Dementia is strongly associated with 90-day mortality in lobar cerebral amyloid angiopathy related intra-cerebral haemorrhage

Journal of the Neurological Sciences 322 (2012) 161–165

Contents lists available at SciVerse ScienceDirect

Journal of the Neurological Sciences journal homepage: www.elsevier.com/locate/jns

Dementia is strongly associated with 90-day mortality in lobar cerebral amyloid angiopathy related intra-cerebral haemorrhage Elena I. Jamieson a,⁎, David Newman b, Anthony K. Metcalf b, Magdi F. Naguib b, Janak Saada b, John F. Potter b, c, Phyo Kyaw Myint b, c a b c

Ipswich Hospital, Heath Road, Ipswich, IP4 5PD Suffolk, UK Norfolk and Norwich University Hospital, Norwich, NR4 7UY Norfolk, UK Norwich Medical School, University of East Anglia, Norwich Research Park, Norwich, NR4 7TJ Norfolk, UK

a r t i c l e

i n f o

Article history: Received 30 January 2012 Received in revised form 10 June 2012 Accepted 19 July 2012 Available online 9 August 2012 Keywords: Lobar intracerebral haemorrhage Non-lobar intracerebral haemorrhage Cerebral amyloid angiopathy Dementia Cognitive impairment Mortality

a b s t r a c t Background: While evidence suggests that lobar intracerebral haemorrhage (ICH) is linked with dementia and cognitive impairment, the association between cognition and mortality risk from ICH is unclear. Aims: To examine the association between dementia or cognitive impairment and short- and medium-term mortality post ICH. Methods: Patients with primary ICH were classified into lobar and non-lobar ICH using radiological criteria. Patients' characteristics and radiological measures were collected at the baseline along with history of dementia and cognitive impairment. Mortality risks at 7, 30, 60, and 90 days were assessed using multiple logistic regression adjusting for potential confounders identified as significant associates in univariate models. Results: A total of 136 patients (males 50%, mean age 77 years, SD 10) were included in this study. Out of 53 (39%) patients with lobar ICH 47 (89%) were classified as having possible and 6 (11%) as probable cerebral amyloid angiopathy (CAA). In lobar ICH the prevalence of history of dementia or cognitive impairment, confusion at presentation, previous ICH, multiple haemorrhages, and initial haematoma volume were significantly higher (p b 0.05). In lobar ICH the significant mortality predictors (p b 0.05) were history of dementia or cognitive impairment (90 days), prior antiplatelet use (60 and 90 days), initial haematoma volume (60 days), male sex (30 and 60 days), age (30, 60, 90 days), and low Glasgow Coma Scale (GCS) (7 and 30 days). In non-lobar ICH prior use of anticoagulation, initial haematoma volume, low GCS and age were significant mortality predictors (p b 0.05). Conclusion: A history of dementia or cognitive impairment is more common in lobar CAA-related ICH and it is a medium-term mortality predictor in lobar ICH but not in deep non-lobar ICH. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Spontaneous intracerebral haemorrhage (ICH) accounts for 10 to 20% of all strokes with a higher incidence in the older population [3,16]. It is one of the most devastating cerebro-vascular diseases with 30-day mortality reported up to 55% in the elderly populations [6,19]. Spontaneous ICH can be categorised into lobar (cortico-subcortical) and non-lobar (deep white matter) haemorrhage using radiological findings [4]. The underlying pathophysiological processes in lobar and non-lobar haemorrhage differ considerably. Non-lobar ICH is thought to be strongly associated with hypertension and lobar ICH is associated with cerebral amyloid angiopathy (CAA) [19,11]. The prevalence of CAA on postmortem examination or brain biopsy in lobar ICH was reported as high as 74% in a clinico-pathological correlation study [12]. ⁎ Corresponding author at: Ipswich Hospital, Heath Road, Ipswich, IP4 5PD Suffolk, UK. Tel.: +44 7734466890. E-mail address: [email protected] (E.I. Jamieson). 0022-510X/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jns.2012.07.047

When compared with ischaemic stroke, there is a relative paucity of research on the link between ICH and dementia. There is evidence suggesting that lobar ICH is a risk factor for subsequent cognitive impairment, and that pre-existing dementia appears to be higher in lobar haemorrhage [5,9]. However, whether cognitive impairment and dementia is associated with a higher risk of lobar ICH compared to non-lobar haemorrhage and whether it has any impact on the outcome of ICH are less well examined. This study aims to investigate the role of dementia and cognitive impairment prior to ICH as a risk factor and mortality predictor in lobar and non-lobar ICH.

