Small vessel disease burden and risk of recurrent cerebrovascular events in patients with lacunar stroke and intracerebral haemorrhage attributable to deep perforator arteriolopathy

Abstract

Introduction:

Deep perforator arteriolopathy (DPA) causes intracerebral haemorrhage (ICH) and lacunar strokes (LS). We compare patient characteristics, MRI findings and clinical outcomes among patients with deep ICH and LS.

Patients and methods:

We included patients with MRI-confirmed LS or ICH in the basal ganglia, thalamus, internal capsule or brainstem from the Bernese Stroke Registry. We assessed MRI small vessel disease (SVD) markers, SVD burden score, modified Rankin Scale (mRS) and ischaemic stroke or ICH at 3 months.

Results:

We included 716 patients, 117 patients (16.3%) with deep ICH (mean age (SD) 65.1 (±15.2) years, 37.1% female) and 599 patients (83.7%) with LS (mean age (SD) 69.7 (±13.6) years, 39.9% female). Compared to LS, deep ICH was associated with a higher SVD burden score (median (IQR) 2 (1–2) vs 1 (0–2)), aOR_shift 3.19, 95%CI 2.15–4.75). Deep ICH patients had more often cerebral microbleeds (deep ICH: 71.6% vs LS: 29.2%, p < 0.001, median count (IQR) 4(2–12) vs 2(1–6)) and a higher prevalence of lacunes (deep ICH: 60.5% vs LS: 27.4% p < 0.001). At 3 months, deep ICH was associated with higher mRS (aOR_shift 2.16, 95%CI 1.21–3.87). Occurrence of ischaemic stroke was numerically but not significantly higher in deep ICH (4.3% vs 2.9%; p = 0.51). One patient (1.1%) with ICH but none with LS suffered ICH recurrence.

Discussion/Conclusion:

DPA manifesting as ICH is associated with more severe MRI SVD burden and worse outcome compared to LS. The short-term risks of subsequent ischaemic stroke and recurrent ICH are similar in ICH and LS patients. This implies potential consequences for future secondary prevention strategies.

Graphical abstract

Keywords

Lacunar stroke intracerebral haemorrhage deep perforator arteriopathy hypertension neuroimaging MRI clinical outcomes

Introduction

Small vessel disease of the deep perforating arterioles (deep perforator arteriolopathy; DPA) is a chronic neurovascular disease causing intracerebral haemorrhage (ICH) and lacunar ischaemic stroke.¹ Prior reports found DPA is the underlying cause of one out of two cases of ICH^2,3 and of about 16% of ischaemic strokes.⁴ DPA mainly affects small vessels in the deep supratentorial structures (basal ganglia, thalamus) and the brainstem.^3,5 Characteristic MRI features of DPA include periventricular white matter hyperintensities,⁶ lacunes of presumed vascular origin⁷ and deep cerebral microbleeds.⁸ Although deep ICH and lacunar strokes (LS) are both considered to be consequences of DPA, comparisons of clinical and imaging characteristics between the two manifestations are scarce.⁹ However, additional knowledge on similarities and differences in risk factors and outcomes might help guide preventive treatment strategies and identify diverging risk profiles.

With this study, we aimed to compare clinical and neuroimaging characteristics of patients with DPA-associated deep ICH versus LS and determine the association of the event type with small vessel disease burden, recurrent events and functional outcome.

Methods

Study design and setting

We conducted a retrospective analysis using data from the prospective Bernese Stroke Registry. The study design adhered to the recommendations provided in the Framework for Clinical Trials in Cerebral Small Vessel Disease (FINESSE).¹⁰ This registry enrols all consecutive patients with acute ischaemic stroke or non-traumatic intracerebral haemorrhage treated at the Inselspital Bern University Hospital admitted within 7 days of symptom onset. Since we are a tertiary referral centre, peripheral hospitals and stroke units refer patients who are potentially eligible for interventions (endovascular therapy, surgical revascularization etc.). Among patients enrolled between 2014 and 2019, we included all patients aged ⩾18 years with an MRI-confirmed diagnosis of a DPA-related cerebrovascular event (intracerebral haemorrhage or isolated lacunar ischaemic stroke) in deep supratentorial structures (basal ganglia, internal and external capsule, thalamus) or the brainstem. Locations were defined according to the Cerebral Haemorrhage Anatomical Rating Scale (CHARTS).¹¹ Lacunar stroke was defined according to the STRIVE criteria,¹² based on imaging (DWI lesion in acute MRI or CT perfusion deficit in acute imaging plus follow-up MRI with corresponding DWI and/or FLAIR lesion in the same localization) and sudden-onset corresponding clinic.

We excluded patients where the index event was not attributable to DPA, that is, ischaemic stroke of non-DPA aetiology (i.e. cardioembolic stroke, large artery disease, vasculitis) or ICH aetiology other than DPA (i.e. cerebral amyloid angiopathy, arteriovenous malformation, Moya-Moya, vasculitis). Despite being considered as a ‘deep and infratentorial’ region according to CHARTS,¹¹ we excluded patients with cerebellar ICH as recent evidence suggests a major role of CAA in cerebellar ICH.¹³ Further, we excluded patients that did not undergo MRI within 3 months of baseline event. Reporting of this manuscript follows the RECORD reporting guideline.¹⁴

Clinical variables

Clinical variables were collected by trained research staff using pre-defined variables and electronic case report forms as reported in prior research.^2,15,16 We extracted the following variables: demographic information (age, sex), clinical presentation (NIHSS, blood pressure and glucose on admission), medical history (pre-stroke modified Rankin Scale (mRS), known arterial hypertension, diabetes, hyperlipidaemia, atrial fibrillation, previous ischaemic stroke or intracerebral haemorrhage) and medication (antiplatelet, anticoagulation, antihypertensives or lipid-lowering drug intake prior to event). Hypertension was defined as previously known hypertension and/or use of antihypertensive treatment without concurrent indication. History of stroke was defined as a clinically overt ischaemic stroke event prior to the index event.

