Background: Predisposing and precipitating factors for delirium for the elderly, over the age of 65 years, are known, but not for the very old, over 80 years. As the society is getting older and evermore patients will reach >80 years, more evidence of the factors and their contribution to delirium is required in this patient group. Methods: In the course of 1 year, 3,076 patients above 80 years were screened prospectively for delirium based on a Delirium Observation Screening (DOS) scale, Intensive Care Delirium Screening Checklist (ICDSC), and a DSM (Diagnostic and Statistical Manual)-5 nursing instrument (ePA-AC) construct. Relevant predisposing and precipitating factors for delirium were assessed with a multiple regression analysis. Results: Of 3,076 patients above 80 years, 1,285 (41.8%) developed a delirium, which led to twice prolonged hospitalization (p < 0.001), requirement for subsequent assisted living (OR 2.2, CI: 1.73–2.8, p < 0.001), and increased mortality (OR 24.88, CI: 13.75–45.03, p < 0.001). Relevant predisposing factors were dementia (OR 15.6, CI: 10.17–23.91, p < 0.001), pressure sores (OR 4.61, CI: 2.74–7.76, p < 0.001), and epilepsy (OR 3.65, CI: 2.12–6.28, p < 0.0001). Relevant precipitating factors were acute renal failure (4.96, CI: 2.38–10.3, p < 0.001), intracranial hemorrhage (OR 8.7, CI: 4.27–17.7, p < 0.001), and pleural effusions (OR 3.25, CI: 1.77–17.8, p < 0.001). Conclusion: Compared to the general delirium rate of approximately 20%, the prevalence of delirium doubled above the age of 80 years (41.8%) due to predisposing factors uncommon in younger patients.
Delirium is an acute neuropsychiatric syndrome associated with fluctuating orientation, attention, awareness, perception, higher order thinking, cognition, sleep-wake cycle, and emotionality, that is, anxiety or depression . In the general hospital setting, delirium is particularly common in older patients. Existing studies on delirium in the elderly investigate patient groups ≥65 years old, but not specifically the very old, very elderly, or oldest old patients >80 years of age. Despite the increasing life expectancy of the population and the fact that health care institutions will be increasingly confronted with patients over 80 years of age in the future, little is known about delirium in this growing patient group.
In general, the underlying etiology of delirium is often multifactorial and can be described by differentiating predisposing and precipitating factors. Predisposing factors exist prior to the onset of the delirium, such as dementia, age, and gender. Conversely, precipitating factors contribute to the activation of a delirium, for example, through metabolic or electrolyte abnormalities or inflammatory processes. In general, the more predisposing factors (e.g., age, dementia, and epilepsy) a patient accumulates, the fewer precipitating factors (e.g., mild inflammatory process such as a cold) are necessary for the development of a delirium [2, 3]. Since aging itself is marked by predisposing factors such as lessened coronary and renal function and diminished cognitive reserve, it is natural that younger adults are less vulnerable to delirium. The interrelationship of delirium and dementia is an important field, as the diagnosis and treatment of delirium superimposed on dementia is challenging. Many studies suggest that delirium could unmask pre-existing, but undiagnosed dementia [4‒6].
The early detection, prevention, and treatment of delirium is of crucial importance regarding the short- and long-term consequences: delirium causes increased morbidity and mortality, prolonged hospitalization, increased health care costs, and increased subsequent institutionalization. In addition, delirium causes a deterioration in general functional ability, cognitive decline, and the associated risk of loss of independence. The direct and indirect economic costs of delirium are high and could challenge the resources of future health care systems in an aging society, as age itself is one of the most important predisposing factors for the development of delirium. Prevalence rates vary among medical service : in obstetrics 2% , in neurology 35% , and in intensive care 70% and higher . A general prevalence of delirium can therefore be seen as an average between all age groups and medical services and is estimated to be 20% . In anticipation of the increasing number of patients ≥80 years, we aimed to (1) estimate the prevalence of delirium in the very old, (2) investigate previously reported risk factors in this age group, and (3) identify relevant predisposing and precipitating risk factors of delirium for care planning in advance.
Study Design, Patients, and Procedures
Over the course of 1 year (January 1–December 31, 2014), all patients (n = 39,442) were prospectively screened for delirium due to a delirium detection initiative (Delir-Path, Fig. 1) at the University Hospital Zurich, a tertiary care center. The reasons why or in which medical field they were admitted are analyzed in a previous work of our group. Patients were excluded if age was below 18 years, the length of stay was below 1 day, and missing data, including the electronic patient’s assessment, leaving 28,806 eligible patients. Of these eligible patients, 3,076 (12.8%) were above 80 years.