2. Methods The project was approved by the Cambridgeshire Research Ethics Committee. The ICH cases were retrospectively identified using the prospective institutional database, the Stroke Register — a part of the European Basic Stroke Register. Part of the data for the current study

162

E.I. Jamieson et al. / Journal of the Neurological Sciences 322 (2012) 161–165

(e.g. age, sex, pre-stroke disability) were collected prospectively by a specialist stroke nurse and verified by a stroke physician and the remaining data were collected retrospectively. In brief, patients who suffered spontaneous ICH and who were admitted to a large university hospital in the east of England from January 2003 to March 2004 were included into the study and followed for 90 days. The definition of ICH was adopted from the Classification of Cerebral-vascular Disease III (1990). The inclusion criteria were evidence of ICH on CT or MR images. Depending on the radiological appearances, CT and MR brain images were initially categorised into lobar and non-lobar distribution [4,19] by a researcher (E.I.J.), and independently reviewed by a stroke physician (A.K.M.) and neuroradiologists (J.S. and D.N.). The time between the onset of the symptoms and brain scan was taken into consideration in interpreting the neuroradiological findings. Investigators were blinded to any information on patient's characteristics when reporting the images. Lobar ICH was defined as occurring in temporal, parietal, occipital or frontal lobes and involving predominantly cortex and underlying white matter. Multiple haemorrhages were defined as two or more separate haematomata. Non-lobar ICH was defined as basal ganglionic, thalamic, cerebellar and brain stem haemorrhages and ICH involving peri-ventricular white matter, or internal capsule. The Boston criteria were used to categorise the probability of CAA in the lobar ICH group, taking into account the patient's age, number of haemorrhages and location [12,19]. Without histological samples, patients with lobar ICH could be classified as having possible or probable CAA. Patients with identified secondary causes such as coagulopathy, INR >3.0, cerebral metastases, vasculitis, subarachnoid haemorrhage, subdural haemorrhage, traumatic cerebral haemorrhage, haemorrhagic transformation of cerebral infarct and haemorrhage due to vascular malformations were excluded from the study. Images were discussed in order to reach mutual agreement on the category of the ICH. The primary outcome measure was patient death (all cause mortality) and mortality data were used to establish significant predictors of short-term mortality (7 and 30 days) and medium-term (60 and 90 days) mortality. Dates of patients' death were obtained from clinical notes and from the hospital's Patient Administrative System (PAS), which collects reports of death from the community as well as inhospital deaths. In all cases where patients were reported alive by the PAS, GP surgeries were contacted to ascertain the patient's status at follow‐up. Baseline characteristics including age, sex, symptomatology at the onset (e.g. confusion) and premorbid medical history were collected. Information on history of vascular disease, Alzheimer's disease, mixed dementia or cognitive impairment was obtained from the medical records and coded as dementia/cognitive impairment. Antiplatelet use was recorded as “yes” when patient was taking any dose of aspirin, clopidogrel, dipyridamole or any combination of these medications. The anticoagulant use category included warfarin, low and high doses of low molecular weight and unfractionated heparin. Other factors such as pre-stroke modified Rankin score (MRS), blood pressure (BP), and Glasgow Coma Scale (GCS) were also recorded. For haematoma volume calculation, the ABC/2 method was used [13]. This involved calculating the volume of the haemorrhage using 3-dimensional measurements and slice thickness. A is the greatest haemorrhage diameter by CT, B is the diameter 90° to A, and C is the number of slices that the haemorrhage involves multiplied by CT slice thickness. The extent of cerebral vascular disease was assessed using the Fazekas scale to provide an overall impression of the level of small vessel ischemic disease [7]. Fazekas 0 represents none or a single white matter lesion; Fazekas 1 represents multiple punctuate lesions; Fazekas 2 represents beginning confluency of lesions and Fazekas 3 represents large confluent lesions.

