- Email: [email protected]
Delirium Risk Factors and Associated Outcomes in a Neurosurgical Cohort: A Case-Control Study
Original Article
Delirium Risk Factors and Associated Outcomes in a Neurosurgical Cohort: A Case-Control Study Ramin A. Morshed1, Jacob S. Young1, Michael Safaee1, Sujatha Sankaran2, Mitchel S. Berger1, Michael W. McDermott1, Shawn L. Hervey-Jumper1
OBJECTIVE: There are limited reports examining delirium in cohorts of neurosurgical patients across inpatient settings without separation based on subspecialty distinction. It is of interest to identify consistent delirium risk factors across various cranial pathologies and inpatient settings that will inform future interventional studies.
-
METHODS: Delirium rates, patient and hospitalization risk factors, and clinical outcomes in 235 patients undergoing a cranial procedure were examined in a retrospective fashion.
neurologic deficit being consistent risk factors across inpatient settings. These results help identify at-risk patients for delirium on a neurosurgical service to enact interventions preemptively.
-
RESULTS: Fifty-two (22.1%) patients experienced delirium during their hospital stay. Patient factors predictive of delirium on univariate logistic regression were older age, a diagnosis of hydrocephalus or intracranial infection, transfer from an outside hospital, and admission through the emergency department. Hospitalization factors predictive of delirium included longer length of intensive care unit (ICU) stay, abnormal sodium values preceding delirium, a new postoperative infection, and the presence of a neurologic deficit. Using recursive partitioning, age ‡72.56 years and ICU length of stay ‡5 days were identified as critical thresholds for predicting delirium (odds ratio [OR] 4.61 and 18.2, respectively). On multivariate logistic regression analysis, age (unit OR 1.05), length of ICU stay (unit OR 1.2), and a neurologic deficit (OR 5.4) were predictive of delirium. Furthermore, delirium was also significantly associated with a longer length of admission as well as decreased likelihood for discharge home.
-
CONCLUSIONS: Delirium is a frequent occurrence after neurosurgery with older age, longer ICU stay, and a
-
Key words Cranial - Craniotomy - Delirium - Neurosurgery -
Abbreviations and Acronyms CAM-ICU: Confusion Assessment Method for the Intensive Care Unit ICU: Intensive care unit NuDESC: Nursing Delirium Screening Scale OR: Odds ratio
WORLD NEUROSURGERY -: e1-e7, - 2019
INTRODUCTION
D
elirium is defined as an acute confusional state with a fluctuating course that impairs patient cognition and ability to comply safely with medical care. It is increasingly being recognized as a major complication following surgery that has significant long-term morbidity for patients and is one of the most common complications after surgery in older adults.1 Preventing and treating delirium is challenging with little evidence of efficacy supporting commonly used pharmacologic interventions such as atypical antipsychotics. Rather, reorientation techniques, minimization of sedating medications, and non-pharmacological measures such as encouraging eyeglasses/hearing aids and mobilization are generally employed to reduce the rates of hospitalization-associated delirium. Risk factors for delirium are usually classified as either predisposing or precipitating factors.2,3 Older age, cognitive impairment, functional disabilities (e.g., poor vision, hearing loss), preexisting comorbidities, alcohol use, and depression have all been associated with an increased risk of delirium.4-6 Precipitating factors that have been previously identified include surgery, anesthesia, infection, sedative and anticholinergic medications, high pain levels, and anemia.7-9 Specifically, 1 large trial of over 1000 patients aged >50 years who underwent elective non-cardiac surgery identified postoperative delirium in 9% of patients, with delirium being associated with greater intraoperative blood
From the Departments of 1Neurological Surgery and 2Medicine, University of California, San Francisco, California, USA To whom correspondence should be addressed: Ramin A. Morshed, M.D. [E-mail: [email protected]] Citation: World Neurosurg. (2019). https://doi.org/10.1016/j.wneu.2019.03.012 Journal homepage: www.journals.elsevier.com/world-neurosurgery Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2019 Elsevier Inc. All rights reserved.
www.journals.elsevier.com/world-neurosurgery
e1
ORIGINAL ARTICLE RAMIN A. MORSHED ET AL.
DELIRIUM RISK FACTORS AND ASSOCIATED OUTCOMES
volume loss, more blood transfusions, and postoperative anemia but not intraoperative hemodynamic complications or the route of anesthesia.8 In the setting of lumbar spine surgery, preoperative comorbidities have been shown to correlate with risk of delirium, and postoperative delirium was associated with increased length of stay, increased costs, and higher mortality.10 Delirium has also been found to be a risk factor for discharge to a postacute nursing facility.11 Neurosurgical patients may be at increased risk of postoperative delirium given that these patients often have prolonged intensive care unit (ICU) care, require frequent neurologic checks with interruption of sleep, possess neurologic or metabolic derangements associated with neurosurgical disease, and have neurologic deficits that may impair communication or mental status. Given the severity and impact hospital-associated delirium has on patients and family caregivers as well as its long-term consequences on outcomes, a better understanding of the prevalence of and the risk factors for delirium in patients who undergo a cranial neurosurgical procedure is needed. Although there are prior reports that have examined delirium rates and risk factors in specific subgroups of neurosurgical patients, there is a lack of delirium data pertaining to general neurosurgical cohorts. Here, we examined the prevalence of delirium, associated patient and hospitalization risk factors, as well as delirium-associated outcomes in patients undergoing a cranial procedure. METHODS Patient Selection, Data Collection, and Variable Definitions After obtaining institutional approval, 251 consecutive patients undergoing a cranial procedure over a 2-month period were identified. Inclusion criteria for the final cohort of patient included patients who had undergone at least 1 cranial procedure, were admitted to the hospital postoperatively, had documentation of delirium testing during the patient’s hospital stay, and were aged 18 years. Of all patients undergoing a cranial procedure during this cohort, 16 patients were excluded as either the procedure performed was an outpatient procedure and allowed for the patient to be discharged home directly afterward or the patient was in a persistent comatose state making delirium testing irrelevant. After exclusion of these patients, the final cohort included 235 patients. Patient, treatment, and outcome characteristics were determined from operative, radiology, pathology, and other clinical reports that were available through the electronic medical record. Steroid and benzodiazepine use were deemed positive if medications were given prior to a positive delirium score with the exception of medications given in the preoperative or intraoperative setting. Abnormal sodium was also noted only if the abnormal value preceded a positive delirium score. Postoperative infections excluded patients in which the primary diagnosis was an infection (i.e., falling under the infection category). Delirium Testing Delirium tests were administered by bedside nursing staff as part of the routine clinical care provided for patients at our institution and were not conducted for the sole purpose of this study. Delirium screening was performed using the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU) scoring
e2
www.SCIENCEDIRECT.com
system if patients were in the ICU or the Nursing Delirium Screening Scale (NuDESC) system if patients were in a medical/ surgical unit. Both scoring systems have been used previously to identify patients with delirium. The CAM-ICU test has demonstrated a sensitivity of 93%e100% and a specificity of 89%e100%, with a consistency evaluation of 0.79e0.96.12,13 The CAM-ICU evaluates delirium in patients from 4 different aspects: 1) acute changes in consciousness, or repeated fluctuations; 2) attention deficit; 3) thinking disorders; and 4) a change in the clarity of consciousness. The standard positive judgment, indicating a CAM-outcome suggestive of delirium, involves 1) and 2), plus either 3) or 4) (or both). The NuDESC evaluates delirium based on observation of the following 5 features: 1) disorientation; 2) inappropriate behavior; 3) inappropriate communication; 4) illusions/hallucinations; and 5) psychomotor retardation. Each item is scored based on its severity (score of 0e2). Original validation studies used a total score 2 to define delirium.14 Patients were deemed to screen positive for delirium if either the NuDESC or CAM-ICU score was above threshold at least once during their hospital admission. Patients did not undergo screening with the CAM-ICU system if they were triaged to a medical/surgical unit postoperatively and did not spend time in the ICU. Statistical Analysis Statistical analysis was performed using JMP Pro 13 software (SAS Institute, Cary, North Carolina, USA). Descriptive statistics were used to define the patient cohort. A univariate generalized logistic regression was first performed to identify patient and treatment characteristics predictive of delirium. Predictive variables from the univariate analysis with a P value 0.2 were then included in a multivariate generalized logistic regression. Excluded variables included new postoperative neurologic deficit that was highly correlated with presence of neurologic deficit (preoperative or postoperative), percentage of hospital stay in the ICU that was highly correlated with length of ICU stay, and movement disorder diagnosis that was composed of a small group of patients without any screening positive for delirium. The recursive partitioning algorithm in JMP Pro 13 was used to determine optimal cut-offs for age and length of ICU stay for predicting delirium. The binary variable produced was then included in the univariate logistic regression analysis. The Fisher exact test and the Wilcoxon ranksum test were performed to compare categorical and continuous clinical outcomes with delirium, respectively. The level of significance was 0.05 for all analyses. RESULTS Delirium Prevalence and Patient Characteristics A total of 235 patients who underwent a cranial-based surgery over a 2-month period were included in the cohort. Of these patients, 52 (22.13%) suffered delirium at some point during their hospital stay, scoring positive on either the NuDESC or CAM-ICU delirium screening tests at least once during their hospital admission. Patient characteristics for the cohort are depicted in Table 1. Several patient factors were associated with delirium on univariate logistic regression analysis (Table 2). When compared with patients without delirium, patients with delirium were older (unit odds ratio [OR] 1.03), more likely to have a
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.03.012
ORIGINAL ARTICLE RAMIN A. MORSHED ET AL.