Work-up

MRI is the preferred acute neurovascular imaging modality in the emergency department at our centre for all patients admitted with a suspicion of stroke (ischaemic or haemorrhagic) regardless of onset time, and available 24/7. Thus, MRI is performed in most patients unless there is a contraindication (i.e. pacemaker) directly upon admission as first-line imaging. At our hospital, all patients admitted with ischaemic stroke undergo routine aetiological work-up, including vascular imaging, echocardiography and at least three times 7 day electrocardiography follow-up to detect potential cardioembolic sources for ischaemic events. Patients with ICH undergo vascular imaging to detect potentially underlying macrovascular ICH causes and MRI as soon as clinically feasible (if admission imaging modality was not MRI).

MRI analysis

We included all clinical MRIs performed within 3 months of the index event for the current analysis. MRIs were performed as part of clinical routine diagnostic and follow-up using Siemens MAGNETOM Avanto, Avanto Fit and Era at 1.5T and Siemens MAGNETOM Verio, Skyra Fit and Vida and TrioTim at 3T. Two trained neurovascular research fellows (MBG and JV) assessed the following MRI markers as defined by the STRIVE guidelines¹² and previously published scales and scores^11,17–20: Cerebral microbleeds (CMB)¹⁸ and cortical superficial siderosis (cSS)²⁰ both on susceptibility weighted imaging (SWI); white matter hyperintensities (WMH, Fazekas score¹⁷) and lacunes (corresponding to a modified MARS classification as performed in previous research^18,21) and enlarged perivascular spaces (PVS)¹⁹ on T2, if available, or fluid attenuated inversion recovery imaging (FLAIR), which is part of the stroke protocol. Haematoma/ischaemic stroke location was determined using the CHARTS.¹¹ Raters were blinded to clinical data but not to the type of index event. We calculated the overall disease burden according to the small vessel disease burden score (SVD-burden) described by Staals et al.²² To ensure consistency of rating, a random sample of 10 MRIs (five with ICH and five with LS) were assessed by both raters.

Clinical follow-up

All patients enrolled in the Bernese Stroke Registry receive standardized follow-up visits at 3 months via in-person visits or structured telephone interviews complemented by outpatient/rehabilitation facility reports. The following outcomes are assessed: modified Rankin Scale score (mRS) and clinically symptomatic ischaemic stroke or ICH at 3 months.

Informed consent and ethical board review

As by Swiss law on highly specialized medicine, all patients with cerebrovascular diseases who are treated at a Swiss Stroke Center have to be enrolled into the Bernese Stroke Registry for quality control purposes. All our patients are informed about this registry and can refuse the further use of their health-related data for research purposes. In compliance with national regulations, we included all patients who did not opt out. The analysis was approved by the ethical board of the Canton of Bern (KEK-BE 2019-00689 and EK BE 231/14/PB_2016-01905).

Outcome definitions

The primary outcome of this analysis was SVD burden, as defined according to the SVD burden score by Staals et al. (range 0–4, 1 point each is attributed for the presence of CMBs or lacunes, respectively, 1 point for >20 basal ganglia perivascular spaces, 1 point for periventricular white matter hyperintensities Fazekas 3 or deep white matter hyperintensities Fazekas 2 or 3).²² Secondary outcomes included mRS and recurrent cerebrovascular events (ischaemic stroke, ICH) within 3 months after the index event.

Statistical analysis

Statistical analysis was performed by MBG according to a prespecified statistical analysis plan developed by MBG, JV and DS (unpublished) using STATA MP/16.0 (Stata Corp., Inc.). We compared baseline characteristics of patients with deep ICH versus lacunar strokes using the Table 1_mc package in Stata.²³ For binary and categorical data, we used Pearson’s chi² test or Fisher’s exact test, as appropriate. For continuous data, we report mean and standard deviation (SD) for normally distributed, and median and interquartile range (IQR) for non-normally distributed data. The significance level for all analyses was set to α < 0.05.

Table 1.

Baseline characteristics.

	Intracerebral haemorrhage	Lacunar stroke	p-Value
	N = 117	N = 599
Demographics and baseline characteristics
Age (years); mean (SD)	65.1 (15.2)	69.7 (13.6)	0.001
Female sex	43 (37.1%)	239 (39.9%)	0.57
Pre-stroke mRS			<0.001
0	44 (41.5%)	482 (80.5%)
1	29 (27.4%)	58 (9.7%)
2	23 (21.7%)	24 (4.0%)
3	9 (8.5%)	35 (5.8%)
4	1 (0.9%)	0 (0.0%)
Known hypertension	82 (71.3%)	468 (78.1%)	0.11
Diabetes mellitus	22 (22.9%)	133 (22.2%)	0.88
Hyperlipidaemia	63 (64.9%)	424 (70.8%)	0.24
Atrial fibrillation	23 (20.0%)	72 (12.0%)	0.021
History of stroke	10 (8.7%)	84 (14.0%)	0.12
History of ICH	8 (7.0%)	5 (0.8%)	<0.001
Medication prior to event
Antiplatelet therapy	17 (15.3%)	202 (33.7%)	<0.001
Anticoagulation			0.11
none	83 (83.8%)	543 (90.7%)
Direct oral anticoagulants	11 (11.1%)	31 (5.2%)
Vitamin K antagonists	5 (5.1%)	22 (3.7%)
Parenteral anticoagulation	0 (0.0%)	3 (0.5%)
Antihypertensive drugs pre-stroke	58 (52.3%)	321 (53.6%)	0.8
Lipid-lowering therapy	15 (19.7%)	154 (25.7%)	0.26
Intravenous thrombolysis	0 (0.0%)	129 (21.5%)	<0.001
Clinical findings
Median NIHSS on admission (IQR)	11 (4–17)	3 (1–5)	<0.001
First systolic blood pressure (mmHg)	183.1 (35.0)	173.8 (31.1)	0.004
First diastolic blood pressure (mmHg)	101.5 (24.2)	88.8 (18.4)	<0.001
Blood glucose on admission	7.1 (2.2)	6.9 (2.4)	0.29