3. Statistical analysis Characteristics and outcome comparison between lobar and non-lobar ICH were performed with the unpaired t-test (for continuous variables) and its non-parametric alternative (for non‐normally distributed data), and Chi-squared test was used for categorical data. For the primary research question the study population was dichotomised into lobar and non-lobar ICH groups. Binary logistic regression was applied to examine the odds ratios (ORs) for 7-day, 30-day, 60-day, and 90-day mortality after the ICH in both groups. Pre-morbid characteristics and clinical variables were tested in unadjusted univariate analysis first and carried forward to the multivariate model if an association was significant at p b 0.1 level at any time point. We had 100% follow‐up (no censoring) and used all cause mortality (no competing risks). Univariate and multivariate logistic regression models were also performed to examine the mortality outcome at the same time points comparing lobar and non-lobar ICH (using non-lobar haemorrhage as the reference category). 4. Results A total of 136 patients (males 50%; mean age 77 years, SD 10; range 33–95 years) were included in this study. Table 1 shows the characteristics and outcome of all participants and provides the comparison between lobar and non-lobar ICH groups. Out of 53 patients with lobar ICH (39%), 47 (89%) had possible and 6 (11%) had probable CAA. Initial haematoma volume was significantly larger in lobar ICH group (p = 0.005) and there were no significant differences in severity of small vessel disease between patients with lobar and non-lobar ICH (p = 0.1). History of dementia or cognitive impairment, previous ICH, confusion at presentation, and multiple haemorrhages were significantly higher in lobar ICH (p b 0.05). Headache, collapse, higher systolic and diastolic BP at presentation were significantly commoner (p b 0.05) in non-lobar ICH (Table 1). Overall, ICH had high 90-day mortality (49%) but it was much higher (81%) in ICH patients with a history of dementia or cognitive impairment (Table 2). There were no differences in mortality between lobar and non-lobar ICH groups at any follow‐up time point by both unadjusted and adjusted analyses controlling for age, sex, initial haematoma volume, dementia or cognitive impairment, diabetes, hypertension and GCS. In lobar ICH, variables independently associated with increased mortality in univariate analyses were a history of dementia or cognitive impairment, age, sex, prior antiplatelet use, initial haematoma volume and low GCS. In non-lobar ICH these factors were age, prior use of anticoagulants, systolic and diastolic BP, initial haematoma volume and low GCS. The likelihood of mortality at different time points for lobar and non-lobar ICH using a multivariate logistic regression model is presented in Table 3. In lobar ICH, history of dementia or cognitive impairment significantly predicted 90-day mortality (OR 33.3, 95% CI 2.16–394.01, p = 0.01) and other significant mortality predictors (p b 0.05) were prior antiplatelet agent use (60 and 90 days), male gender (30 and 60 days) and low GCS (7 and 30 days). In multivariate models for non-lobar ICH, significant mortality predictors (pb 0.05) were low GCS (at all time points), prior anticoagulation (60 and 90 days) and higher diastolic BP (30, 60 and 90 days). In both lobar and non-lobar ICH, smaller initial haematoma volume and younger age were associated with lower mortality ORs. 5. Discussion Although there is a clear link between vascular dementia and cerebro-vascular disease, there is a lack of research on the role of dementia in ICH. In our study the prevalence of pre-existing dementia in lobar ICH was 19%. A recent study reported 23% premorbid dementia prevalence rate in lobar ICH and factors associated with dementia

E.I. Jamieson et al. / Journal of the Neurological Sciences 322 (2012) 161–165

163

Table 1 Baseline clinical characteristics and outcome of all patients with spontaneous ICH and comparison between lobar and non-lobar ICH groups.

Number of patients Sex (male = M, female = F) Age (years) Boston criteria Possible Probable Initial haematoma volume (mls) Degree of small vessel disease (Fazekas scale) 0 1 2 3 History of diabetes Previous hypertension Use of antiplatelets Type Aspirin Dipyridamole Clopidogrel Use of anticoagulants Previous TIA Previous stroke Previous haemorrhagic stroke Confusion at presentation Dementia/cognitive impairment Pre-morbid MRS 0 1 2 3 4 5 Systolic BP (mm Hg) Diastolic BP (mm Hg) GCS LOC/collapse Headache Multiple haemorrhage Discharge MRS 0 1 2 3 4 5 Inhospital death

All ICH

Lobar ICH

Non-lobar ICH

p Valuea

Mean (SD), n (%) 136 M 68, F 68 77 (10)

53 (39) M 27, F 26 77 (8)

83 (61) M 41, F 42 76 (11)

0.59 0.61

N/A

N/A

25.9 (25.8)