diagnosis of hydrocephalus (OR 4.69) or intracranial infection (OR 15.17) (Figure 1A), and more likely to be admitted as a transfer from an outside facility or through the emergency department (OR 8.03). Figure 1B depicts a logistic probability plot of age as a predictor of delirium. Sex, minority status, number of prior-toadmission medications, prior craniotomies, body mass index, supra- versus infratentorial location, and a diagnosis of diabetes were not correlated with delirium. Recursive partitioning identified age 72.56 years as an optimal cut-off for predicting delirium with an OR of 4.61 (Table 2). Hospitalization Details Several hospitalization factors were associated with delirium on univariate logistic regression (Table 2). On average, patient’s screening positive for delirium had a longer ICU stay (9.0 vs. 1.6 days, unit OR 1.32). Figure 1C depicts a logistic probability plot of length of ICU stay as a predictor of delirium. A length of ICU stay of 8.3 days (95% CI, 6.5e11.2) conferred a 50% probability of delirium. Recursive partitioning identified length of ICU stay 5 days as optimal for predicting delirium (OR 18.2). Urgent or emergent surgery was also predictive of delirium (OR 6.52). Unexpectedly, a shorter length of surgery was associated with higher rates of delirium (P < 0.05) but was also significantly correlated with a diagnosis of hydrocephalus and intracranial infection (P < 0.001 and P < 0.05, respectively). In terms of medication use, the use of benzodiazepines was not found to correlate with delirium. Greater steroid use was associated with a decreased rate of delirium (OR 0.29) but was also highly associated with a tumor diagnosis that was conveyed as a decreased risk of delirium (OR 0.5). Abnormal sodium values (either hyper or hyponatremia) as well as a new postoperative infection were both predictive of delirium. The presence of a neurologic deficit (either present preoperatively or postoperatively) was also significantly associated with delirium (OR 6.8). Treatment with an awake craniotomy and estimated blood loss from surgery were not predictive of delirium. Further subgroup analyses were performed to determine other risk factors associated with specific pathologies. Within the group of patients with vascular pathology, vasospasm and hemorrhage (subarachnoid, extra-axial, or intraparenchymal) were both significantly associated with delirium (P ¼ 0.01 and P ¼ 0.005, respectively). Within the group of patients with tumors, intra-axial lesions had high rates of delirium when compared with extra-axial lesions (21.5% vs. 9.6%, respectively), with a trend toward significance (P ¼ 0.08). Multivariate Analysis Patient and hospitalization factors that were found to correlate with delirium on univariate logistic regression analysis were included in a multivariate generalized logistic regression analysis (Table 3). Age (unit OR 1.05), length of ICU stay (unit OR 1.23), and the presence of a neurologic deficit at any point during admission (OR 5.41) were significantly associated with the presence of delirium. When performing a multivariate analysis using the same variables within the cohorts of elective surgeries and urgent/emergent surgeries separately, length of ICU stay and age were both still significantly associated with delirium. However, the presence of a neurologic deficit was only found to
WORLD NEUROSURGERY -: e1-e7, - 2019
DELIRIUM RISK FACTORS AND ASSOCIATED OUTCOMES
Table 1. Patient Demographics Patients
Number [ 235
Sex (number, %) Male
116 (49.4%)
Female
119 (50.6%)
Age (number, 95% CI) (years)
52.6 (50.7e54.6)
BMI (number, 95% CI)
27.7 (26.9e28.5)
Race (number, %)* African American
9 (4.0%)
Asia
25 (11.2%)
Caucasian
142 (63.7%)
Hispanic
45 (20.2%)
Hawaiian/Pacific Islander
2 (0.9%)
Category (number, %) Tumor
117 (49.8%)
Vascular
56 (23.8%)
Hydrocephalus
23 (9.8%)
Infection
5 (2.1%)
Epilepsy
7 (3.0%)
Movement disorder
12 (5.1%)
Othery
15 (6.4%)
Pathology location (number, %) Supratentorial
200 (85.1%)
Infratentorial
35 (14.9%)
Prior cranial procedure (number, %)
75 (31.9%)
Transfer from OSH or ED admission (number, %)
80 (34.0%)
Diabetes (number, %) Prior-to-admission medications (number, 95% CI) Follow-up (number, 95% CI) (days)
32 (13.6%) 4.1 (3.7e4.6) 139 (128.0e149.4)
CI, confidence interval; BMI, body mass index; OSH, outside hospital; ED, emergency department. *Twelve patients with unknown race/ethnicity or declined to answer. yThis category included an arachnoid cyst fenestration, Chiari malformation decompression, cranioplasty, cerebrospinal fluid leak repair, idiopathic intracranial hypertension management with shunt placement, Rathke cleft cyst removal, and biopsies for sarcoidosis and other inflammatory processes.
be significantly associated with delirium in cases of elective surgeries. Outcomes Key hospital outcomes were also associated with delirium (Table 4). Patients with delirium had a mean length-of-admission >3.5-times that of the group without delirium (16.3 vs. 4.4 days). Furthermore, patients who were delirium-positive were significantly less likely to discharge home, and instead were placed in an
www.journals.elsevier.com/world-neurosurgery
e3
ORIGINAL ARTICLE RAMIN A. MORSHED ET AL.
DELIRIUM RISK FACTORS AND ASSOCIATED OUTCOMES
Table 2. Univariate Generalized Logistic Regression of Patient, Hospitalization, and Surgery Characteristics as Predictors of Delirium Characteristics Age (years) Age 72.56 years
Delirium Positive (N [ 52)
Delirium Negative (N [ 183)
OR [95% CI]
P Value
58.0 [53.9e62.0]
51.1 [48.9e53.3]
1.03 [1.01e1.06]
0.005
11 (52.4%)
10 (47.6%)
4.61 [1.84e11.6]
0.001
Sex
0.75
Male Female BMI Racial/ethnic minority*
27 (23.3%)
89 (76.7%)
1.14 [0.62e2.11]
25 (21.0%)
94 (79.0%)
Ref
27.0 [25.2e28.7]
27.9 [26.9e28.8]
0.98 [0.93e1.03]
0.37
22 (22.4%)
59 (72.8%)
1.52 [0.80e2.88]
0.24
19 (16.2%)
98 (83.8%)
0.50 [0.26e0.94]
0.04
Pathology Tumor Vascular
15 (26.8%)
41 (73.2%)
1.40 [0.70e2.81]
0.36
Hydrocephalus
12 (52.2%)
11 (47.8%)
4.69 [1.93e11.4]
0.0008
Infection
4 (80.0%)
1 (20.0%)
15.17 [1.66e138.8]
0.009
Epilepsy
0 (0%)
7 (100%)
0.00 [0.00e0.00]
0.35
Movement disorder
0 (0%)
12 (100%)
0.00 [0.00e0.00]
0.07
Supratentorial
46 (23.0%)
154 (77.0%)
Infratentorial
6 (17.1%)
29 (82.9%)
Pathology location
1.44 [0.56e3.7]
0.51
Prior cranial procedure
21 (28.0%)
54 (72.0%)
1.62 [0.85e3.06]
0.18
Transfer from OSH or ED admission
37 (46.3%)
43 (53.7%)
8.03 [4.03e16.02]
<0.0001
Diabetes
11 (34.4%)
21 (65.6%)
2.07 [0.92e4.63]
0.11
Psychiatric disorder diagnosis
16 (28.6%)
40 (71.4%)
1.59 [0.80e3.15]
0.19
4.6 [3.6e5.6]
4.0 [3.5e4.5]
1.05 [0.96e1.14]
0.30
6.52 [3.3e12.9]
<0.0001
Prior-to-admission medications (number) Urgent/emergent surgery
27 (50.9%)
26 (49.1%)
Non-urgent/elective surgery
25 (13.7%)
157 (86.3%)
Length of ICU stay (days)
9.0 [7.6e10.4]
1.6 [0.8e2.3]
1.32 [1.20e1.44]
<0.0001
Length ICU stay 5 days
28 (71.8%)
11 (28.2%)
18.24 [8.05e41.33]
<0.0001
47.2 [38.3e56.1]
33.9 [29.1e38.7]
1.01 [1.003e1.02]
0.01
% of hospital stay in ICU (%) Length of surgery (hours)
3.8 [3.2e4.3]
4.5 [4.2e4.8]
0.82 [0.69e0.97]
0.02
Estimated blood loss (mL)
192.2 [91.3e293.0]
159.2 [105.8e212.6]
1.00 [0.99e1.00]
0.58
Postoperative steroid use
30 (16.6%)
151 (83.4%)
0.29 [0.15e0.56]
0.0004
Postoperative benzodiazepine use
20 (27.8%)
52 (72.2%)
1.57 [0.83e3.00]
0.18
Abnormal sodiumy
31 (43.7%)
40 (56.3%)
5.79 [2.96e11.3]
<0.0001
New postoperative infectionz
8 (66.7%)
4 (33.3%)
10.23 [2.92e35.83]
Presence of neurologic deficitx
39 (41.1%)
56 (58.9%)
6.8 [3.37e13.73]
<0.0001
New/worsened postoperative neurologic deficit
15 (38.5%)
24 (61.5%)
2.69 [1.28e5.62]
0.009
1 (6.3%)
15 (93.7%)
0.22 [0.03e1.70]
0.21
Awake craniotomy
0.0003
OR, odds ratio; CI, confidence interval; BMI, body mass index; OSH, outside hospital; ED, emergency department; ICU, intensive care unit. *Twelve patients with unknown race/ethnicity or declined to answer. yEither hyponatremia or hypernatremia included. zPatients within the “infection” pathology category needed to have developed a new postoperative infection for inclusion. xNeurologic deficit that was present at any time during hospital stay, even preoperatively.
