Abstract

Background and purpose: several studies have assessed delirium post-stroke but conflicting results have been obtained. Also, the natural history and outcome of delirium post-stroke need to be fully elucidated.

Methodology: eligible stroke patients were assessed for delirium on admission and for four consecutive weeks using the Confusion Assessment Method (CAM). Risk factors for delirium were recorded. Our outcome measures were length of stay, inpatient mortality and discharge destination.

Results: of 110 eligible patients, 82 were recruited over 7 months. Delirium was detected in 23 patients (28%); 21 of these were delirious on their first assessment. Sixty-nine per cent of patients who had four weekly assessments were delirious at 4 weeks. Multivariate logistic regression analysis was performed, and two models were identified. With unsafe swallow in the analysis, delirium was associated with an unsafe swallow on admission (OR 28.4, P<0.001), Barthel score < 10 (OR 32.1, P = 0.004) and poor vision pre-stroke (OR 110.8, P = 0.01). With unsafe swallow removed from the analysis, delirium was associated with an admission C-reactive protein (CRP) > 5 mg/l (OR 10.2, P = 0.009), Barthel score < 10 (OR 46.5, P = 0.001) and poor vision pre-stroke (OR 85.2, P = 0.01). Delirious patients had a higher mortality (30.4% vs. 1.7%, P<0.001), longer length of stay (62.2 vs. 28.9 days, P<0.001) and increased risk of institutionalisation (43.7 vs. 5.2%, OR 14, P<0.001).

Conclusions: delirium is common post-stroke. Most cases develop at stroke onset and remain delirious for an appreciable period. Delirium onset is associated with stroke severity (low admission Barthel), unsafe swallow on admission, poor vision pre-stroke and a raised admission CRP. Delirium is a marker of poor prognosis.

Background

Although stroke is a known predisposing factor for delirium, there have only been a few prospective studies of delirium in the acute stroke setting and these have given conflicting results with prevalence estimates ranging from 13 to 48% [1–5]. In addition, different independent risk factors for delirium post-stroke have been identified including left-sided strokes [2], intracerebral haemorrhages [3, 4], cardioembolic stroke [4], total anterior circulation infarction (TACI) [4], age [4], neglect [3], pre-existing cognitive impairment [4] and metabolic disorders post-stroke [4]. These studies used different screening tools and different methodologies but did not include the natural history of the delirium after onset although one study has 12-month follow-up data [4]. The purpose of this prospective, observational study was to study the prevalence of delirium in the acute stroke setting, the risk factors and natural history of delirium post-stroke and determinants of outcome of patients who develop delirium post-stroke.

Patients and methods

Patients

The study population was recruited over a 7-month period. All consecutive stroke patients admitted to the Stroke Unit at King's College Hospital, London, were eligible for the study. The unit admits patients whose pathology is felt to be vascular in origin (ischaemic or haemorrhagic) with the exclusion of subarachnoid haemorrhage. The hospital has a policy of admitting all stroke patients directly to the Stroke Unit. The inclusion criteria for the study were (i) an admission diagnosis of cerebral infarction or intracerebral haemorrhage and (ii) a delirium assessment within 4 days of admission. Our exclusion criteria were (i) patients whose symptoms lasted <24 h, (ii) patients who did not speak English and (iii) patients with a Glasgow Coma Scale (GCS) score of <8.

Risk factor assessment

Potential predisposing risk factors for delirium were considered. We recorded the following patient characteristics factors: age, sex, pre-stroke medical history, including dementia, psychiatric history, delirium, alcohol abuse defined as >21 units per week for males and 15 units per week for females, reduced hearing, reduced vision, smoking, hypertension, diabetes, history of stroke/transient ischaemic attack (TIA), atrial fibrillation, hyperlipidaemia and ischaemic heart disease. An a priori category of four or more medications on admission was created to assess this construct. Stroke type was defined using the Bamford classification [6] and the Trial of Org 10172 in Acute Stroke Treatment (TOAST) criteria [7].