Characterization of Predisposing and Precipitating Factors for Delirium
Several predisposing and precipitating factors for the development of delirium have been described previously [3, 4, 6, 10]. In our cohort of very old patients, predisposing and precipitating factors for the development of a delirium were based on the formation of diagnostic clusters, according to the 10th revision of the International Statistical Classification of Diseases (ICD-10)  (Table 1). We calculated multiple logistic regressions to evaluate the impact of individual diagnostic clusters for delirium . Clusters relevant to delirium were included in the model, even if they did not reach significance and excluded only in instances of weakening the model, which was verified with the respective Cox-Snell’s and Nagelkerke’s r2.
A set of delirium screening scales was used to determine delirium: (1) the Delirium Observation Screening Scale , (2) the Intensive Care Delirium Screening Checklist , and (3) a DSM-5-criteria-based construct  from a nursing tool, called the Ergebnisorientiertes PflegeAssessment Acute-Care (ePA-AC) [16, 17]. On regular floors, the DOS scale was administered routinely 3 times daily during the first 3 days of admission in 65 years or older patients and based on the clinical evidence of incident delirium for all patients, that is, also in those <65 years. On the intensive care units (ICUs), the ICDSC was routinely performed thrice daily. Furthermore, nursing staff daily assessed various medical and functional parameters with the ePA-AC in all patients, regardless of whether a delirium was present or not. Since ePA-AC can also be used to measure delirium , information regarding the presence of delirium was therefore available for all patients and not only for older patients.
Validation of this delirium screening approach was done; delirium diagnoses were in 91% of patients correct, as determined by a reference-rater from a consultation-liaison psychiatry service. Further, this construct was tested against the validated DOS and ICDSC approach and achieved almost perfect agreement (Cohen’s κ 0.83, p < 0.001). On regular floors, patients ≥80 years were screened daily with the DOS and ePA-AC. On ICUs, the ICDSC was conducted 3 times per day. DOS, ICDSC, and ePA-AC were conducted by nursing staff specifically trained in a 4-h course with achievement tests; literature research and eLearnings were made. Additionally, part of the education was performed via case reports, state-of-the-art lessons on epidemiology, and knowledge about delirium, including diagnostic criteria.
The DOS is a 13-item scale validated to indicate delirium according to DSM-IV criteria (DOS, cutoff ≥3) . Items include disturbances of consciousness (1), attention (2–4), thought processes (5 and 6), orientation (7 and 8), memory (9), psychomotor behavior (10, 11, and 13), and affect (12). Symptoms are rated on a scale (0–1) as not existent (0), sometimes to always existent (1), and unable to assess (–). The cutoff score for delirium is ≥3, and values were aggregated throughout recordings. This approach proved to be valid and correctly identified 91% of delirium diagnoses as determined by the consultation-liaison psychiatry service.
The ICDSC is a screening instrument with 8 items (ICDSC, cutoff ≥ 4) based on the DSM-IV criteria specifically designed for the intensive care setting with 2 points: absent or present. This scale was designed for patients with limited communication abilities such as intubated patients. The items include the assessment of (1) consciousness (comatose, soporose, awake, or hypervigilant), (2) orientation, (3) hallucinations or delusions, (4) psychomotor activity, (5) inappropriate speech or mood, (6) attentiveness, (7) sleep-wake cycle disturbances, and (8) fluctuation of symptomatology. The maximum score is 8; scores of >3 indicate the presence of delirium. Each item is rated on the patient’s behavior over the previous 8 h .
The ePA-AC is a nursing instrument administered daily assessing mobility, personal care and dressing, feeding, elimination, cognition and alertness, communication and interaction, sleeping, breathing, pain, pressure ulcers, and wounds . DOS, ICDSC, and ePA-AC values as well as medical data were obtained from the electronic medical chart (Klinikinformationssystem, KISIM; CisTec AG, Zurich). This study was approved by the Ethics Committee of the Canton of Zurich (KEK-ZH-Nr. 2012-0263). A waiver of informed consent was obtained from the committee. Our reporting is in line with the STROBE statement .
Data analyses were performed with the Statistical Package for the Social Sciences version 25 and R statistical software version 3.5.0 for Windows. Descriptive characteristics are summarized depending on parametric properties using means and standard deviations or medians and interquartile ranges for continuous variables and percentages for categorical variables.
The data were tested with Shapiro-Wilk’s test for distribution of normal distribution. Intergroup differences for continuous variables were computed using Student’s t test and Mann-Whitney U test depending on their parametric properties and for categorical variables with Pearson’s-χ2 test.
In a first step, the delirium construct based on DSM-5 was tested, and its agreement with the validated approach – with a DOS cutoff ≥3 or ICDSC ≥4 – with Cohen’s κ was determined as measure of concordance. The agreement was defined as >0.80 near perfect .