p-values in bold indicate that these values were below the significance level.

Missing values excluded from the total for all comparative tables.

Regression analyses

We performed univariable and multivariable regressions, of which we present odds ratio (OR) point estimates and 95%-confidence intervals (95%CI). To assess the consistency of our findings, we developed two models adjusting for pre-specified, clinically plausible confounders, taking into account the risk of overadjustment by including a maximum of 1 covariable per 10 outcome events in at least one of the two models. We performed ordered logistic regressions to determine the association of event type (deep ICH vs LS) with the SVD burden score and mRS at 3 months. The models were adjusted for (a) age, arterial hypertension, antiplatelet therapy, anticoagulation therapy and (b) age, arterial hypertension and pre-stroke mRS. Further, we performed firth logistic regressions to determine the association of event type with (1) recurrent ICH and (2) ischaemic stroke at 3 months adjusting for (a) age and arterial hypertension and (b) atrial fibrillation and SVD burden score.

To account for the potentially elevated risk of recurrent cerebrovascular events in patients with previous cerebrovascular events, we performed a sensitivity analysis, limiting the dataset to patients with first-ever cerebrovascular event.

Missing data handling

We compared baseline characteristics of patients with versus without available outcomes using descriptive statistical methods. Due to the high risk of introducing additional bias, we did not impute outcome data. Thus, patients with missing outcomes were excluded of all regression analyses (listwise deletion).

The results of this study are reported according to the STROBE guidelines and RECORD extensions.¹⁴

Results

Baseline characteristics

Amongst all consecutive 914 patients with ICH (any location) and 6645 patients with all-type ischaemic strokes, we included 716 patients (see Figure 1 patient flowchart): 599 patients with MRI-confirmed LS (84%) and 117 patients with deep ICH (16%). For 610 patients (85%), the index event was the first cerebrovascular event. Patients with deep ICH were younger, had a higher pre-stroke mRS and were less often on antiplatelet therapy. The prevalence of prior ischaemic strokes was similar amongst both groups, but deep ICH patients had a significantly higher prevalence of previous ICH (7.0% vs 0.8%, p < 0.001). On admission, NIHSS, systolic and diastolic blood pressure were significantly higher in deep ICH patients. Baseline clinical findings are displayed in Table 1. Median time between symptom onset and MRI was 0 days (IQR 0–3) for ICH patients, but this information was not available for patients with LS. MRI is the preferred imaging modality on admission in all patients with suspected stroke, therefore the vast majority of patients with LS received MRI on admission, similar to patients with ICH. While 599/652 patients with LS (92%) underwent MRI, this was the case for 117/322 patients with deep ICH. A comparison between ICH patients with versus without MRI is provided in Supplemental Table 1. Patients with deep ICH had a significantly higher SVD burden score with a median of 2 points (IQR 1–2) versus 1 point (IQR 0–2) in LS (p < 0.001, Figure 2). In only 9.8% of patients with deep ICH versus 37.1% with LS (p < 0.001), there were no visible signs of SVD. In particular, we observed a higher total burden of cerebral microbleeds (for prevalence and count). While supratentorial and brainstem lacunes were more prevalent in deep ICH, prevalence of cerebellar lacunes was similar in both subgroups. On the other hand, basal ganglia PVS were more prevalent in LS than deep ICH. We did not observe any differences in white matter hyperintensity severity according to Fazekas grading. Table 2 displays neuroimaging small vessel disease findings.

Figure 1.

Study flowchart.

Figure 2.

Small vessel disease burden score.

Table 2.

Neuroimaging findings.