47 (81) 6 (11) 33.8 (25.9)

21.1 (24.7)

0.005c 0.1b

20 66 35 15 14 73 45

(15) (48) (26) (11) (10) (54) (33)

12 (23) 24 (45) 10 (19) 7 (13) 5 (9) 30 (57) 26 (58)

8 (10) 42 (51) 25 (30) 8 (10) 9 (10) 43 (52) 19 (42)

40 (30) 8 (6) 5 (4) 22 (16) 15 (11) 24 (17) 5 (4) 37 (27) 16 (12)

16 (30) 3 (6) 4 (8) 10 (19) 9 (17) 11 (13) 4 (8) 27 (51) 10(19)

24 (29) 5 (6) 1 (1) 12 (15) 6 (7) 13 (24) 1 (1.2) 10 (12) 6 (7)

86 (63) 23 (17) 16 (12) 7 (5) 4 (3) 0 (0) 179 (31) 93 (21) 11.5 (3.9) 79 (58) 45 (33) 9 (7)

37(70) 6 (11) 6 (11) 2 (4) 2 (2) 0 (0) 162 (28) 87 (22) 12 (3.3) 23 (43) 11 (20) 8 (15)

49(59) 17(20) 10 (12) 5 (6) 2 (4) 0 (0) 179 (31) 97 (20) 11 (4.4) 56 (68) 34 (41) 1 (1)

1 (0.7) 3 (2) 8 (6) 15 (11) 25 (18) 20(15) 64 (47)

1 (2) 2 (4) 5 (9) 3 (6) 10 (19) 7 (13) 25 (47)

0 (0) 1 (1.2) 3 (4) 12 (15) 15 (18) 13 (16) 39 (47)

N/A

0.79 0.58 0.58 0.72 0.93 0.06 0.49 0.08 0.09 0.05 b0.0001 0.04 0.60b

0.001 0.006 0.2 0.006 0.02 0.01 0.38b

0.61

Modified Rankin Scale (MRS) 0: No symptoms at all 1: No significant disability despite symptoms; able to carry out all usual duties and activities 2: Slight disability; unable to carry out all previous activities but able to look after own affairs without assistance 3: Moderate disability; requiring some help, but able to walk without assistance 4: Moderately severe disability; unable to walk without assistance and unable to attend to own bodily needs without assistance 5: Severe disability; bedridden, incontinent and requiring constant nursing care and attention 6: Dead Fazekas scale 0: 1: 2: 3:

None or a single white matter lesion Multiple punctuate lesions Beginning of confluency of lesions Large confluent lesions a Data presented are mean (SD) for continuous and or number (%) for categorical data. Independent sample t test was used to compare means and Chi square test was used to compare categorical data. Data presented are number (%) for categorical data and mean (SD) for continuous data. b Overall p value. c Significant associations (p ≤ 0.05) are highlighted in bold.

were older age (OR 1.09, 95% CI 1.02–1.15) and increasing cortical atrophy (OR 3.35, 95% CI 1.40–7.96) [5]. We found that association with premorbid dementia or cognitive impairment exists in lobar ICH but not in non-lobar ICH. While the mechanism of this association remains unclear, this finding needs to be replicated, perhaps in larger prospective cohorts. A higher prevalence of dementia or cognitive impairment (19% vs. 7%, p = 0.04) in

the lobar ICH group when compared with the non-lobar ICH group and direct dementia-related factors contributing death could be one plausible explanation. Generally, acutely ill patients with dementia or cognitive impairment have a higher inpatient mortality compared to patients without [18] and mortality increases with dementia severity [1]. Interestingly, only mid- but not short term mortality showed association with dementia or cognitive impairment in lobar ICH. There is a

164

E.I. Jamieson et al. / Journal of the Neurological Sciences 322 (2012) 161–165

Table 2 Mortality in all patients with ICH; and lobar and non-lobar ICH groups at different follow up time points and likelihood of mortality in patients with lobar ICH compare to patients with non-lobar ICHa. Follow‐up time

0–7 days 8–30 days 31–60 days 61–90 days a b

All ICH, n= 136

Dementia or cognitive impairment, n = 16

Lobar ICH, n = 53

Non-lobar ICH, n = 83

Odds ratios of mortality in lobar compared to non-lobar haemorrhage

n at risk of death

n of deaths (%)

n at risk of death

n of deaths (%)

n at risk of death

n of deaths (%)

n at risk of death

n of deaths (%)