e4
www.SCIENCEDIRECT.com
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.03.012
ORIGINAL ARTICLE RAMIN A. MORSHED ET AL.
Figure 1. Delirium rates among different neurosurgical pathology (A), and logistic regression curves of age (B), and length of intensive care unit stay (C) as predictors
acute rehabilitation or skilled nursing facility, long-term acute care facility, or an outside hospital. Delirium was still significantly associated with longer length-of-admission and lower rates of discharging home when examining elective and urgent/emergent cases separately. Thirty-day unexpected readmission rates were not significantly different between groups within the entire cohort. However, when looking specifically within the subgroup of elective surgeries, delirium was associated with high rates of 30-day readmission (24.0% vs. 6.4% for patients without delirium) (P ¼ 0.004). When examining rates of death by end of follow-up, there was no difference between the groups with and without delirium. When subgroup analyses were performed within the vascular and tumor pathology groups as well as between elective versus urgent/emergent procedures, there was no difference in rates of death. However, for patients with age 72.56 years (the critical age threshold for predicting delirium), delirium was associated with a higher rate of death by end of follow-up (36.4% vs. 0% for patients without delirium). DISCUSSION Delirium is now being recognized as a major complication following surgery that has significant impact on clinical outcomes. Several reports demonstrate that delirium is associated with
WORLD NEUROSURGERY -: e1-e7, - 2019
DELIRIUM RISK FACTORS AND ASSOCIATED OUTCOMES
of delirium in a neurosurgical cohort. ICU, intensive care unit.
increased complication rates, prolonged hospital stay, and increased hospital costs.15-19 It is therefore important to identify at-risk populations so that preventative interventions may be instituted early during a patient’s hospital stay. Rates of delirium range from 31.8%e81.3% in both medical and surgical ICU patient populations.15,18,20-22 Neurosurgical patients may be at increased risk of delirium owing to a number of patient and hospitalization factors including long intensive care requirements as well as disease, metabolic, or medication-related derangements that impair mental status. However, specific structural pathology within the brain has also been associated with delirium.23 In this report, the rate of delirium was found to be approximately 22% but, of note, included a significant number of patients who did not require ICU admission. This is within the realm of prior reports pertaining to specific subgroups of neurosurgical patients with delirium rates ranging from 13.2%e42.2%.23-26 Our results provide insight into the prevalence of delirium within a typical neurosurgical service by including patients with a variety of pathologies requiring either elective operations or urgent/emergent interventions across all levels of care within the hospital. By including all patients undergoing a cranial procedure, this provides the opportunity to identify specific subpopulations at risk. Based on multivariate analysis, age, length of ICU stay, and the presence of a neurologic deficit were most predictive of
www.journals.elsevier.com/world-neurosurgery
e5
ORIGINAL ARTICLE RAMIN A. MORSHED ET AL.
DELIRIUM RISK FACTORS AND ASSOCIATED OUTCOMES
Table 3. Multivariate Generalized Logistic Regression for Predictors of Delirium Characteristic
OR [95% CI]
P Value
Presence of neurologic deficit
5.31 [1.87e15.11]
0.002
Length of ICU stay
1.23 [1.07e1.43]
0.004
Age
1.05 [1.01e1.08]
0.006
Abnormal sodium
0.21
Tumor diagnosis
0.39
Length of surgery
0.40
Postoperative benzodiazepine use
0.45
Postoperative steroid use
0.46
New postoperative infection
0.46
Transfer from OSH or ED admission
0.55
Diabetes
0.61
Hydrocephalus diagnosis
0.78
Infection diagnosis
0.99
Unit OR reported for length of ICU stay and age. OR, odds ratio; CI, confidence interval; ICU, intensive care unit; OSH, outside hospital; ED, emergency department.
delirium. We hypothesize that having a neurologic deficit is a harbinger for larger areas of cortical dysfunction or injury that predispose a patient to delirium. Although previously existing neurologic deficits may be a static risk factor, new postoperative deficits require investigation and may be reversible in nature depending on the etiology (e.g., vasospasm and worsening focal edema). They therefore represent a target for delirium prevention or reversal. Although the presence of a neurologic deficit was not a significant risk factor for delirium in the patients in which an urgent or emergent surgery was performed, this may be owing to the factor that there was a high rate of neurologic deficits in this subgroup (60.4% vs. 34.6% in the elective surgery cohort), potentially making this risk factor irrelevant. Based on univariate analysis, other risk factors included a diagnosis of hydrocephalus or intracranial infection, urgent or emergent procedural intervention, admission from an outside facility or through the emergency department, abnormal sodium, and a new postoperative extra-cranial infection. We hypothesize that a diagnosis of hydrocephalus and intracranial infection were strong
predictors of delirium because these diagnoses are often associated with generalized mental status changes that may precipitate a delirious state. Diagnosis with a movement disorder was associated with no cases of delirium in the current cohort, that is a contrast to prior reports.24,27 This may be because patients at our institution undergoing placement of deep-brain stimulation admitted electively for surgery had undergone extensive outpatient screening to ensure they would tolerate such a procedure and cognitively be able to adjust pulse-generator settings based on severity of symptoms. Although greater steroid use was associated with lower rates of delirium, steroid use was less likely to be used among patients with hydrocephalus or intracranial infection. Similarly, although shorter surgery duration was associated with delirium, these tended to be patients with hydrocephalus or intracranial infection undergoing shorter procedures including shunt placement, needle aspiration of abscess, or intracranial biopsy. Many of the prior reports examining delirium prevalence and associated risk factors in neurosurgical patients have been done in specific cohorts with traumatic, cerebrovascular, or movement disorder pathology. For patients with cerebral hematomas, Naidech et al.28 found that hematoma location (specifically within the right cerebral subcortical white matter or parahippocampus) correlated with delirium. Wang et al.26 examined delirium rates in a cohort of ICU cerebrovascular neurosurgical patients and found severity of illness, fever, use of physical restraints, and sleep deprivation were independent predictors of delirium. In a cohort of cerebrovascular surgery patients, Harasawa and Mizuno25 reported risk factors including dehydration, age, underlying disease, disturbance of consciousness, and anxiety or depression. In patients undergoing deep brain stimulation lead placement for Parkinson disease, Carlson et al.24 reported delirium risk factors including a history of delirium, age, and disease duration. However, in their cohort, delirium occurrence was not related to complex medication regimens or dementia. Furthermore, delirium did not seem to impact hospital stay as most patients in their cohort went home on postoperative day 1.24 Other reports in neurosurgical patients, including our present study, suggest that delirium impacts clinical outcomes. For example, Lukasiewicz et al.29 found that delirium was associated with higher rates of returning to the operating room, mortality, and longer length-of-stay in a cohort of patients undergoing subdural hematoma evacuation. There are several limitations associated with this retrospective study. Clinical history and outcomes required documentation in the medical chart. Screening was performed for clinical and not research purposes so the ability to detect differences in delirium rates based on patient or hospitalization factors may not be
Table 4. Outcomes Associated with Delirium
Length of admission (days, 95% CI)
Delirium Positive
Delirium Negative
P Value
16.3 (14.3e18.2)
4.36 (3.3e5.4)
<0.0001
Disposition home (number, %)
17 (32.7%)
164 (89.6%)
<0.0001
30-day unexpected readmission (number, %)
7 (13.5%)
13 (7.1%)
0.16
Death by end of follow-up
8 (15.4%)
22 (12.0%)
0.52
CI, confidence interval.
e6
www.SCIENCEDIRECT.com
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.03.012
ORIGINAL ARTICLE RAMIN A. MORSHED ET AL.