		Intracerebral haemorrhage	Lacunar stroke	p-value
		N = 117	N = 599
Lesion localization
Supratentorial		109 (93.2%)	434 (72.5%)	<0.001
Brainstem		8 (6.8%)	165 (27.5%)	<0.001
Cerebral microbleeds (CMBs) (n (%))
Any CMB		83 (71.6%)	175 (29.2%)	<0.001
Median CMB count if CMBs present (IQR)		4 (2–12)	2 (1–6)	0.002
Brainstem CMBs	0	95 (81.9%)	575 (96.0%)	<0.001
	1	8 (6.9%)	16 (2.7%)
	2-5	11 (9.5%)	8 (1.3%)
	6-10	2 (1.7%)	0 (0.0%)
Median brainstem CMB count if CMBs present (IQR)		2 (1–3)	1 (1–2)	0.044
Cerebellar CMBs	0	74 (63.8%)	557 (93.0%)	<0.001
	1	42 (36.2%)	19 (3.2%)
	2-5	0 (0.0%)	17 (2.8%)
	6-10	0 (0.0%)	3 (0.5%)
	11-20	0 (0.0%)	2 (0.3%)
	>20	0 (0.0%)	1 (0.2%)
Median cerebellar CMB count if CMBs present (IQR)		1 (1–1)	2 (1–4)	<0.001
Deep CMBs	0	50 (43.1%)	504 (84.1%)	<0.001
	1	20 (17.2%)	37 (6.2%)
	2-5	30 (25.9%)	47 (7.8%)
	6-10	9 (7.8%)	6 (1.0%)
	11-20	7 (6.0%)	5 (0.8%)
Median deep CMB count if CMBs present (IQR)		2 (1–4)	2 (1–4)	0.18
Lobar CMBs	0	63 (54.3%)	478 (79.8%)	<0.001
	1	13 (11.2%)	53 (8.8%)
	2-5	20 (17.2%)	43 (7.2%)
	6-10	9 (7.8%)	12 (2.0%)
	11-20	8 (6.9%)	7 (1.2%)
	>20	3 (2.6%)	6 (1.0%)
Median lobar CMB count if CMBs present (IQR)		4 (2–8)	2 (1–5)	0.004
White matter hyperintensities (WMH, according to Fazekas grading)
Periventricular WMH
0		34 (30.1%)	179 (29.9%)	0.87
1		38 (33.6%)	182 (30.4%)
2		27 (23.9%)	151 (25.2%)
3		14 (12.4%)	87 (14.5%)
Deep WMH
0		31 (27.2%)	166 (27.7%)	0.85
1		50 (43.9%)	247 (41.2%)
2		22 (19.3%)	135 (22.5%)
3		11 (9.6%)	51 (8.5%)
Lacunes
Any lacune		69 (60.5%)	153 (27.4%)	<0.001
Median lacunes count if lacunes present (IQR)		1 (1–1)	1 (1–2)	<0.001
Supratentorial lacunes	0	48 (41.4%)	458 (76.5%)	<0.001
	1	21 (18.1%)	92 (15.4%)
	2-5	30 (25.9%)	45 (7.5%)
	6-10	12 (10.3%)	3 (0.5%)
	11-20	4 (3.4%)	1 (0.2%)
	>20	1 (0.9%)	0 (0.0%)
Median supratentorial lacunes count (if lacunes present (IQR)		3 (1–5.5)	1 (1–2)	<0.001
Brainstem lacunes	0	100 (87.0%)	568 (94.8%)	<0.001
	1	10 (8.7%)	28 (4.7%)
	2-5	5 (4.3%)	3 (0.5%)
Median brainstem lacunes count if lacunes present (IQR)		1 (1–2)	1 (1–1)	0.076
Cerebellar lacunes	0	103 (88.8%)	531 (88.6%)	1
	1	10 (8.6%)	52 (8.7%)
	2-5	3 (2.6%)	16 (2.7%)
Median cerebellar lacunes count if lacunes present (IQR)		1 (1–1)	1 (1–1)	0.92
Perivascular spaces
Basal ganglia-PVS
no PVS		21 (17.9%)	0 (0.0%)	<0.001
1–10 PVS		63 (53.8%)	354 (73.3%)
11–20 PVS		24 (20.5%)	111 (23.0%)
21–40 PVS		9 (7.7%)	18 (3.7%)
Total small vessel disease burden
SVD burden score (median (IQR))		2 (1–2)	1 (0–2)	<0.001
Small vessel disease burden score
0		11 ( 9.8%)	222 (37.1%)	<0.001
1		40 (35.7%)	176 (29.4%)
2		38 (33.9%)	114 (19.0%)
3		20 (17.9%)	54 (9.0%)
4		3 (2.7%)	33 (5.5%)

p-values in bold indicate that these values were below the significance level.

Missing values excluded from the total for all comparative tables.

Sensitivity analysis with patients with first-ever cerebrovascular event

Baseline characteristics of patients with first-ever cerebrovascular events were similar to the full dataset, but with slightly lower pre-stroke mRS (for a full baseline table, see Supplemental Table 2). The small vessel disease score was higher in deep ICH than LS (Supplemental Table 3, p < 0.001) and patients with deep ICH had poorer functional outcome at 3 months than those with LS (Supplemental Table 4).

Association of SVD burden score with vascular risk factors and pre-stroke mRS

In the univariable, ordinal shift analysis, deep ICH, higher age, known hypertension, antiplatelet therapy, anticoagulation, admission NIHSS and higher pre-stroke mRS were associated with higher SVD burden score. After adjustment for age, hypertension and antithrombotic therapy, deep ICH was associated with SVD burden score severity (model 1, complete case analysis with n = 698 patients: aOR for ordinal shift 3.19, 95%CI 2.15–4.75). This was consistent in model 2, where we adjusted for age, hypertension and pre-stroke mRS (model 2, complete case analysis with n = 692 patients: aOR 2.94, 95%CI 1.97–4.36). The sensitivity analyses restricted to first-ever cerebrovascular events corroborated these findings. Results of the regression analyses are displayed in Table 3.

Table 3.

Regression analyses.