Unadjusted

104 79 71 69

32 57 65 67

15 9 5 3

1 (6) 7 (44) 11 (69) 13 (81)

43 33 28 26

10 20 25 27

61 46 43 43

22 37 40 40

0.66 0.75 0.97 1.12

(23.5) (42) (47) (49)

(19) (38) (47) (51)

(27) (45) (48) (48)

Adjusted

b

OR (95% CI), p value (0.27–1.49); (0.37–1.54); (0.48–1.91); (0.56–2.23);

0.31 0.43 0.91 0.75

1.19 0.92 1.20 1.52

(0.39–3.57); 0.76 (0.37–2.27), 0.85 (0.48–3.02); 0.69 (0.59–3.85); 0.38

Non lobar ICH is reference category. Adjusted for age, gender, previous history of diabetes, hypertension, dementia or cognitive impairment, initial haematoma volume and GCS on admission.

lack of research into short- and medium-term mortality in ICH patients with dementia or cognitive impairment and further studies should therefore be powered to examine the specific cause of mortality. In our study, most ICH deaths occurred within the first 60 days with a high 90-day mortality (49%) and this is in agreement with previous reports [8,19]. We found ICH patients with a previous history of dementia or cognitive impairment had a very high 90-day mortality rate (81%). This is an important finding in the view of the rising prevalence of dementia in an ageing United Kingdom population. Antiplatelet and anticoagulant agents are widely used for primary and secondary prevention of cardio-vascular diseases and the prevalence of antiplatelet use prior to ICH is reported to be 20–60% [20]. In our study the antiplatelet/anticoagulant use was 46% and therefore our study population appears to be a representative sample. A recent meta-analysis showed that patients on antiplatelet drugs who suffered ICH had an increased mortality in multivariable-adjusted pooled analyses (OR 1.27, 95% CI 1.10–1.47) [20]. Other studies confirmed a higher mortality from ICH up to one year post stroke in patients with prior antiplatelet/anticoagulant use [15] and independently, an increased risk of ICH recurrence in patients using aspirin after lobar ICH [2]. In our study antiplatelet agent use was a significant medium-term mortality predictor for lobar ICH but not for non-lobar ICH. This would fit in with the theory that CAA is the underlying pathology in the majority of lobar ICHs and that patients with CAA-related ICH who have fragile small cortical-subcortical vessels loaded with amyloid may be at higher risk of larger ICH and subsequently higher mortality when they receive antiplatelet agents. A low level of consciousness, older age, hypertension, and a larger initial haematoma volume are reported to be markers of poor prognosis

[10,14,21]. Our study confirmed that low GCS consistently predicts early mortality (7 and 30 days) in both types of ICH and a greater association was observed for non-lobar ICH. These findings correlate with existing literature [10]. Although the haematoma volume was significantly larger in the lobar ICH group, it was much more strongly associated with mortality at all follow‐up points in deep non-lobar ICH in our study. It may be that the central deep bleeds seen in non-lobar ICH are most likely to cause a low GCS in non-lobar haemorrhage and larger haematoma volume is associated with higher mortality [10]. A younger age in our study was associated with lower mortality ORs in lobar and non-lobar ICH, in agreement with previous reports [10,14] showing that the incidence of lobar haemorrhage was age dependent: 2.3% for patients younger than 75 years old and 12.1% for over 85 years old [10]. Our study does have some limitations. This is a relatively small study to draw definite conclusions on patients' characteristics and factors predicting mortality; and the fact that some of the ORs did not reach statistical significance could be due to a small sample size. We used radiological criteria to assess probability of CAA in lobar ICH and did not apply “gold standard” tests as autopsy or brain biopsy to diagnose definite CAA. Another weakness of this study is the retrospective assessment of the patient's pre-morbid characteristics. Although it is unlikely that objective measures (e.g. systolic BP) were influenced by the retrospective nature of data collection, a prior history of dementia or cognitive impairment is most likely to be underreported. It is well known that in the UK dementia and cognitive impairment are not formally diagnosed in 40–64% of those affected [17]. There may be other unknown residual confounding effects such as microbleeds, secondary hydrocephalus, heamatoma expansion,