Diagnosis of delirium

Patients were screened for delirium within 4 days of admission using the Confusion Assessment Method (CAM) [8]. As the researcher collecting the data was based in the Stroke Unit for the duration of the study, in most cases, the initial assessment was carried out within 1 day of admission. The CAM was developed in 1990, to be a simple test that general health professionals could use to identify delirium rapidly and accurately. The algorithm was devised from the DSM-III-R criteria for the diagnosis of delirium. According to this algorithm, the diagnosis of delirium is based on four features: (i) acute onset and fluctuating course; (ii) inattention; (iii) disorganised thinking and (iv) altered level of consciousness. CAM-positive delirium requires the presence of (i) and (ii) with either (iii) or (iv). The CAM has high sensitivity and specificity in general medical inpatients (0.9). Patients were screened at weekly intervals, using the CAM for a maximum of 4 weeks or until they were discharged. We screened for pre-stroke cognitive impairment, a known risk factor for the development of delirium post-stroke using a shortened 16-question version of the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) [9]. We defined cognitive impairment as an IQCODE score >3. All the assessments were carried out by one assessor–-a fourth year Specialist Registrar in Geriatric Medicine, who had received specific training in the use of the CAM by a psychiatrist who has an interest in delirium research.

Follow-up

Patient outcome measures for 1 month for each patient included inpatient mortality, length of stay in hospital and discharge destination (home or institutionalisation).

Statistics

Data were analysed using SPSS, version 14 (SPSS Inc., Chicago, IL, USA). We firstly used binary logistic regression to test for associations between delirium and predisposing and precipitating risk factors. Odds ratios for risk factors associated with delirium that were significantly associated with delirium according to the binary logistic regression were then entered into multiple logistic regression (MLR) analyses to build a model for determining predictors of delirium. We developed two MLR models, both of which explained similar amounts of variance in the development of delirium (the dependent variable). Only predictors that were significant were kept in the final model. The level of significance was set at 0.05.

Results

Patient characteristics

A total of 110 patients were eligible for the study of whom 82 were recruited. Twenty-eight patients were excluded. Consent or assent was not obtained in 13 cases; 6 patients did not speak English; there was a time delay in recruiting 5 patients and 4 patients had a Glasgow Coma Scale < 8. A total of 51 patients were male; 31 patients were female. The mean age of the group was 66.4 years (range 24–97 years, SD 15.9).

Prevalence of delirium

Within 1 month of the stroke, delirium was detected in 23 patients (28%): 8 women and 15 men. In 21 of these cases, delirium was detected on the initial assessment. No significant difference was found in the prevalence of delirium between male and female stroke patients (Table 1). Patients who developed delirium were older than patients who did not. A total of 34 patients (42%) were African Caribbean. Among African-Caribbean patients, 9 of 34 patients (26.5%) developed delirium versus 14 of 48 (29.2%) of non-African-Caribbean patients. The mean age of the African-Caribbean group was 64.4 years (41–83 years), while the mean age of the non-African-Caribbean group which was mainly Caucasian was 67.8 years (24–97 years). Of the 23 patients with delirium, 16 had four consecutive weekly assessments. At week 4, delirium was present in 9 of the 14 patients in whom delirium was detected at week 1 (64.3%). Only two patients developed delirium after the first assessment.

Table 1

Characteristics of patients with stroke divided into those with and without delirium