	Univariable analysis		Model 1		Model 2		Sensitivity analysis: First-ever event, model 1		Sensitivity analysis: First-ever event, model 2
	OR	95%CI	aOR	95%CI	aOR	95%CI	aOR	95%CI	aOR	95%CI
SVD burden score (ordinal shift)
Intracerebral haemorrhage	2.39,*	1.69–3.38	3.19,*	2.15–4.75	2.94,*	1.97–4.36	3.59,*	2.33–5.53	3.68,*	2.37–5.72
Age (years)	1.04,*	1.03–1.05	1.04,*	1.02–1.05	1.04,*	1.03–1.05	1.04,*	1.03–1.06	1.04,*	1.03–1.06
Hypertension	3.32,*	2.38–4.63	2.27,*	1.58–3.26	2.34,*	1.64–3.33	2.25,*	1.52–3.32	2.23,*	1.51–3.28
Atrial fibrillation	1.76,*	1.19–2.60
Antiplatelet therapy	1.72,*	1.29–2.84	1.36	0.99–1.87			1.02	0.71–1.46
Anticoagulation	2.35,*	1.51–3.68	1.56	0.97–2.50			1.61,*	0.94–2.78
Admission NIHSS	1.07,*	1.04–1.10
Pre-stroke mRS	1.54,*	1.33–1.80			1.15	0.97–1.36			1.02	0.84–1.25
Modified Rankin Scale at 3 months (ordinal shift)
Intracerebral haemorrhage	9.86,*	6.32–15.39	2.16,*	1.21–3.87	2.79,*	1.46–5.32	2.09,*	1.12–3.90	2.20,*	1.11–4.37
Age	1.02,*	1.01–1.03	1.01	1.00–1.02	1.00	0.99–1.02	1.01	0.99–1.02	1.00	0.98–1.01
Hypertension	1.54,*	1.10–2.17	1.36	0.92–2.00	1.18	0.79–1.78	1.48	0.97–2.24	1.30	0.84–2.00
Admission NIHSS	1.27,*	1.22–1.33	1.22,*	1.16–1.28	1.22,*	1.16–1.28	1.23,*	1.17–1.29	1.25,*	1.18–1.32
Pre-stroke mRS	2.52,*	2.09–3.03	2.10,*	1.72–2.57	1.93,*	1.57–2.38	2.30,*	1.82–2.92	2.15,*	1.69–2.74
Antiplatelet	1.69,*	1.23–2.33			1.76,*	1.21–2.56			1.69,*	1.10–2.59
Anticoagulation	2.14,*	1.33–3.45			2.16,*	1.13–4.10			2.18,*	1.05–4.51
Atrial fibrillation	1.96,*	1.30–2.97			1.08	0.62–1.88			1.25	0.67–2.33
SVD burden score	1.54,*	1.36–1.76			1.18,*	1.02–1.37			1.18	1.00–1.41
Intracerebral haemorrhage at follow-up (logistic regression)
Intracerebral haemorrhage	13.52	0.55–334.66	12.04	0.45–320.71	19.54	0.49–773.10	3.75	0.07–210.73	2.11	0.04–113.32
Age	0.96	0.88–1.06	0.97	0.85–1.10			0.99	0.82–1.20
Hypertension	0.98	0.04–24.18	1.55	0.03–77.29			0.33	0.00–43.63
Antiplatelet therapy	2.66	0.05–134.87
Anticoagulation	2.09	0.08–51.87
SVD burden score	1.10	0.29–4.13			0.63	0.09–4.36			1.60	0.20–13.01
Atrial fibrillation	1.58	0.06–39.08			2.06	0.07–57.24			3.31	0.04–264.11
Ischaemic stroke at follow-up (logistic regression)
Intracerebral haemorrhage	1.63	0.54–4.90	1.61	0.53–4.92	1.00	0.29–3.39	2.50	0.74–8.38	1.65	0.44–6.10
Age	0.99	0.96–1.02	0.98	0.95–1.02			0.97	0.93–1.02
Hypertension	1.93	0.50–7.50	2.49	0.56–11.01			2.00	0.40–9.94
Antiplatelet therapy	1.28	0.46–3.61
Anticoagulation	2.34	0.77–7.09
SVD burden score	1.65,*	1.12–2.43			1.62,*	1.10–2.40			1.86,*	1.14–3.04
Atrial fibrillation	1.74	0.58–5.24			1.55	0.50–4.80			1.00	0.22–4.44

p < 0.05.

Functional outcome at 3 months

Functional outcome data (mRS) was available for 91 patients with deep ICH (77.8%) and for 476 patients with LS (79.5%). Patients with available mRS were younger, more often male, had lower pre-stroke mRS, a higher prevalence of atrial fibrillation and a higher admission NIHSS (Supplemental Table 5). There were no differences in prevalence and distribution of SVD markers between patients with missing and available mRS (Supplemental Table 5). Deep ICH and LS both lead to an increase in mRS at 3 months compared to baseline (Figure 3(a)). Functional outcome was significantly worse in deep ICH compared to LS. At 3 months, 29.7% of patients with deep ICH and 74.6% of patients with LS were functionally independent (mRS 0–2). While 28/117 (23.9% with available mRS) deep ICH patients had died, this was the case in 16/540 (3.0%) LS patients with available vital status.

Figure 3.

Outcomes in patients with deep ICH versus lacunar stroke: (a) Comparison of pre-event and 3 month modified Rankin Scale (mRS) in patients with lacunar stroke (n = 476) versus deep ICH (n = 91). Pre-event and 3 month mRS were significantly higher in patients with deep ICH (p < 0.001 for both timepoints) and (b) Event rates in patients with lacunar stroke versus deep intracerebral haemorrhage (ICH) at 3 month follow-up. Patients with deep ICH had numerically higher rates of cerebrovascular events compared to patients with LS but this difference was not statistically significant (p for ischaemic stroke at follow-up 0.51, p for ICH at follow-up 0.18).

All investigated variables were associated with poor functional outcome (mRS) at 3 months in univariable analysis. After adjustment for age, hypertension, stroke severity (NIHSS) on admission and pre-stroke mRS, deep ICH (vs LS) was significantly associated with poorer functional outcome (model 1, complete case analysis with n = 543 patients: aOR for higher mRS 2.16, 95%CI 1.21–3.87). This association persisted when adjusting for age, hypertension, admission NIHSS, pre-stroke mRS, previous antithrombotic therapy, pre-existing atrial fibrillation and SVD burden score (model 2, complete case analysis with n = 526 patients: aOR 2.79, 95%CI 1.46–5.32). Results of the sensitivity analyses were comparable (Table 3).