DELIRIUM RISK FACTORS AND ASSOCIATED OUTCOMES
optimized based on the size of the cohort. Furthermore, although the variables identified (i.e., age, length of ICU stay, and neurologic deficits) were predictive of delirium, it is unclear if a causal relationship exists. There are several avenues of further investigation. Whether delirium treatment or preventative methods in these at-risk subgroups of patients actually addresses delirium rates and prevents complications remains to be seen and is a point of further investigation. Furthermore, as longer ICU length of stay was a significant predictor of delirium, we are currently trialing triaging patients with similar neurosurgical pathology to step-down units to determine if the rate of delirium and other complications are lower.
2. Inouye SK, Charpentier PA. Precipitating factors for delirium in hospitalized elderly persons. Predictive model and interrelationship with baseline vulnerability. JAMA. 1996;275:852-857. 3. Marcantonio ER. Delirium in hospitalized older adults. N Engl J Med. 2017;377:1456-1466. 4. Inouye SK, Viscoli CM, Horwitz RI, Hurst LD, Tinetti ME. A predictive model for delirium in hospitalized elderly medical patients based on admission characteristics. Ann Intern Med. 1993; 119:474-481. 5. Inouye SK, Westendorp RGJ, Saczynski JS. Delirium in elderly people. Lancet (London, England). 2014;383:911-922. 6. Marcantonio ER, Goldman L, Mangione CM, et al. A clinical prediction rule for delirium after elective noncardiac surgery. JAMA. 1994;271: 134-139. 7. Lynch EP, Lazor MA, Gellis JE, Orav J, Goldman L, Marcantonio ER. The impact of postoperative pain on the development of postoperative delirium. Anesth Analg. 1998;86:781-785. 8. Marcantonio ER, Goldman L, Orav EJ, Cook EF, Lee TH. The association of intraoperative factors with the development of postoperative delirium. Am J Med. 1998;105:380-384. 9. Marcantonio ER, Juarez G, Goldman L, et al. The relationship of postoperative delirium with psychoactive medications. JAMA. 1994;272:1518-1522. 10. Fineberg SJ, Nandyala SV, Marquez-Lara A, Oglesby M, Patel AA, Singh K. Incidence and risk factors for postoperative delirium after lumbar spine surgery. Spine (Phila Pa 1976). 2013;38: 1790-1796. 11. Gleason LJ, Schmitt EM, Kosar CM, et al. Effect of delirium and other major complications on outcomes after elective surgery in older adults. JAMA Surg. 2015;150:1134-1140. 12. Ely EW, Inouye SK, Bernard GR, et al. Delirium in mechanically ventilated patients: validity and
Delirium is a common occurrence in neurosurgical patients undergoing cranial surgery. Several risk factors, especially older age, longer ICU stay, and the presence of a neurologic deficit are predictive of delirium occurrence. Age >72 years or length of ICU stay >5 days were identified as critical thresholds associated with a higher risk of delirium. Neurosurgical patients with delirium were found to have longer hospital admissions and were discharged home less frequently than patients without delirium. These findings may aide in identifying at-risk patients for delirium on a neurosurgical service to enact precautions preemptively.
reliability of the confusion assessment method for the intensive care unit (CAM-ICU). JAMA. 2001; 286:2703-2710.
REFERENCES 1. Marcantonio ER. Postoperative delirium: a 76year-old woman with delirium following surgery. JAMA. 2012;308:73-81.
CONCLUSIONS
13. Ely EW, Margolin R, Francis J, et al. Evaluation of delirium in critically ill patients: validation of the confusion assessment method for the intensive care unit (CAM-ICU). Crit Care Med. 2001;29: 1370-1379. 14. Gaudreau J-D, Gagnon P, Harel F, Tremblay A, Roy M-A. Fast, systematic, and continuous delirium assessment in hospitalized patients: the nursing delirium screening scale. J Pain Symptom Manage. 2005;29:368-375. 15. Ely EW, Gautam S, Margolin R, et al. The impact of delirium in the intensive care unit on hospital length of stay. Intensive Care Med. 2001;27: 1892-1900. 16. Girard TD, Jackson JC, Pandharipande PP, et al. Delirium as a predictor of long-term cognitive impairment in survivors of critical illness. Crit Care Med. 2010;38:1513-1520.
23. Matano F, Mizunari T, Yamada K, Kobayashi S, Murai Y, Morita A. Environmental and clinical risk factors for delirium in a neurosurgical center: a prospective study. World Neurosurg. 2017;103: 424-430. 24. Carlson JD, Neumiller JJ, Swain LDW, Mark J, McLeod P, Hirschauer J. Postoperative delirium in Parkinson’s disease patients following deep brain stimulation surgery. J Clin Neurosci Off J Neurosurg Soc Australas. 2014;21:1192-1195. 25. Harasawa N, Mizuno T. A novel scale predicting postoperative delirium (POD) in patients undergoing cerebrovascular surgery. Arch Gerontol Geriatr. 2014;59:264-271. 26. Wang J, Ji Y, Wang N, et al. Risk factors for the incidence of delirium in cerebrovascular patients in a neurosurgery intensive care unit: a prospective study. J Clin Nurs. 2018;27:407-415. 27. Tanaka M, Tani N, Maruo T, et al. Risk factors for postoperative delirium after deep brain stimulation surgery for Parkinson disease. World Neurosurg. 2018;114:e518-e523.
17. Koster S, Oosterveld FGJ, Hensens AG, Wijma A, van der Palen J. Delirium after cardiac surgery and predictive validity of a risk checklist. Ann Thorac Surg. 2008;86:1883-1887.
28. Naidech AM, Polnaszek KL, Berman MD, Voss JL. Hematoma locations predicting delirium symptoms after intracerebral hemorrhage. Neurocrit Care. 2016;24:397-403.
18. Milbrandt EB, Deppen S, Harrison PL, et al. Costs associated with delirium in mechanically ventilated patients. Crit Care Med. 2004;32:955-962.
29. Lukasiewicz AM, Grant RA, Basques BA, Webb ML, Samuel AM, Grauer JN. Patient factors associated with 30-day morbidity, mortality, and length of stay after surgery for subdural hematoma: a study of the American College of Surgeons National Surgical Quality Improvement Program. J Neurosurg. 2016;124:760-766.
19. Shehabi Y, Riker RR, Bokesch PM, Wisemandle W, Shintani A, Ely EW. Delirium duration and mortality in lightly sedated, mechanically ventilated intensive care patients. Crit Care Med. 2010;38:2311-2318. 20. Ouimet S, Kavanagh BP, Gottfried SB, Skrobik Y. Incidence, risk factors and consequences of ICU delirium. Intensive Care Med. 2007;33:66-73. 21. Pandharipande P, Cotton BA, Shintani A, et al. Prevalence and risk factors for development of delirium in surgical and trauma intensive care unit patients. J Trauma. 2008;65:34-41. 22. Thomason JWW, Shintani A, Peterson JF, Pun BT, Jackson JC, Ely EW. Intensive care unit delirium is an independent predictor of longer hospital stay: a prospective analysis of 261 non-ventilated patients. Crit Care. 2005;9:R375-R381.