Characteristic Total With delirium Without delirium OR 95% CI for OR P-value 
Demographics        
 Male 51 (62%) 15 (29%) 36 (71%) 0.838 0.306 2.281 0.73 
 Female 31 (38%) 8 (26%) 23 (74%)     
 Afro-Caribbean 34 (42%) 9 (26%) 25 (74%) 0.874 0.327 2.338 0.8 
 Age 66.41 75.04 63.05    0.008 
Predisposing factors        
 Poor vision 6 (7%) 5 (83%) 1 (17%) 16.11 1.77 147.04 0.006 
 Poor hearing 8 (10%) 6 (75%) 2 (25%) 10.06 1.86 54.49 0.005 
 IQCODE > 3 21 (28%) 13 (62%) 8 (38%) 7.31 2.4 22.32 0.001 
 Small-vessel disease on brain CT scan 31 (38%) 14 (45%) 17 (55%) 3.84 1.4 10.54 0.011 
 > 4 medications on admission    1.22 1.05 1.42 0.009 
Vascular risk factors        
 History of CVA/TIA 17 (21%) 8 (47%) 9 (53%) 2.96 0.97 9.02 0.05 
 Atrial fibrillation 21 (26%) 9 (43%) 12 (57%) 2.52 0.88 7.2 0.08 
 Total vascular risk factors    1.5 1.06 2.13 0.024 
On admission        
 Admission Barthel    0.74 0.64 0.84 <0.001 
 CRP > 5 mg/l 41 (53%) 19 (46%) 22 (54%) 7.13 2.13 23.79 0.001 
 Unsafe swallow assessment 22 (27%) 18 (82%) 4 (18%) 49.5 11.98 204.46 <0.001 
Stroke        
 Bamford        
 Lacunar 17 (21%) 2 (12%) 15 (88%) 0.28 0.06 1.34 0.09 
 TACI 15 (18%) 11 (73%) 4 (27%) 12.6 3.42 46.42 <0.001 
Endpoints        
 Inpatient mortality 8 (10%) 7 (88%) 1 (13%) 25.38 2.91 221.61 <0.001 
 Length of stay    1.03 1.01 1.05 <0.001 
 Institutionalisation 10 (14%) 7 (70%) 3 (30%) 14 3.05 64.37 <0.001 
Characteristic Total With delirium Without delirium OR 95% CI for OR P-value 
Demographics        
 Male 51 (62%) 15 (29%) 36 (71%) 0.838 0.306 2.281 0.73 
 Female 31 (38%) 8 (26%) 23 (74%)     
 Afro-Caribbean 34 (42%) 9 (26%) 25 (74%) 0.874 0.327 2.338 0.8 
 Age 66.41 75.04 63.05    0.008 
Predisposing factors        
 Poor vision 6 (7%) 5 (83%) 1 (17%) 16.11 1.77 147.04 0.006 
 Poor hearing 8 (10%) 6 (75%) 2 (25%) 10.06 1.86 54.49 0.005 
 IQCODE > 3 21 (28%) 13 (62%) 8 (38%) 7.31 2.4 22.32 0.001 
 Small-vessel disease on brain CT scan 31 (38%) 14 (45%) 17 (55%) 3.84 1.4 10.54 0.011 
 > 4 medications on admission    1.22 1.05 1.42 0.009 
Vascular risk factors        
 History of CVA/TIA 17 (21%) 8 (47%) 9 (53%) 2.96 0.97 9.02 0.05 
 Atrial fibrillation 21 (26%) 9 (43%) 12 (57%) 2.52 0.88 7.2 0.08 
 Total vascular risk factors    1.5 1.06 2.13 0.024 
On admission        
 Admission Barthel    0.74 0.64 0.84 <0.001 
 CRP > 5 mg/l 41 (53%) 19 (46%) 22 (54%) 7.13 2.13 23.79 0.001 
 Unsafe swallow assessment 22 (27%) 18 (82%) 4 (18%) 49.5 11.98 204.46 <0.001 
Stroke        
 Bamford        
 Lacunar 17 (21%) 2 (12%) 15 (88%) 0.28 0.06 1.34 0.09 
 TACI 15 (18%) 11 (73%) 4 (27%) 12.6 3.42 46.42 <0.001 
Endpoints        
 Inpatient mortality 8 (10%) 7 (88%) 1 (13%) 25.38 2.91 221.61 <0.001 
 Length of stay    1.03 1.01 1.05 <0.001 
 Institutionalisation 10 (14%) 7 (70%) 3 (30%) 14 3.05 64.37 <0.001 

Risk factors for delirium

Bivariate analysis demonstrated many significant associations with the development of delirium (Table 1). MLR analysis was performed based on the results of the bivariate regression analysis. Two models were developed which identified independent determinants of developing delirium including unsafe swallow on admission, an admission Barthel score <10, a raised CRP on admission and poor vision pre-stroke (Table 2). Pre-stroke cognitive impairment approached statistical significance as an independent predictor (P = 0.064).