Recurrent events at 3 months

During the follow-up period of 3 months, 12 patients with LS (2.9%, 3 of them with atrial fibrillation) and 4 (4.3%, 1 with atrial fibrillation) with deep ICH suffered a (recurrent) ischaemic stroke, while no patient with index LS and only one deep ICH patient (1.1%) suffered (recurrent) ICH. Rates of recurrent ischaemic stroke and ICH were numerically higher in deep ICH than LS (Figure 3(b)). Information on recurrent cerebrovascular events was available in 92/117 patients with ICH (78.6%) and 412/599 patients with LS (68.8%). Patients with available data were younger, more often male, and the prevalences of diabetes and antiplatelet therapy where lower, while the ones for atrial fibrillation and anticoagulation were higher (Supplemental Table 6). A comparison of baseline and neuroimaging characteristics in patients with missing versus available outcome data is provided in the supplement.

In the univariable analysis, higher SVD burden score was associated with higher odds of ischaemic stroke at follow-up. None of the evaluated parameters was associated with recurrent ICH. In the adjusted analyses, there was no association of the index event type with recurrent ischaemic strokes or ICH at 3 months. We found that higher SVD burden score was independently associated with recurrent ischaemic stroke. Findings of the sensitivity analysis were comparable.

Discussion

In this analysis from the Bernese Stroke Registry, we compared clinical and neuroimaging presentation of DPA-associated ICH to LS and investigated the association of the index event type with the baseline SVD burden score, functional outcome and rate of recurrent cerebrovascular events at 3 months. Compared to LS, deep ICH patients were younger and had a higher pre-event mRS, cerebrovascular risk factors at baseline were similar in both groups. While atrial fibrillation was more prevalent in deep ICH patients, there was no significant difference in baseline anticoagulation. Prevalence and severity of SVD neuroimaging markers were higher in patients with deep ICH compared to LS, which resulted in a higher SVD burden score. The risk of ischaemic stroke in the 3 months following the index event was numerically higher than that of recurrent ICH in deep ICH patients and even outweighed the risk of ischaemic stroke in LS patients.

This comparison of two cerebrovascular complications of DPA contributes towards understanding why some patients with DPA suffer LS while others suffer deep ICH. It brings new insights into the clinical and neuroimaging presentations of DPA and the event-specific risk of recurrent cerebrovascular events. This information is of high clinical relevance as it may influence future studies investigating risk-based secondary prevention strategies. This is of even greater importance in deep ICH, given that there are only few evidence-based treatment options available for these patients and the risk of recurrent events is high with differing risk profiles between CAA and non-CAA ICH.^2,24

Although patients with deep ICH were younger than LS patients, they had poorer pre-stroke mRS and a higher total small vessel disease burden at baseline. It is particularly interesting that patients with lacunar strokes were more frequently on aspirin than those with deep ICH. A potential explanation is that certain comorbidities and risk factors were not diagnosed prior to the index ICH and therefore untreated, resulting in higher morbidity and ultimately a more severe complication of DPA. On the other hand, we cannot rule out that some comorbidities prompting the use of aspirin were more prevalent in patients with LS than deep ICH. It is plausible that certain mechanisms predisposing to development of deep ICH or LS in DPA differ. In particular, genetic factors contributing not only to lobar, but also to deep ICH and progression of small vessel disease have been discussed.^25,26 Apart from cerebrovascular events, SVD is a relevant contributor towards cognitive impairment,²⁷ which might not be adequately represented in the mRS. Given that affected patients are often relatively young (mean age in deep ICH 65.1 years, in LS 69.7 years) and the consequences of both event types disabling, these observations highlight the urgent need for effective primary prevention.

We confirm the higher prevalence of CMBs in non-lobar ICH than in LS and the similar severity and distribution of white matter hyperintensities in these entities, which were reported in a previous study.⁹ While Wiegertjes et al. determined the severity of white matter hyperintensities using volumetry,⁹ we applied the semiquantitative score by Fazekas et al., which is frequently used in clinical routine due to its straight-forward approach.¹⁷ In contrast to previous work,⁹ we found the prevalence of supratentorial and brainstem lacunes to be higher in deep ICH than in LS, which also contributes to the higher overall small vessel disease burden in deep ICH. While the prevalence of lacunes in ischaemic strokes is comparable to the one reported in literature,²² the number of lacunes in deep ICH is higher than previously reported prevalences of around 20%.^28,29 However, data on the prevalence of lacunes according to the ICH location differ and several studies that compared small vessel disease markers in lobar versus non-lobar ICH did not report the presence of lacunes.^30,31 Further differences include a higher prevalence of basal ganglia perivascular spaces in deep ICH compared to previous studies,¹⁹ but a lower prevalence in LS.²² Potential reasons for these differences may include variability in MRI protocols, including MRI field strength (previous studies used 1.5T,^19,22 we used both 1.5T and 3T MRIs), but also patient-related quality of imaging acquisition (e.g. patient was moving in the MRI), which is in our experience poorer in the acute care setting. Higher field strength was shown to be associated with a higher CMB count,³² which is presumably also the case for other SVD markers. Whether this affects the total SVD burden score (which does not take into account the lesion count for CMBs and lacunes, but only their presence) still needs to be investigated. As field strength was not available on pseudonymized imagings, we were not able to adjust for this potential confounder. However, in our centre, patients are assigned to the different field strengths based on scanner availability in a quasi-randomized way. Therefore, we are confident that the observed differences reflect real differences rather than technical variation.