WORLD NEUROSURGERY -: e1-e7, - 2019
Conflict of interest statement: The authors declare that the article content was composed in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Received 23 January 2019; accepted 2 March 2019 Citation: World Neurosurg. (2019). https://doi.org/10.1016/j.wneu.2019.03.012 Journal homepage: www.journals.elsevier.com/worldneurosurgery Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2019 Elsevier Inc. All rights reserved.
www.journals.elsevier.com/world-neurosurgery
e7
Delirium Risk Factors and Associated Outcomes in a Neurosurgical Cohort: A Case-Control Study Ramin A. Morshed1, Jacob S. Young1, Michael Safaee1, Sujatha Sankaran2, Mitchel S. Berger1, Michael W. McDermott1, Shawn L. Hervey-Jumper1
OBJECTIVE: There are limited reports examining delirium in cohorts of neurosurgical patients across inpatient settings without separation based on subspecialty distinction. It is of interest to identify consistent delirium risk factors across various cranial pathologies and inpatient settings that will inform future interventional studies.
-
METHODS: Delirium rates, patient and hospitalization risk factors, and clinical outcomes in 235 patients undergoing a cranial procedure were examined in a retrospective fashion.
neurologic deficit being consistent risk factors across inpatient settings. These results help identify at-risk patients for delirium on a neurosurgical service to enact interventions preemptively.
-
RESULTS: Fifty-two (22.1%) patients experienced delirium during their hospital stay. Patient factors predictive of delirium on univariate logistic regression were older age, a diagnosis of hydrocephalus or intracranial infection, transfer from an outside hospital, and admission through the emergency department. Hospitalization factors predictive of delirium included longer length of intensive care unit (ICU) stay, abnormal sodium values preceding delirium, a new postoperative infection, and the presence of a neurologic deficit. Using recursive partitioning, age ‡72.56 years and ICU length of stay ‡5 days were identified as critical thresholds for predicting delirium (odds ratio [OR] 4.61 and 18.2, respectively). On multivariate logistic regression analysis, age (unit OR 1.05), length of ICU stay (unit OR 1.2), and a neurologic deficit (OR 5.4) were predictive of delirium. Furthermore, delirium was also significantly associated with a longer length of admission as well as decreased likelihood for discharge home.
-
CONCLUSIONS: Delirium is a frequent occurrence after neurosurgery with older age, longer ICU stay, and a
-
Key words Cranial - Craniotomy - Delirium - Neurosurgery -
Abbreviations and Acronyms CAM-ICU: Confusion Assessment Method for the Intensive Care Unit ICU: Intensive care unit NuDESC: Nursing Delirium Screening Scale OR: Odds ratio
WORLD NEUROSURGERY -: e1-e7, - 2019
INTRODUCTION
D
elirium is defined as an acute confusional state with a fluctuating course that impairs patient cognition and ability to comply safely with medical care. It is increasingly being recognized as a major complication following surgery that has significant long-term morbidity for patients and is one of the most common complications after surgery in older adults.1 Preventing and treating delirium is challenging with little evidence of efficacy supporting commonly used pharmacologic interventions such as atypical antipsychotics. Rather, reorientation techniques, minimization of sedating medications, and non-pharmacological measures such as encouraging eyeglasses/hearing aids and mobilization are generally employed to reduce the rates of hospitalization-associated delirium. Risk factors for delirium are usually classified as either predisposing or precipitating factors.2,3 Older age, cognitive impairment, functional disabilities (e.g., poor vision, hearing loss), preexisting comorbidities, alcohol use, and depression have all been associated with an increased risk of delirium.4-6 Precipitating factors that have been previously identified include surgery, anesthesia, infection, sedative and anticholinergic medications, high pain levels, and anemia.7-9 Specifically, 1 large trial of over 1000 patients aged >50 years who underwent elective non-cardiac surgery identified postoperative delirium in 9% of patients, with delirium being associated with greater intraoperative blood
From the Departments of 1Neurological Surgery and 2Medicine, University of California, San Francisco, California, USA To whom correspondence should be addressed: Ramin A. Morshed, M.D. [E-mail: [email protected]] Citation: World Neurosurg. (2019). https://doi.org/10.1016/j.wneu.2019.03.012 Journal homepage: www.journals.elsevier.com/world-neurosurgery Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2019 Elsevier Inc. All rights reserved.
www.journals.elsevier.com/world-neurosurgery
e1
ORIGINAL ARTICLE RAMIN A. MORSHED ET AL.
DELIRIUM RISK FACTORS AND ASSOCIATED OUTCOMES
volume loss, more blood transfusions, and postoperative anemia but not intraoperative hemodynamic complications or the route of anesthesia.8 In the setting of lumbar spine surgery, preoperative comorbidities have been shown to correlate with risk of delirium, and postoperative delirium was associated with increased length of stay, increased costs, and higher mortality.10 Delirium has also been found to be a risk factor for discharge to a postacute nursing facility.11 Neurosurgical patients may be at increased risk of postoperative delirium given that these patients often have prolonged intensive care unit (ICU) care, require frequent neurologic checks with interruption of sleep, possess neurologic or metabolic derangements associated with neurosurgical disease, and have neurologic deficits that may impair communication or mental status. Given the severity and impact hospital-associated delirium has on patients and family caregivers as well as its long-term consequences on outcomes, a better understanding of the prevalence of and the risk factors for delirium in patients who undergo a cranial neurosurgical procedure is needed. Although there are prior reports that have examined delirium rates and risk factors in specific subgroups of neurosurgical patients, there is a lack of delirium data pertaining to general neurosurgical cohorts. Here, we examined the prevalence of delirium, associated patient and hospitalization risk factors, as well as delirium-associated outcomes in patients undergoing a cranial procedure. METHODS Patient Selection, Data Collection, and Variable Definitions After obtaining institutional approval, 251 consecutive patients undergoing a cranial procedure over a 2-month period were identified. Inclusion criteria for the final cohort of patient included patients who had undergone at least 1 cranial procedure, were admitted to the hospital postoperatively, had documentation of delirium testing during the patient’s hospital stay, and were aged 18 years. Of all patients undergoing a cranial procedure during this cohort, 16 patients were excluded as either the procedure performed was an outpatient procedure and allowed for the patient to be discharged home directly afterward or the patient was in a persistent comatose state making delirium testing irrelevant. After exclusion of these patients, the final cohort included 235 patients. Patient, treatment, and outcome characteristics were determined from operative, radiology, pathology, and other clinical reports that were available through the electronic medical record. Steroid and benzodiazepine use were deemed positive if medications were given prior to a positive delirium score with the exception of medications given in the preoperative or intraoperative setting. Abnormal sodium was also noted only if the abnormal value preceded a positive delirium score. Postoperative infections excluded patients in which the primary diagnosis was an infection (i.e., falling under the infection category). Delirium Testing Delirium tests were administered by bedside nursing staff as part of the routine clinical care provided for patients at our institution and were not conducted for the sole purpose of this study. Delirium screening was performed using the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU) scoring
e2
www.SCIENCEDIRECT.com
system if patients were in the ICU or the Nursing Delirium Screening Scale (NuDESC) system if patients were in a medical/ surgical unit. Both scoring systems have been used previously to identify patients with delirium. The CAM-ICU test has demonstrated a sensitivity of 93%e100% and a specificity of 89%e100%, with a consistency evaluation of 0.79e0.96.12,13 The CAM-ICU evaluates delirium in patients from 4 different aspects: 1) acute changes in consciousness, or repeated fluctuations; 2) attention deficit; 3) thinking disorders; and 4) a change in the clarity of consciousness. The standard positive judgment, indicating a CAM-outcome suggestive of delirium, involves 1) and 2), plus either 3) or 4) (or both). The NuDESC evaluates delirium based on observation of the following 5 features: 1) disorientation; 2) inappropriate behavior; 3) inappropriate communication; 4) illusions/hallucinations; and 5) psychomotor retardation. Each item is scored based on its severity (score of 0e2). Original validation studies used a total score 2 to define delirium.14 Patients were deemed to screen positive for delirium if either the NuDESC or CAM-ICU score was above threshold at least once during their hospital admission. Patients did not undergo screening with the CAM-ICU system if they were triaged to a medical/surgical unit postoperatively and did not spend time in the ICU. Statistical Analysis Statistical analysis was performed using JMP Pro 13 software (SAS Institute, Cary, North Carolina, USA). Descriptive statistics were used to define the patient cohort. A univariate generalized logistic regression was first performed to identify patient and treatment characteristics predictive of delirium. Predictive variables from the univariate analysis with a P value 0.2 were then included in a multivariate generalized logistic regression. Excluded variables included new postoperative neurologic deficit that was highly correlated with presence of neurologic deficit (preoperative or postoperative), percentage of hospital stay in the ICU that was highly correlated with length of ICU stay, and movement disorder diagnosis that was composed of a small group of patients without any screening positive for delirium. The recursive partitioning algorithm in JMP Pro 13 was used to determine optimal cut-offs for age and length of ICU stay for predicting delirium. The binary variable produced was then included in the univariate logistic regression analysis. The Fisher exact test and the Wilcoxon ranksum test were performed to compare categorical and continuous clinical outcomes with delirium, respectively. The level of significance was 0.05 for all analyses. RESULTS Delirium Prevalence and Patient Characteristics A total of 235 patients who underwent a cranial-based surgery over a 2-month period were included in the cohort. Of these patients, 52 (22.13%) suffered delirium at some point during their hospital stay, scoring positive on either the NuDESC or CAM-ICU delirium screening tests at least once during their hospital admission. Patient characteristics for the cohort are depicted in Table 1. Several patient factors were associated with delirium on univariate logistic regression analysis (Table 2). When compared with patients without delirium, patients with delirium were older (unit odds ratio [OR] 1.03), more likely to have a
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.03.012
ORIGINAL ARTICLE RAMIN A. MORSHED ET AL.