Table 3 Multivariate logistic regression model showing odds ratios of mortality at different time points (7, 30, 60, and 90 days) for lobar and non-lobar ICH (only significant predictors at univariate models for any of these time points were included in the multivariate model). Mortality: odds ratio (95% CI); p value 7 days Male sex Younger age Dementia or cognitive impairment Antiplatelet use Low GCS Haematoma volume

3.29 1.00 0.31 1.08 1.42 0.43

(0.47–22.90); 0.23 (0.89–1.13); 0.98 (0.02–4.19); 0.37 (0.14–8.44); 0.94 (1.05–1.92); 0.02 (0.09–1.98); 0.28

Younger age Antiplatelet use Anticoagulation Low GCS SBP DBP Haematoma volume

1.00 2.85 7.43 1.67 0.97 0.98 0.24

(0.93–1.08); 0.99 (0.32–25.28); 0.35 (0.53–104.11); 0.14 (1.25–2.21); 0.001 (0.93–1.01); 0.09 (0.90–1.05); 0.48 (0.06–0.94); 0.04

30 days Lobar ICH group 5.19 (1.02–26.36); 0.05 0.89 (0.79–0.99); 0.05 0.83 (0.14–5.15); 0.83 2.55 (0.46–14.20); 0.28 1.30 (1.00–1.68); 0.05 0.55 (0.17–1.76); 0.32 Non-lobar ICH group 0.83 (0.73–0.94); 0.003 4.28 (0.41–44.91); 0.23 37.57 (2.05–689.95); 0.02 2.75 (1.58–4.78); b0.0001 0.96 (0.92–1.00); 0.06 1.11 (1.01–1.22); 0.02 0.03 (0.003–0.25); 0.002

GCS = Glasgow Coma Scale; SBP = systolic blood pressure; DBP = diastolic blood pressure. Significant associations (p ≤0.05) are highlighted in bold.

60 days 8.95 0.86 8.17 8.37 1.27 0.23

0.81 1.15 31.63 2.94 0.98 1.08 0.02

(1.17–68.39); 0.03 (0.75–0.99); 0.05 (0.83–79.89); 0.07 (1.02–68.55); 0.05 (0.94–1.71); 0.12 (0.06–0.93); 0.04

(0.70–0.93); 0.002 (0.14–9.39); 0.89 (1.75–571.18); 0.02 (1.66–5.22); b0.0001 (0.94–1.02); 0.31 (1.00–1.17); 0.05 (0.002–0.19); 0.001

90 days 5.36 0.87 33.3 7.69 1.21 0.30

0.81 1.15 31.63 2.94 0.98 1.08 0.02

(0.77–37.63); 0.09 (0.76–0.99); 0.04 (2.16–394.01); 0.01 (0.95–62.26); 0.06 (0.89–1.62); 0.21 (0.08–1.12); 0.07

(0.70–0.93); 0.002 (0.14–9.39); 0.89 (1.75–571.18); 0.02 (1.66–5.22); b0.0001 (0.94–1.02); 0.31 (1.00–1.17); 0.05 (0.002–0.19); 0.001

E.I. Jamieson et al. / Journal of the Neurological Sciences 322 (2012) 161–165

intraventricular spread of the bleeding, and midline shift displacement volume which have not been assessed but are known to be markers of poor prognosis after spontaneous ICH. To our knowledge this is the first study showing that dementia or cognitive impairment is an important mortality predictor in lobar CAA-related ICH. Dementia and CAA are increasingly common in the older population and further studies are needed to investigate the specific aspects of these diseases that are associated with increased mortality in order to improve the care of this group of patients. Conflict of interest statement All authors declare that there are no conflicts of interests in connection with this study. Acknowledgements We would like to express our thanks to the Department of Medicine for the Elderly, Norfolk and Norwich University Hospital for the administrative support in conducting this study. Authors' contribution E.I.J. designed the study with input from A.K.M., D.N., M.F.N., J.S. and P.K.M. E.I.J., D.N., A.K.M. and J.S. contributed in data collection. E.I.J. analysed the data. E.I.J. and P.K.M. drafted the paper and J.F.P. critically reviewed the manuscript. All authors contributed in writing of the paper. References [1] Andersen K, Lolk A, Martinussen T, Kragh-Sørensen P. Very mild to severe dementia and mortality: a 14-year follow-up: the Odense study. Dement Geriatr Cogn Disord 2010;29:61-7. [2] Biffi A, Halpin A, Towfighi A, Gilson A, Busl K, Rost N, et al. Aspirin and recurrent intracerebral hemorrhage in cerebral amyloid angiopathy. Neurology 2010;75(8): 693-8 [24]. [3] Brainin M, Bornstein N, Boysen G, Demaerin V. Acute neurological stroke care in Europe: results of the European Stroke Care Inventory. Eur J Neurol 2000;7:5–10.