In a next step, data were dichotomized by presence or absence of delirium. Then, simple logistic regressions were calculated to determine the prevalence rates of delirium for medical characteristics and their respective odds ratios (ORs) and corresponding confidence intervals (CIs). For all inferential tests, 2-tailed tests were chosen, and the significance level alpha (α) was set at p < 0.05.
Characteristics of Delirious Patients
Of 3,076 patients above 80 years, 1,285 (41.8%) developed delirium. The sociodemographic and medical characteristics of the delirious patients were typical for this syndrome and are displayed in Table 2. Gender was equally distributed between delirious and nondelirious patients with no significant impact on the development of delirium. Prior to admission, delirious patients resided less often at home (OR 0.36), required more assisted living (OR 2.73), or were transferred from other hospitals (OR 2.05) and admitted as emergencies (OR 2.24). ICU treatment increased the probability of delirium by 5-fold and the need for ventilation by almost 18-fold; there was a significant difference in group size (e.g., 208 ventilated delirious patients vs. 24 ventilated nondelirious patients). In addition, there was a significant difference in the length of stay in ICUs and the duration of ventilation.
Further, these patients had a prolonged hospitalization, that is, stayed almost twice as long, and were managed at an increased rate in the ICUs. At discharge, their physical and functional impairment was greater as indicated by increased requirement for outside hospital assistance (OR 2.37) and rehabilitation (OR 2.2) than returning to their own residence (OR 0.17); they also suffered from more severe illness as indicated by their increased mortality risk (OR 24.88).
Predisposing Factors for Delirium
Overall, predisposing factors for the development of delirium were related primarily to diseases of the nervous system. The most relevant predisposing factor was dementia, which led to a 15.6-fold increased risk of developing delirium (see Table 3; Fig. 2). Diseases such as pressure sores, epilepsy, and valvular heart disease caused an increased risk; formally, brain tumors and subcortical white matter lesions did not cause an increased risk; however, CIs widely ranged.
Precipitating Factors for Delirium
Summary of Main Findings
To date, this study is the largest to describe the predisposing and precipitating factors of delirium in patients over 80 years of age. In the course of 1 year, 41.8% of hospitalized patients ≥80 years old developed delirium; compared to the prevalence of delirium of 20% in the general population [8, 20], the delirium in patients ≥80 years old is therefore twice as high.
Not surprisingly, dementia was the strongest predisposing factor for the development of delirium. Other relevant predisposing factors were pressure ulcers and epilepsy. Important precipitating factors were systemic disorders which could be related to intracranial hemorrhage, acute renal failure, pleural effusion, and SIRS. Considering the different risk factors and the correspondingly affected organ systems, it is evident that especially diseases of the brain (e.g., dementia, subcortical lesions of the white matter, and epilepsy), the heart (e.g., heart valve diseases and pleural effusion), and body homeostasis (e.g., acute renal failure and sepsis/SIRS) are associated with delirium.
In addition, the development of delirium led to increased mortality, twice as long hospitalization, and the need for subsequent assisted living. Both the necessity and duration of treatment in intensive care or assisted ventilation were significantly associated with delirium. Patients with delirium in the course of their illness were more frequently referred from institutions such as nursing homes and assisted living and as emergencies, indicating greater functional and physical impairment, as well as more severe illness. In the aftermath of delirium, functional deterioration occurred as manifested by an increased need for assisted living and rehabilitation versus returning to their residence. Consequently, the cost of care for a patient following delirium increases, emphasizing the socioeconomic costs that follow.
