funding
This research was supported by a grant of Patient-Centered Clinical Research Coordinating Center (PACEN) funded by the Ministry of Health & Welfare, Republic of Korea (grant number: RS-2020-KH094340).