165

[4] Chao CP, Kotsenas AL, Broderick DF. Cerebral amyloid angiopathy: CT and MR imaging findings. Radiographic 2006;26(5):1517-31. [5] Cordonnier C, Leys D, Dumont F, Deramecourt V, Bordet R, Pasquier F, et al. What are the causes of pre-existing dementia in patients with intracerebral haemorrhages? Brain 2010;133(11):3281-9. [6] Dennis MS. Outcome after brain haemorrhage. Cerebrovasc Dis 2003;16(Suppl. 1): 9–13. [7] Fazekas F, Barkhof F, Wahlund LO, Pantoni L, Erkinjuntti T, Scheltens P, et al. CT and MRI rating of white matter lesions. Cerebrovasc Dis 2002;13(Suppl. 2):31-6. [8] Flaherty ML, Haverbusch M, Sekar P. Long term mortality after intracerebral haemorrhage. Neurology 2006;66:1182-6. [9] Greenberg SM, Eng JA, Ning M, Smith EE, Rosand J. Hemorrhage burden predicts recurrent intracerebral hemorrhage after lobar hemorrhage. Stroke 2004;35(6): 1415-20. [10] Hu X, Zhang JH, Qin X. Risk factors of early death in patients with hypertensive intracerebral hemorrhage during hospitalization. Acta Neurochir Suppl 2011;111: 387-91. [11] Jackson CA, Sudlow CL. Is hypertension a more frequent risk factor for deep than for lobar supratentorial intracerebral haemorrhage? J Neurol Neurosurg Psychiatry 2006;77(11):1244-52. [12] Knudsen KA, Rosand J, Karluk D, Greenberg SM. Clinical diagnosis of cerebral amyloid angiopathy: validation of the Boston criteria. Neurology 2001;56(4):537-9. [13] Kothari RU, Brott T, Broderick JP, Barsan WG, Sauerbeck LR, Zuccarello M, et al. The ABCs of measuring intracerebral hemorrhage volumes. Stroke 1996;27:1304-5. [14] Kuramatsu JB, Sauer R, Mauer C, Lucking H, Kloska S, Kiphuth IC, et al. Correlation of age and haematoma volume in patients with spontaneous lobar intra-cerebral haemorrhage. J Neurol Neurosurg Psychiatry 2011;82:144-9. [15] Kwok CS, Skinner J, Metcalf AK, Potter JF, Myint PK. Prior antiplatelet or anticoagulant therapy and mortality in stroke. Heart 2012;98(9):712-7. [16] Lovelock CE, Molyneux AJ, Rothwell PM. Change in incidence and aetiology of intracerebral haemorrhage in Oxfordshire, UK, between 1981 and 2006: a populationbased study. Lancet Neurol 2007;6:487-93. [17] Russ TC, Shenkin SD, Reynish E, Ryan T, Anderson D, MacLullich AMJ. Dementia in acute hospital inpatients: the role of geriatrician. Age and Ageing 2012;41(3): 282-4. [18] Sampson EL, Blanchard MR, Jones L, Tookman A, King M. Dementia in the acute hospital: prospective cohort study of prevalence and mortality. Br J Psychiatry 2009;195:61-6. [19] Smith EE, Greenberg SM. Clinical diagnosis of cerebral amyloid angiopathy: validation of the Boston criteria. Current Atherosclerosis Reports 2003;4:260-6. [20] Thompson BB, Béjot Y, Caso V, Castillo J, Christensen H, Flaherty ML, et al. Prior antiplatelet therapy and outcome following intracerebral hemorrhage: a systematic review. Neurology 2010;12(75):1333-42 [15]. [21] Zia E, Hedblad B, Pessah-Rasmussen H, Berglund G, Janzon L, Engström G. Blood pressure in relation to the incidence of cerebral infarction and intracerebral hemorrhage. Hypertensive hemorrhage: debated nomenclature is still relevant. Stroke 2007;38(10):2681-5.