In our study with more than 700 patients with DPA-associated cerebrovascular events, only one patient suffered recurrent ICH during follow-up. Interestingly, the incidence of ischaemic strokes within 3 months was numerically higher than that for ICH in both subgroups, advocating for shared pathomechanisms in deep ICH and LS. Our data corroborate the findings of recent observational and interventional studies, which observed markedly lower incidences of recurrent ICH and an important risk of ischaemic strokes in ICH patients.^24,33,34 In the light of the high risk of ischaemic strokes and the relative contraindication for intravenous thrombolysis shortly after intracerebral haemorrhage, future studies should further elucidate the safety of antithrombotic medication to prevent ischaemic strokes after ICH. Based on the findings of the RESTART trial, this might include administration of antiplatelets, but other, potentially safer options should be sought as well.³³

This study has the following strengths: Data are provided through a prospectively collected, consecutive hospital-based registry from a large tertiary care hospital providing specialized stroke care for a region of about 1.5 million inhabitants. MRI is widely available and the standard acute care imaging in our hospital. Neuroimaging markers were assessed according to recent guidelines using previously published, well-established scales and scores by experienced neurovascular researchers. Their application is feasible in clinical routine and might inform physicians about the severity of the disease, which correlates with the risk of recurrent events.

Due to the registry-based nature of the study, there are some inherent limitations: First, this is routinely collected data and this study was not planned at the time of the database implementation. This is reflected in the fact that some variables were only collected in one subtype of patients. In particular, smoking (which is associated with a higher SVD burden²²) and coronary heart disease were only available for LS, but not deep ICH patients. Further, questions such as prescription of antithrombotic therapy after the index event remain open. However, patients were treated in a single tertiary centre adhering to international and institutional guidelines. Also, the proportion of deep ICH compared to all ICH and LS to all ischaemic strokes, respectively, was lower in our centre in comparison to other studies, which presumably is related to the fact that we are a tertiary stroke centre, offering interventional therapies, resulting in a patient population enriched with patients with large vessel-occlusion strokes and lobar ICH. Second, analysis was performed in MRIs from clinical routine with different imaging protocols, which – despite adding to heterogeneity – is an argument for the generalizability of our findings. Third, despite structured outcome assessment, 3 month outcomes were only available in 78.6% of patients with deep ICH and 68.8% with LS. We therefore urge caution in the interpretation of our results, although patients with missing outcomes were in a similar health condition at baseline and had a lower NIHSS on admission. On the other hand, given that certain patients can only undergo MRI with special precautions, there is a risk of selection bias resulting from this inclusion criterion. Patients with deep ICH underwent MRI significantly less often than patients with LS, which suggests we might even have underestimated the true difference. The majority of patients who did not undergo MRI were ICH patients, with several prognostically unfavourable characteristics, including higher age, higher prevalences of hyperlipidaemia and anticoagulation^35,36 lower admission GCS and higher admission NIHSS (which was recently demonstrated to correlate well with haematoma volume in ICH³⁷). This was also reflected in poorer functional outcome at 3 months.

Further, lesion volumes for ICH and LS were not available. Nevertheless, the volume alone is not indicative for the stroke severity and patients’ handicaps, as affected regions may vary.³⁸ Therefore, we included the admission NIHSS as a surrogate for the stroke severity. Lastly, we used deep and brainstem location as a surrogate criterion for the diagnosis of DPA, as there are currently no accepted criteria for its non-invasive, in-vivo diagnosis.

Conclusion

Taken together, we present hypothesis-generating data comparing presentation and outcomes in patients with deep ICH and LS. Both are complications of DPA and lead to relevant morbidity. While prevalence of conventional cerebrovascular risk factors is similar in LS and deep ICH, deep ICH have more severe SVD manifestations. The risk of ischaemic strokes is high in deep ICH and LS in the subacute phase, while the risk for ICH after LS seems negligible. An applicable and reproducible definition for DPA (comparable to the Boston criteria in cerebral amyloid angiopathy³⁹) would facilitate research into SVD-specific treatments, which are urgently needed.

Supplemental Material

sj-docx-1-eso-10.1177_23969873231193237 – Supplemental material for Small vessel disease burden and risk of recurrent cerebrovascular events in patients with lacunar stroke and intracerebral haemorrhage attributable to deep perforator arteriolopathy

Supplemental material, sj-docx-1-eso-10.1177_23969873231193237 for Small vessel disease burden and risk of recurrent cerebrovascular events in patients with lacunar stroke and intracerebral haemorrhage attributable to deep perforator arteriolopathy by Martina B Goeldlin, Jan Vynckier, Madlaine Mueller, Boudewijn Drop, Basel Maamari, Noah Vonlanthen, Bernhard M Siepen, Arsany Hakim, Johannes Kaesmacher, Christopher Marvin Jesse, Mandy D Mueller, Thomas R Meinel, Morin Beyeler, Leander Clénin, Jan Gralla, Werner Z’Graggen, David Bervini, Marcel Arnold, Urs Fischer and David J Seiffge in European Stroke Journal

Supplemental Material

sj-docx-2-eso-10.1177_23969873231193237 – Supplemental material for Small vessel disease burden and risk of recurrent cerebrovascular events in patients with lacunar stroke and intracerebral haemorrhage attributable to deep perforator arteriolopathy

Supplemental material, sj-docx-2-eso-10.1177_23969873231193237 for Small vessel disease burden and risk of recurrent cerebrovascular events in patients with lacunar stroke and intracerebral haemorrhage attributable to deep perforator arteriolopathy by Martina B Goeldlin, Jan Vynckier, Madlaine Mueller, Boudewijn Drop, Basel Maamari, Noah Vonlanthen, Bernhard M Siepen, Arsany Hakim, Johannes Kaesmacher, Christopher Marvin Jesse, Mandy D Mueller, Thomas R Meinel, Morin Beyeler, Leander Clénin, Jan Gralla, Werner Z’Graggen, David Bervini, Marcel Arnold, Urs Fischer and David J Seiffge in European Stroke Journal

Footnotes

Acknowledgements

The following collaborators participated in the Swiss Stroke Registry and contributed in a significant and documentable manner but do not qualify as regular co-authors to the present manuscript: Mirjam Heldner, Simon Jung, Leonidas Panos, Hakan Sarikaya, Marianne Kormann, Christine Brülisauer, Liselotte McEvoy, Stefanie Chang.