Table 2

Multivariate logistic regression analysis: independent factors associated with delirium post stroke

Characteristic Total With delirium Without delirium OR 95% CI for OR P-value 
Model 1a 
Unsafe swallow 22 (27%) 18 (82%) 4 (18%) 28.4 4.6 174.1 <0.001 
Admission Barthel < 10 35 (43%) 21 (60% 14 (40%) 32.1 3.1 332.8 0.004 
Poor vision 6 (7%) 5 (83%) 1 (17%) 110.8 3.0 4049.9 0.01 
Model 2b 
CRP > 5 mg/l 41 (53%) 19 (46%) 22 (54%) 10.2 1.8 58.7 0.009 
Admission Barthel < 10 35 (43% 21 (60%) 14 (40%) 46.5 5.3 403.8 0.001 
Poor vision 6 (7%) 5 (83%) 1 (17%) 85.2 2.8 2564.5 0.011 
Characteristic Total With delirium Without delirium OR 95% CI for OR P-value 
Model 1a 
Unsafe swallow 22 (27%) 18 (82%) 4 (18%) 28.4 4.6 174.1 <0.001 
Admission Barthel < 10 35 (43%) 21 (60% 14 (40%) 32.1 3.1 332.8 0.004 
Poor vision 6 (7%) 5 (83%) 1 (17%) 110.8 3.0 4049.9 0.01 
Model 2b 
CRP > 5 mg/l 41 (53%) 19 (46%) 22 (54%) 10.2 1.8 58.7 0.009 
Admission Barthel < 10 35 (43% 21 (60%) 14 (40%) 46.5 5.3 403.8 0.001 
Poor vision 6 (7%) 5 (83%) 1 (17%) 85.2 2.8 2564.5 0.011 

aThe Cox and Snell r2 value: 51.8%.

bThe Cox and Snell r2 value: 46.2%.

Outcomes

Delirious patients had a significantly higher mortality, longer length of stay in hospital and higher risk of institutionalisation (Table 3).

Table 3

Outcomes for both delirious and non-delirious patients

 Group with Group without  
Endpoint delirium delirium P-value 
Mortality 30.4% 1.7% <0.001 
Length of stay (LOS) 62.2 days 28.9 days <0.001 
Institutionalisation 43.7% 5.2% <0.001 
 Group with Group without  
Endpoint delirium delirium P-value 
Mortality 30.4% 1.7% <0.001 
Length of stay (LOS) 62.2 days 28.9 days <0.001 
Institutionalisation 43.7% 5.2% <0.001 

Discussion

This study is the first prospective study with 4 weeks of sequential follow-up data of delirium in the acute stroke setting and the first British study to assess delirium post-stroke. We found that delirium is a common problem in the acute stroke setting with a prevalence of 28%, within the range of previous studies. The difference in prevalence rates in these studies may be due to several factors. For example, with regard to age, in the Caeiro et al. study [3], the mean age was 57.3 years, significantly lower than that in the other studies where the mean age was over 70 years. They found a prevalence of 13%. The mean age in our group at 66.4 years is less than that one would expect in a stroke population, and is not entirely accounted for by the large African-Caribbean group. It has been shown that the average age of stroke is 11 years less in African-Caribbean than white patients [10]. Fourteen (17%) of our patients were under the age of 50, and seven of these were African Caribbean. The two previous studies with the highest prevalence figures (42 and 48%) were by Gustafson et al. who performed serial assessments [1, 2]. It is not surprising, in view of the fluctuating nature of delirium, that more cases of delirium will be diagnosed with more frequent monitoring.

We found that 21 of 23 cases of delirium during the first month post-stroke developed within 4 days of the stroke and would, therefore, be picked up by one delirium screening assessment within 4 days of admission. In all cases of delirium in week 1, the acute stroke was felt to be the primary precipitant of delirium. In the two cases that developed subsequently, one case (week 2) was felt to be due to septicaemia and the other case (week 4) was felt to be due to hyponatraemia. This has important implications in that it suggests that if delirium screening is incorporated into the initial stroke assessments within the first few days of admission, most cases of delirium will be identified.

The majority of patients in whom delirium was detected at the first assessment and who were screened for four consecutive weeks were still delirious at week 4. This finding is consistent with other recent studies that have shown that delirium is slow to resolve. McCusker et al. have shown that delirium is still present in 32% of medical inpatients on discharge [11]. Our study is the first study to assess delirious stroke patients for four consecutive weeks. There is a need for further studies to clarify if the duration of delirium in the stroke setting is longer than that in other general medical inpatients with comparable risk factors.