diagnosis of hydrocephalus (OR 4.69) or intracranial infection (OR 15.17) (Figure 1A), and more likely to be admitted as a transfer from an outside facility or through the emergency department (OR 8.03). Figure 1B depicts a logistic probability plot of age as a predictor of delirium. Sex, minority status, number of prior-toadmission medications, prior craniotomies, body mass index, supra- versus infratentorial location, and a diagnosis of diabetes were not correlated with delirium. Recursive partitioning identified age 72.56 years as an optimal cut-off for predicting delirium with an OR of 4.61 (Table 2). Hospitalization Details Several hospitalization factors were associated with delirium on univariate logistic regression (Table 2). On average, patient’s screening positive for delirium had a longer ICU stay (9.0 vs. 1.6 days, unit OR 1.32). Figure 1C depicts a logistic probability plot of length of ICU stay as a predictor of delirium. A length of ICU stay of 8.3 days (95% CI, 6.5e11.2) conferred a 50% probability of delirium. Recursive partitioning identified length of ICU stay 5 days as optimal for predicting delirium (OR 18.2). Urgent or emergent surgery was also predictive of delirium (OR 6.52). Unexpectedly, a shorter length of surgery was associated with higher rates of delirium (P < 0.05) but was also significantly correlated with a diagnosis of hydrocephalus and intracranial infection (P < 0.001 and P < 0.05, respectively). In terms of medication use, the use of benzodiazepines was not found to correlate with delirium. Greater steroid use was associated with a decreased rate of delirium (OR 0.29) but was also highly associated with a tumor diagnosis that was conveyed as a decreased risk of delirium (OR 0.5). Abnormal sodium values (either hyper or hyponatremia) as well as a new postoperative infection were both predictive of delirium. The presence of a neurologic deficit (either present preoperatively or postoperatively) was also significantly associated with delirium (OR 6.8). Treatment with an awake craniotomy and estimated blood loss from surgery were not predictive of delirium. Further subgroup analyses were performed to determine other risk factors associated with specific pathologies. Within the group of patients with vascular pathology, vasospasm and hemorrhage (subarachnoid, extra-axial, or intraparenchymal) were both significantly associated with delirium (P ¼ 0.01 and P ¼ 0.005, respectively). Within the group of patients with tumors, intra-axial lesions had high rates of delirium when compared with extra-axial lesions (21.5% vs. 9.6%, respectively), with a trend toward significance (P ¼ 0.08). Multivariate Analysis Patient and hospitalization factors that were found to correlate with delirium on univariate logistic regression analysis were included in a multivariate generalized logistic regression analysis (Table 3). Age (unit OR 1.05), length of ICU stay (unit OR 1.23), and the presence of a neurologic deficit at any point during admission (OR 5.41) were significantly associated with the presence of delirium. When performing a multivariate analysis using the same variables within the cohorts of elective surgeries and urgent/emergent surgeries separately, length of ICU stay and age were both still significantly associated with delirium. However, the presence of a neurologic deficit was only found to
WORLD NEUROSURGERY -: e1-e7, - 2019
DELIRIUM RISK FACTORS AND ASSOCIATED OUTCOMES
Table 1. Patient Demographics Patients
Number [ 235
Sex (number, %) Male
116 (49.4%)
Female
119 (50.6%)
Age (number, 95% CI) (years)
52.6 (50.7e54.6)
BMI (number, 95% CI)
27.7 (26.9e28.5)
Race (number, %)* African American
9 (4.0%)
Asia
25 (11.2%)
Caucasian
142 (63.7%)
Hispanic
45 (20.2%)
Hawaiian/Pacific Islander
2 (0.9%)
Category (number, %) Tumor
117 (49.8%)
Vascular
56 (23.8%)
Hydrocephalus
23 (9.8%)
Infection
5 (2.1%)
Epilepsy
7 (3.0%)
Movement disorder
12 (5.1%)
Othery
15 (6.4%)
Pathology location (number, %) Supratentorial
200 (85.1%)
Infratentorial
35 (14.9%)
Prior cranial procedure (number, %)
75 (31.9%)
Transfer from OSH or ED admission (number, %)
80 (34.0%)
Diabetes (number, %) Prior-to-admission medications (number, 95% CI) Follow-up (number, 95% CI) (days)
32 (13.6%) 4.1 (3.7e4.6) 139 (128.0e149.4)
CI, confidence interval; BMI, body mass index; OSH, outside hospital; ED, emergency department. *Twelve patients with unknown race/ethnicity or declined to answer. yThis category included an arachnoid cyst fenestration, Chiari malformation decompression, cranioplasty, cerebrospinal fluid leak repair, idiopathic intracranial hypertension management with shunt placement, Rathke cleft cyst removal, and biopsies for sarcoidosis and other inflammatory processes.
be significantly associated with delirium in cases of elective surgeries. Outcomes Key hospital outcomes were also associated with delirium (Table 4). Patients with delirium had a mean length-of-admission >3.5-times that of the group without delirium (16.3 vs. 4.4 days). Furthermore, patients who were delirium-positive were significantly less likely to discharge home, and instead were placed in an
www.journals.elsevier.com/world-neurosurgery
e3
ORIGINAL ARTICLE RAMIN A. MORSHED ET AL.
DELIRIUM RISK FACTORS AND ASSOCIATED OUTCOMES
Table 2. Univariate Generalized Logistic Regression of Patient, Hospitalization, and Surgery Characteristics as Predictors of Delirium Characteristics Age (years) Age 72.56 years
Delirium Positive (N [ 52)
Delirium Negative (N [ 183)
OR [95% CI]
P Value
58.0 [53.9e62.0]
51.1 [48.9e53.3]
1.03 [1.01e1.06]
0.005
11 (52.4%)
10 (47.6%)
4.61 [1.84e11.6]
0.001
Sex
0.75
Male Female BMI Racial/ethnic minority*
27 (23.3%)
89 (76.7%)
1.14 [0.62e2.11]
25 (21.0%)
94 (79.0%)
Ref
27.0 [25.2e28.7]
27.9 [26.9e28.8]
0.98 [0.93e1.03]
0.37
22 (22.4%)
59 (72.8%)
1.52 [0.80e2.88]
0.24
19 (16.2%)
98 (83.8%)
0.50 [0.26e0.94]
0.04
Pathology Tumor Vascular
15 (26.8%)
41 (73.2%)
1.40 [0.70e2.81]
0.36
Hydrocephalus
12 (52.2%)
11 (47.8%)
4.69 [1.93e11.4]
0.0008
Infection
4 (80.0%)
1 (20.0%)
15.17 [1.66e138.8]
0.009
Epilepsy
0 (0%)
7 (100%)
0.00 [0.00e0.00]
0.35
Movement disorder
0 (0%)
12 (100%)
0.00 [0.00e0.00]
0.07
Supratentorial
46 (23.0%)
154 (77.0%)
Infratentorial
6 (17.1%)
29 (82.9%)
Pathology location
1.44 [0.56e3.7]
0.51
Prior cranial procedure
21 (28.0%)
54 (72.0%)
1.62 [0.85e3.06]
0.18
Transfer from OSH or ED admission
37 (46.3%)
43 (53.7%)
8.03 [4.03e16.02]
<0.0001
Diabetes
11 (34.4%)
21 (65.6%)
2.07 [0.92e4.63]
0.11
Psychiatric disorder diagnosis
16 (28.6%)
40 (71.4%)
1.59 [0.80e3.15]
0.19
4.6 [3.6e5.6]
4.0 [3.5e4.5]
1.05 [0.96e1.14]
0.30
6.52 [3.3e12.9]
<0.0001
Prior-to-admission medications (number) Urgent/emergent surgery
27 (50.9%)
26 (49.1%)
Non-urgent/elective surgery
25 (13.7%)
157 (86.3%)
Length of ICU stay (days)
9.0 [7.6e10.4]
1.6 [0.8e2.3]
1.32 [1.20e1.44]
<0.0001
Length ICU stay 5 days
28 (71.8%)
11 (28.2%)
18.24 [8.05e41.33]
<0.0001
47.2 [38.3e56.1]
33.9 [29.1e38.7]
1.01 [1.003e1.02]
0.01
% of hospital stay in ICU (%) Length of surgery (hours)
3.8 [3.2e4.3]
4.5 [4.2e4.8]
0.82 [0.69e0.97]
0.02
Estimated blood loss (mL)
192.2 [91.3e293.0]
159.2 [105.8e212.6]
1.00 [0.99e1.00]
0.58
Postoperative steroid use
30 (16.6%)
151 (83.4%)
0.29 [0.15e0.56]
0.0004
Postoperative benzodiazepine use
20 (27.8%)
52 (72.2%)
1.57 [0.83e3.00]
0.18
Abnormal sodiumy
31 (43.7%)
40 (56.3%)
5.79 [2.96e11.3]
<0.0001
New postoperative infectionz
8 (66.7%)
4 (33.3%)
10.23 [2.92e35.83]
Presence of neurologic deficitx
39 (41.1%)
56 (58.9%)
6.8 [3.37e13.73]
<0.0001
New/worsened postoperative neurologic deficit
15 (38.5%)
24 (61.5%)
2.69 [1.28e5.62]
0.009
1 (6.3%)
15 (93.7%)
0.22 [0.03e1.70]
0.21
Awake craniotomy
0.0003
OR, odds ratio; CI, confidence interval; BMI, body mass index; OSH, outside hospital; ED, emergency department; ICU, intensive care unit. *Twelve patients with unknown race/ethnicity or declined to answer. yEither hyponatremia or hypernatremia included. zPatients within the “infection” pathology category needed to have developed a new postoperative infection for inclusion. xNeurologic deficit that was present at any time during hospital stay, even preoperatively.