Comparison with the Existing Literature
To our knowledge, there are no studies to date describing the predisposing and precipitating factors for the development of delirium exclusively in patients ≥80 years of age. Therefore, existing studies on risk factors for delirium ≥65 years of age were included as reference in this study. Besides the high prevalence rate of delirium aforementioned, it is striking that the risks – as described with the corresponding ORs – of several known predisposing and precipitating factors were higher than in previous studies on delirium. For example, the most relevant predisposing factor was dementia, which represented one of the most relevant factors (OR 15.6). In previous studies on delirium superimposed on dementia, lower risks were reported (OR 2.8–4.5) [21‒24] for several reasons: (1) the prevalence of dementia in patients ≥80 years of age ranges between 18 and 29% , in contrast to patients <80 years of age between 2 and 7% , which might affect the corresponding odds for the development of delirium. Since previous studies have examined a higher proportion of patients <80 years of age, the prevalence of dementia, and thus corresponding odds, might also be lower. (2) Dementia is not a static disease/condition  but worsens over time or with increasing age, and it seems natural that patients ≥80 years of age have a more advanced stage of dementia than patients <80 years old. It is conceivable that an advanced stage of dementia is more likely to predispose to the development of delirium, as cognitive abilities or resources are much lower . This circumstance could additionally contribute to the high OR for dementia in our patient population. To our knowledge, there is a lack of specific studies on the extent to which advanced versus mild dementia leads to delirium or not. Unfortunately, we had no information about the severity of the respective dementia. (3) Multimorbidity increases with age  and is itself a risk factor for the development of delirium [29, 30], and dementia patients often exhibit high co- or multimorbidity ; it is conceivable that a demented patient ≥80 years of age has a higher multimorbidity than a patient <80 years of age, which could also increase the risk of developing delirium. In addition, multimorbidity not only increases with age, but also with the diagnosis of dementia . One could speculate that a multimorbid patient with dementia has a higher risk of developing delirium than a patient without dementia. (4) In this context, the probability of polypharmacy also increases the risk of developing delirium . Taking these reasons (1–4) into account, the comparatively high risk of dementia in patients ≥80 years seems plausible for the development of delirium. In this sense, it also seems plausible why the risks of several predisposing and precipitating factors are higher in patients ≥80 years than in patients <80 years. In addition, a not to be underestimated rate of possibly undiagnosed dementia patients in our very old patient group would also be possible, which would additionally increase the risks of all factors.
In addition, there are several factors besides dementia that represent both multimorbidity and frailty and thus lead to an increased susceptibility for delirium. Consequently, these factors can be interpreted as indicators of the presence of multimorbidity and frailty. As an example, pressure sores are known to be associated with high frailty and immobility . According to this, it seems conclusive that in our study the presence of pressure sores is strongly associated with delirium; this circumstance also fits the existing literature . Furthermore, several precipitating factors can also be interpreted as an indicator of multimorbidity and frailty. Pleural effusions are more likely to occur in patients who are already multimorbid and have severe cardiac diseases. Head trauma often occurs as a result of a tendency to fall, which many frail patients tend to do. In summary, it can therefore be concluded that in addition to the high age of over 80 years and the presence of dementia, multimorbidity and frailty are highly associated with delirium.
In contrast to previous studies that focused on “older” people with different age definitions, this study comprehensively describes predisposing and precipitating risk factors for delirium in a large sample of patients between ≥80 and 102 years of age. In an aging society, the studies must focus on the growing number of these patients among hospitalized patients. These findings suggest that delirium is very common in patients ≥80 years of age, particularly in patients who are admitted from assisted living or as emergencies, stay longer in the hospital, and require assisted living on discharge or death. These results could serve as a basis for advanced care planning and suggest that it might be useful to screen for delirium in every patient over 80 years of age and – based on the literature to date – also in younger patients.
Strengths and Limitations
A strength of this study is the large sample size, including successive recordings over 1 year. Generalizability to other health care facilities is limited, since our patient population was representative for a tertiary care center. With regard to statistical correlations, it appears problematic if diseases (1) are rare or (2) have varying degrees of severity, which leads to a spread of the CI. The larger or wider the CI, the lower the statistical significance of the risk of contributing to delirium. Consequently, diseases may not be statistically significantly associated with the development of delirium, although it would be obvious on the basis of pathophysiological considerations and clinical experience. This is clearly shown by the example of poisoning and peritonitis, where the CIs ranged from 0.68 to 66.8.
Delirium in patients >80 years of age is common, likely underdiagnosed, and an indicator of later loss of functional independency and mortality. In practical terms, this means that a large proportion of patients do not return home after a delirium. This study is so far the largest of its kind in terms of predisposing and precipitating risk factors of delirium in very old patients and extends the limited knowledge about delirium in an aging society.
We want to thank all clinical staff, who made this study possible!
Statement of Ethics
This study was approved by the Ethics Committee of the Canton of Zurich (KEK-ZH-Nr. 2012-0263). A waiver of informed consent was obtained from the committee. The study was conducted ethically in accordance with the World Medical Association Declaration of Helsinki.
Conflict of Interest Statement
J.M., L.B., S.F., J.E., R.v.K., and S.B. report no disclosures or competing interests.
This work was supported by the Clinician Scientist Program of the Medical Faculty of the University of Tübingen (Program No. 45800).
J.M. analyzed clinical and diagnostic data and drafted and revised the manuscript. L.B. acquired clinical and diagnostic data and revised the manuscript. S.F. acquired clinical and diagnostic data and revised the manuscript. J.E. and R.v.K. analyzed clinical and diagnostic data and revised the manuscript. S.B. conceptualized the study, acquired and analyzed clinical and diagnostic data, and revised the manuscript. All authors read and approved the revised manuscript.
Availability of Data and Material
The anonymized data and materials are stored locally, and any raw data from the statistical analysis can be made available on reasonable request.