We found that delirium is independently associated with poor vision pre-stroke, TACI stroke, an admission Barthel < 10 and either a raised CRP on admission or unsafe swallow on admission. Poor vision is a well-recognised risk factor for delirium [12]. With regard to stroke subtype, we found that delirium was more likely associated with TACI stroke; however, in the final regression analysis model, we found that a low admission Barthel score was a stronger risk factor for delirium than stroke subtype. It is not surprising that stroke severity is associated with the risk of developing delirium. In addition, severe strokes are more likely to be associated with medical complications, which by themselves could precipitate delirium. The primary precipitant for the onset of delirium may differ from case to case. More research is needed to clarify the contribution of the stroke to the risk of developing delirium compared with the risk of developing delirium post-stroke because of medical complications. As only two of our patients developed delirium beyond 4 days, this suggests that the stroke itself was largely responsible for the delirium instead of medical complications.

Unsafe swallow on admission is also likely to be a marker of stroke severity. We did not find the TOAST classification to be helpful in predicting those patients at risk of developing delirium, although Sheng et al. did find that patients who did have an ichaemic stroke due to a cardioembolic source were more likely to develop delirium [4]. Interestingly, we found that a raised CRP on admission is independently associated with delirium onset. Recently, a raised admission CRP has also been shown to be a predictor of the incidence of delirium in medical inpatients [13]. However, a raised CRP is very common in older patients admitted to hospital. One potential explanation for CRP being high in delirious patients on admission was because these patients might have had larger strokes and therefore were more likely to have developed infective complications. Whether this could have happened by the first assessment on day 4 is unclear. Henon et al. found that metabolic or infectious disorders were independently associated with delirium post-stroke [5].

We have shown in this study that delirium post-stroke is associated with a poor prognosis, increased mortality, longer length of stay in hospital and higher rate of institutionalisation post-discharge. This is similar to delirium in the setting of other medical illnesses [14, 15]. There are few data on the consequences of delirium post-stroke, in particular the long-term sequelae. Only one report has 12-month follow-up data [4]. However, the data that are available indicate similar prognostic associations to those found in other clinical samples. While it seems clear that delirium is a poor prognostic indicator post-stroke, what is less clear is whether this is because of its association with severe hemispheric strokes or whether it is by itself an independent marker of poor outcome post-stoke [16]. There is a need to clarify this–-it may be that the prognosis with delirium post-stroke may be worse than that in medical patients in general as both delirium and stroke are acute brain syndromes.

As there is no definitive treatment for delirium, despite its high morbidity and mortality, the prevention of delirium should be a priority in all hospital wards, including stroke units. To date, there have been no studies that have evaluated the management of delirium post-stroke. While the incidence of delirium can be reduced in non-stroke patients [17], significant prevention may not be possible in stroke, as we have shown that most cases of delirium develop early, probably at onset as a result of the stroke itself. What may be possible is to reduce the length of the delirium by some of the interventions shown to be useful in non-stroke patients.

There are several limitations of our study. We might have underestimated the incidence of delirium as we excluded six non-English speakers and four patients with a GCS score <8. We assessed patients within 4 days of admission. Delirium might have resolved in some stroke patients by 4 days. Another potential limitation is that patients were not assessed at the same time of the day over the 4-week period—this would not be feasible in the context of a busy Stroke Unit. It is also possible that some stroke patients with delirium might have been delirious before their stroke. Finally, the study did not have sufficient power to investigate sub-threshold delirium.

Conclusion

We found that delirium is a common complication in the acute stroke setting with grave prognostic implications. Most patients who develop delirium post-stroke will be delirious on admission; therefore, screening for delirium on admission in all stroke patients will diagnose the vast majority of cases. Most patients who develop delirium remain delirious for at least a month. The onset of delirium is associated with a much higher mortality, a longer length of stay and increased risk of institutionalisation. We have identified the best predictors of delirium development post-stroke, the knowledge of which may improve the future recognition of delirium in the stroke setting.

Key points

  • Delirium is a common complication in the acute stroke setting.

  • Most cases occur early and are slow to resolve.

  • Delirium post-stroke is associated with a poor prognosis.

Conflicts of interest

None.

Funding

No external sources of funding.

Ethical approval

Ethics approval for the study was obtained from the Central Office for Research Ethics Committees (COREC) prior to the commencement of the study. Informed consent was taken from all patients. In patients unable to give consent, a next-of-kin gave informed assent.