e4
www.SCIENCEDIRECT.com
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.03.012
ORIGINAL ARTICLE RAMIN A. MORSHED ET AL.
Figure 1. Delirium rates among different neurosurgical pathology (A), and logistic regression curves of age (B), and length of intensive care unit stay (C) as predictors
acute rehabilitation or skilled nursing facility, long-term acute care facility, or an outside hospital. Delirium was still significantly associated with longer length-of-admission and lower rates of discharging home when examining elective and urgent/emergent cases separately. Thirty-day unexpected readmission rates were not significantly different between groups within the entire cohort. However, when looking specifically within the subgroup of elective surgeries, delirium was associated with high rates of 30-day readmission (24.0% vs. 6.4% for patients without delirium) (P ¼ 0.004). When examining rates of death by end of follow-up, there was no difference between the groups with and without delirium. When subgroup analyses were performed within the vascular and tumor pathology groups as well as between elective versus urgent/emergent procedures, there was no difference in rates of death. However, for patients with age 72.56 years (the critical age threshold for predicting delirium), delirium was associated with a higher rate of death by end of follow-up (36.4% vs. 0% for patients without delirium). DISCUSSION Delirium is now being recognized as a major complication following surgery that has significant impact on clinical outcomes. Several reports demonstrate that delirium is associated with
WORLD NEUROSURGERY -: e1-e7, - 2019
DELIRIUM RISK FACTORS AND ASSOCIATED OUTCOMES
of delirium in a neurosurgical cohort. ICU, intensive care unit.
increased complication rates, prolonged hospital stay, and increased hospital costs.15-19 It is therefore important to identify at-risk populations so that preventative interventions may be instituted early during a patient’s hospital stay. Rates of delirium range from 31.8%e81.3% in both medical and surgical ICU patient populations.15,18,20-22 Neurosurgical patients may be at increased risk of delirium owing to a number of patient and hospitalization factors including long intensive care requirements as well as disease, metabolic, or medication-related derangements that impair mental status. However, specific structural pathology within the brain has also been associated with delirium.23 In this report, the rate of delirium was found to be approximately 22% but, of note, included a significant number of patients who did not require ICU admission. This is within the realm of prior reports pertaining to specific subgroups of neurosurgical patients with delirium rates ranging from 13.2%e42.2%.23-26 Our results provide insight into the prevalence of delirium within a typical neurosurgical service by including patients with a variety of pathologies requiring either elective operations or urgent/emergent interventions across all levels of care within the hospital. By including all patients undergoing a cranial procedure, this provides the opportunity to identify specific subpopulations at risk. Based on multivariate analysis, age, length of ICU stay, and the presence of a neurologic deficit were most predictive of
www.journals.elsevier.com/world-neurosurgery
e5
ORIGINAL ARTICLE RAMIN A. MORSHED ET AL.
DELIRIUM RISK FACTORS AND ASSOCIATED OUTCOMES
Table 3. Multivariate Generalized Logistic Regression for Predictors of Delirium Characteristic
OR [95% CI]
P Value
Presence of neurologic deficit
5.31 [1.87e15.11]
0.002
Length of ICU stay
1.23 [1.07e1.43]
0.004
Age
1.05 [1.01e1.08]
0.006
Abnormal sodium
0.21
Tumor diagnosis
0.39
Length of surgery
0.40
Postoperative benzodiazepine use
0.45
Postoperative steroid use
0.46
New postoperative infection
0.46
Transfer from OSH or ED admission
0.55
Diabetes
0.61
Hydrocephalus diagnosis
0.78
Infection diagnosis
0.99
Unit OR reported for length of ICU stay and age. OR, odds ratio; CI, confidence interval; ICU, intensive care unit; OSH, outside hospital; ED, emergency department.
delirium. We hypothesize that having a neurologic deficit is a harbinger for larger areas of cortical dysfunction or injury that predispose a patient to delirium. Although previously existing neurologic deficits may be a static risk factor, new postoperative deficits require investigation and may be reversible in nature depending on the etiology (e.g., vasospasm and worsening focal edema). They therefore represent a target for delirium prevention or reversal. Although the presence of a neurologic deficit was not a significant risk factor for delirium in the patients in which an urgent or emergent surgery was performed, this may be owing to the factor that there was a high rate of neurologic deficits in this subgroup (60.4% vs. 34.6% in the elective surgery cohort), potentially making this risk factor irrelevant. Based on univariate analysis, other risk factors included a diagnosis of hydrocephalus or intracranial infection, urgent or emergent procedural intervention, admission from an outside facility or through the emergency department, abnormal sodium, and a new postoperative extra-cranial infection. We hypothesize that a diagnosis of hydrocephalus and intracranial infection were strong
predictors of delirium because these diagnoses are often associated with generalized mental status changes that may precipitate a delirious state. Diagnosis with a movement disorder was associated with no cases of delirium in the current cohort, that is a contrast to prior reports.24,27 This may be because patients at our institution undergoing placement of deep-brain stimulation admitted electively for surgery had undergone extensive outpatient screening to ensure they would tolerate such a procedure and cognitively be able to adjust pulse-generator settings based on severity of symptoms. Although greater steroid use was associated with lower rates of delirium, steroid use was less likely to be used among patients with hydrocephalus or intracranial infection. Similarly, although shorter surgery duration was associated with delirium, these tended to be patients with hydrocephalus or intracranial infection undergoing shorter procedures including shunt placement, needle aspiration of abscess, or intracranial biopsy. Many of the prior reports examining delirium prevalence and associated risk factors in neurosurgical patients have been done in specific cohorts with traumatic, cerebrovascular, or movement disorder pathology. For patients with cerebral hematomas, Naidech et al.28 found that hematoma location (specifically within the right cerebral subcortical white matter or parahippocampus) correlated with delirium. Wang et al.26 examined delirium rates in a cohort of ICU cerebrovascular neurosurgical patients and found severity of illness, fever, use of physical restraints, and sleep deprivation were independent predictors of delirium. In a cohort of cerebrovascular surgery patients, Harasawa and Mizuno25 reported risk factors including dehydration, age, underlying disease, disturbance of consciousness, and anxiety or depression. In patients undergoing deep brain stimulation lead placement for Parkinson disease, Carlson et al.24 reported delirium risk factors including a history of delirium, age, and disease duration. However, in their cohort, delirium occurrence was not related to complex medication regimens or dementia. Furthermore, delirium did not seem to impact hospital stay as most patients in their cohort went home on postoperative day 1.24 Other reports in neurosurgical patients, including our present study, suggest that delirium impacts clinical outcomes. For example, Lukasiewicz et al.29 found that delirium was associated with higher rates of returning to the operating room, mortality, and longer length-of-stay in a cohort of patients undergoing subdural hematoma evacuation. There are several limitations associated with this retrospective study. Clinical history and outcomes required documentation in the medical chart. Screening was performed for clinical and not research purposes so the ability to detect differences in delirium rates based on patient or hospitalization factors may not be
Table 4. Outcomes Associated with Delirium
Length of admission (days, 95% CI)
Delirium Positive
Delirium Negative
P Value
16.3 (14.3e18.2)
4.36 (3.3e5.4)
<0.0001
Disposition home (number, %)
17 (32.7%)
164 (89.6%)
<0.0001
30-day unexpected readmission (number, %)
7 (13.5%)
13 (7.1%)
0.16
Death by end of follow-up
8 (15.4%)
22 (12.0%)
0.52
CI, confidence interval.
e6
www.SCIENCEDIRECT.com
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.03.012
ORIGINAL ARTICLE RAMIN A. MORSHED ET AL.