References

1
Gustafson
Y
Olsson
T
Asplund
K
Hagg
E
Acute confusional state (delirium) soon after stroke is associated with hypercortisolism
Cerebrovasc Dis
 , 
1993
, vol. 
3
 (pg. 
33
-
8
)
2
Gustafson
Y
Olsson
T
Erikkson
S
Asplund
K
Bucht
G
Acute confusional states (delirium) in stroke patients
Cerebrovasc Dis
 , 
1991
, vol. 
1
 (pg. 
257
-
64
)
3
Caeiro
L
Ferro
J
Albuquerque
R
Figueira
ML
Delirium in the first days of acute stroke
J Neurol
 , 
2004
, vol. 
251
 (pg. 
171
-
8
)
4
Sheng
AZ
Shen
Q
Cordato
D
Zhang
YY
Kam Yin Chan
D
Delirium within three days of stroke in a cohort of elderly patients
J Am Geriatr Soc
 , 
2006
, vol. 
54
 (pg. 
1192
-
8
)
5
Henon
H
Lebert
F
Durieu
I
, et al.  . 
Confusional state in stroke. Relation to preexisting dementia, patient characteristics and outcome
Stroke
 , 
1999
, vol. 
30
 (pg. 
773
-
9
)
6
Bamford
J
Sandercock
P
Dennis
M
Burn
J
Warlow
C
Classification and natural history of clinically identifiable subtypes of cerebral infarction
Lancet
 , 
1991
, vol. 
337
 (pg. 
1521
-
6
)
7
Adams
HP
Jr
Bendixen
BP
Kappelle
LJ
, et al.  . 
Classification of subtype of acute ischaemic stroke. Definitions for use in a multicentre clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment
Stroke
 , 
1993
, vol. 
24
 (pg. 
35
-
41
)
8
Inouye
SK
Dyck
CH
Alessi
CA
Balkin
S
Siegal
AP
Horwitz
RI
Clarifying confusion: the confusion assessment method. A new method for detection of delirium
Ann Intern Med
 , 
1990
, vol. 
113
 (pg. 
941
-
8
)
9
Jorm
AF
A short form of the Informant Questionnaire on Cognitive Impairment in the Elderly (IQCODE): development and cross-validation
Psychol Med
 , 
1994
, vol. 
24
 (pg. 
145
-
53
)
10
Wolfe
CDA
Rudd
AG
Howard
R
, et al.  . 
Incidence and case fatality rates of stroke subtypes in a multiethnic population: the South London Stroke Register
J Neurol Neurosurg Psychiatry
 , 
2002
, vol. 
72
 (pg. 
211
-
6
)
11
McCusker
J
Cole
M
Dendukuri
N
Belzile
E
Primeau
F
Delirium in older medical inpatients and subsequent cognitive and functional status: a prospective study
Can Med Assoc J
 , 
2001
, vol. 
165
 (pg. 
575
-
83
)
12
The prevention, diagnosis and management of delirium in older people
2006
June
 
National Guidelines, Royal College of Physicians
13
Macdonald
A
Adamis
A
Treloar
A
Martin
F
C-reactive protein levels predict the incidence of delirium and recovery from it
Age Ageing
 , 
2007
, vol. 
36
 (pg. 
222
-
5
)
14
Siddiqi
N
House
AO
Holmes
JD
Occurrence and outcome of delirium in medical in-patients: a systematic literature review
Age Ageing
 , 
2006
, vol. 
35
 (pg. 
350
-
64
)
15
Young
J
Inouye
S
Delirium in older people
Br Med J
 , 
2007
, vol. 
334
 (pg. 
842
-
6
)
16
Mc Manus
J
Pathansali
R
Stewart
R
Macdonald
A
Jackson
SHJ
Delirium, post-stroke
Age Ageing
 , 
2007
, vol. 
36
 (pg. 
613
-
8
)
17
Inouye
SK
Bogardus
ST
Charpentier
PA
, et al.  . 
A multicomponent intervention to prevent delirium in hospitalized older patients
N Engl J Med
 , 
1999
, vol. 
340
 (pg. 
669
-
76
)

Comments

0 Comments