DELIRIUM RISK FACTORS AND ASSOCIATED OUTCOMES
optimized based on the size of the cohort. Furthermore, although the variables identified (i.e., age, length of ICU stay, and neurologic deficits) were predictive of delirium, it is unclear if a causal relationship exists. There are several avenues of further investigation. Whether delirium treatment or preventative methods in these at-risk subgroups of patients actually addresses delirium rates and prevents complications remains to be seen and is a point of further investigation. Furthermore, as longer ICU length of stay was a significant predictor of delirium, we are currently trialing triaging patients with similar neurosurgical pathology to step-down units to determine if the rate of delirium and other complications are lower.
2. Inouye SK, Charpentier PA. Precipitating factors for delirium in hospitalized elderly persons. Predictive model and interrelationship with baseline vulnerability. JAMA. 1996;275:852-857. 3. Marcantonio ER. Delirium in hospitalized older adults. N Engl J Med. 2017;377:1456-1466. 4. Inouye SK, Viscoli CM, Horwitz RI, Hurst LD, Tinetti ME. A predictive model for delirium in hospitalized elderly medical patients based on admission characteristics. Ann Intern Med. 1993; 119:474-481. 5. Inouye SK, Westendorp RGJ, Saczynski JS. Delirium in elderly people. Lancet (London, England). 2014;383:911-922. 6. Marcantonio ER, Goldman L, Mangione CM, et al. A clinical prediction rule for delirium after elective noncardiac surgery. JAMA. 1994;271: 134-139. 7. Lynch EP, Lazor MA, Gellis JE, Orav J, Goldman L, Marcantonio ER. The impact of postoperative pain on the development of postoperative delirium. Anesth Analg. 1998;86:781-785. 8. Marcantonio ER, Goldman L, Orav EJ, Cook EF, Lee TH. The association of intraoperative factors with the development of postoperative delirium. Am J Med. 1998;105:380-384. 9. Marcantonio ER, Juarez G, Goldman L, et al. The relationship of postoperative delirium with psychoactive medications. JAMA. 1994;272:1518-1522. 10. Fineberg SJ, Nandyala SV, Marquez-Lara A, Oglesby M, Patel AA, Singh K. Incidence and risk factors for postoperative delirium after lumbar spine surgery. Spine (Phila Pa 1976). 2013;38: 1790-1796. 11. Gleason LJ, Schmitt EM, Kosar CM, et al. Effect of delirium and other major complications on outcomes after elective surgery in older adults. JAMA Surg. 2015;150:1134-1140. 12. Ely EW, Inouye SK, Bernard GR, et al. Delirium in mechanically ventilated patients: validity and
Delirium is a common occurrence in neurosurgical patients undergoing cranial surgery. Several risk factors, especially older age, longer ICU stay, and the presence of a neurologic deficit are predictive of delirium occurrence. Age >72 years or length of ICU stay >5 days were identified as critical thresholds associated with a higher risk of delirium. Neurosurgical patients with delirium were found to have longer hospital admissions and were discharged home less frequently than patients without delirium. These findings may aide in identifying at-risk patients for delirium on a neurosurgical service to enact precautions preemptively.
reliability of the confusion assessment method for the intensive care unit (CAM-ICU). JAMA. 2001; 286:2703-2710.
REFERENCES 1. Marcantonio ER. Postoperative delirium: a 76year-old woman with delirium following surgery. JAMA. 2012;308:73-81.
CONCLUSIONS
13. Ely EW, Margolin R, Francis J, et al. Evaluation of delirium in critically ill patients: validation of the confusion assessment method for the intensive care unit (CAM-ICU). Crit Care Med. 2001;29: 1370-1379. 14. Gaudreau J-D, Gagnon P, Harel F, Tremblay A, Roy M-A. Fast, systematic, and continuous delirium assessment in hospitalized patients: the nursing delirium screening scale. J Pain Symptom Manage. 2005;29:368-375. 15. Ely EW, Gautam S, Margolin R, et al. The impact of delirium in the intensive care unit on hospital length of stay. Intensive Care Med. 2001;27: 1892-1900. 16. Girard TD, Jackson JC, Pandharipande PP, et al. Delirium as a predictor of long-term cognitive impairment in survivors of critical illness. Crit Care Med. 2010;38:1513-1520.
23. Matano F, Mizunari T, Yamada K, Kobayashi S, Murai Y, Morita A. Environmental and clinical risk factors for delirium in a neurosurgical center: a prospective study. World Neurosurg. 2017;103: 424-430. 24. Carlson JD, Neumiller JJ, Swain LDW, Mark J, McLeod P, Hirschauer J. Postoperative delirium in Parkinson’s disease patients following deep brain stimulation surgery. J Clin Neurosci Off J Neurosurg Soc Australas. 2014;21:1192-1195. 25. Harasawa N, Mizuno T. A novel scale predicting postoperative delirium (POD) in patients undergoing cerebrovascular surgery. Arch Gerontol Geriatr. 2014;59:264-271. 26. Wang J, Ji Y, Wang N, et al. Risk factors for the incidence of delirium in cerebrovascular patients in a neurosurgery intensive care unit: a prospective study. J Clin Nurs. 2018;27:407-415. 27. Tanaka M, Tani N, Maruo T, et al. Risk factors for postoperative delirium after deep brain stimulation surgery for Parkinson disease. World Neurosurg. 2018;114:e518-e523.
17. Koster S, Oosterveld FGJ, Hensens AG, Wijma A, van der Palen J. Delirium after cardiac surgery and predictive validity of a risk checklist. Ann Thorac Surg. 2008;86:1883-1887.
28. Naidech AM, Polnaszek KL, Berman MD, Voss JL. Hematoma locations predicting delirium symptoms after intracerebral hemorrhage. Neurocrit Care. 2016;24:397-403.
18. Milbrandt EB, Deppen S, Harrison PL, et al. Costs associated with delirium in mechanically ventilated patients. Crit Care Med. 2004;32:955-962.
29. Lukasiewicz AM, Grant RA, Basques BA, Webb ML, Samuel AM, Grauer JN. Patient factors associated with 30-day morbidity, mortality, and length of stay after surgery for subdural hematoma: a study of the American College of Surgeons National Surgical Quality Improvement Program. J Neurosurg. 2016;124:760-766.
19. Shehabi Y, Riker RR, Bokesch PM, Wisemandle W, Shintani A, Ely EW. Delirium duration and mortality in lightly sedated, mechanically ventilated intensive care patients. Crit Care Med. 2010;38:2311-2318. 20. Ouimet S, Kavanagh BP, Gottfried SB, Skrobik Y. Incidence, risk factors and consequences of ICU delirium. Intensive Care Med. 2007;33:66-73. 21. Pandharipande P, Cotton BA, Shintani A, et al. Prevalence and risk factors for development of delirium in surgical and trauma intensive care unit patients. J Trauma. 2008;65:34-41. 22. Thomason JWW, Shintani A, Peterson JF, Pun BT, Jackson JC, Ely EW. Intensive care unit delirium is an independent predictor of longer hospital stay: a prospective analysis of 261 non-ventilated patients. Crit Care. 2005;9:R375-R381.
WORLD NEUROSURGERY -: e1-e7, - 2019
Conflict of interest statement: The authors declare that the article content was composed in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Received 23 January 2019; accepted 2 March 2019 Citation: World Neurosurg. (2019). https://doi.org/10.1016/j.wneu.2019.03.012 Journal homepage: www.journals.elsevier.com/worldneurosurgery Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2019 Elsevier Inc. All rights reserved.
www.journals.elsevier.com/world-neurosurgery
e7