Declaration of conflicting interest

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Dr. Goeldlin: Grants from Swiss Academy of Medical Sciences/Bangerter-Rhyner-Foundation, Mittelbauvereinigung der Universität Bern (outside the submitted work), Swiss Stroke Society (Förderpreis), European Stroke Organisation (department to department visit grant), Insel Gruppe, and Pfizer congress grant (outside the submitted work). Dr. Siepen: Funding by the Bangerter-Rhyner-Foundation (outside the submitted work). Dr. Fischer: research support from Swiss National Science Foundation, Swiss Heart Foundation, Medtronic (BEYOND SWIFT, SWIFT DIRECT), Stryker, Rapid medical, Penumbra, Phenox (DISTAL), consultancies for Medtronic, Stryker, and CSL Behring (fees paid to institution), participation in an advisory board for Alexion/Portola and Boehringer Ingelheim (fees paid to institution).

Dr. Seiffge: Advisory board for Bayer and Portola/Alexion, Research grant from the Bangerter-Rhyner Foundation. All other authors do not report conflicts of interest.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Dr. Goeldlin has received a protected research time grant from the Swiss Academy of Medical Sciences/Bangerter-Rhyner-Foundation for the submitted work (YTCR 13/18). The funder did not have any influence on the study design, conduction, data analysis or interpretation of the results.

Ethical approval

The analysis was approved by the ethical board of the Canton of Bern (KEK-BE 2019-00689 and EK BE 231/14/PB_2016-01905).

Informed consent

Guarantor

DJS.

Contributorship

Study design: MBG, JV, UF, DJS. Data collection: MBG, JV, MM, BD, BM, BS, CMJ, TRM, MB, LC. Statistical analysis: MBG, based on a plan by MBG, JV and DJS.

Manuscript draft: MBG, DJS. Critical revision of the manuscript: all authors.

Trial Registration (where applicable)

Not applicable.

Data availability statement

Upon reasonable request to the corresponding author, a restricted dataset may be shared with qualified researchers after previous clearance from the responsible ethical review board.

ORCID iDs

Martina B Goeldlin

Boudewijn Drop

Arsany Hakim

Thomas R Meinel

Morin Beyeler

Werner Z’Graggen

Urs Fischer

David J Seiffge

Supplemental material

Supplemental material for this article is available online.

References

Poon

MTC

Fonville

Al-Shahi Salman

. Long-term prognosis after intracerebral haemorrhage: systematic review and meta-analysis. J Neurol Neurosurg Psychiatry 2014; 85: 660–667.

Goeldlin

Mueller

Siepen

, et al. Etiology, 3-month functional outcome and recurrent events in non-traumatic intracerebral hemorrhage. J Stroke 2022; 24: 266–277.

Hostettler

Seiffge

Werring

. Intracerebral hemorrhage: an update on diagnosis and treatment. Expert Rev Neurother 2019; 19: 679–694.

Petty

Brown

Whisnant

, et al. Ischemic stroke subtypes. Stroke 1999; 30: 2513–2516.

Wardlaw

Smith

Dichgans

. Small vessel disease: mechanisms and clinical implications. Lancet Neurol 2019; 18: 684–696.

Fazekas

Kleinert

Offenbacher

, et al. Pathologic correlates of incidental MRI white matter signal hyperintensities. Neurology 1993; 43: 1683–1689.

Duering

Csanadi

Gesierich

, et al. Incident lacunes preferentially localize to the edge of white matter hyperintensities: insights into the pathophysiology of cerebral small vessel disease. Brain 2013; 136: 2717–2726.

Charidimou

Imaizumi

Moulin

, et al. Brain hemorrhage recurrence, small vessel disease type, and cerebral microbleeds: a meta-analysis. Neurology 2017; 89: 820–829.

Wiegertjes

Jansen

Jolink

, et al. Differences in cerebral small vessel disease magnetic resonance imaging markers between lacunar stroke and non-lobar intracerebral hemorrhage. Eur Stroke J 2021; 6: 236–244.

10.

Markus

van Der Flier

Smith

, et al. Framework for clinical trials in cerebral small vessel disease (FINESSE): a review. JAMA Neurol 2022; 79: 1187–1198.

11.

Charidimou

Schmitt

Wilson

, et al. The Cerebral Haemorrhage Anatomical RaTing inStrument (CHARTS): Development and assessment of reliability. J Neurol Sci 2017; 372: 178–183.

12.

Wardlaw

Smith

Biessels

, et al. Neuroimaging standards for research into small vessel disease and its contribution to ageing and neurodegeneration. Lancet Neurol 2013; 12: 822–838.

13.

Pasi

Pongpitakmetha

Charidimou

, et al. Cerebellar microbleed distribution patterns and cerebral amyloid angiopathy. Stroke 2019; 50: 1727–1733.

14.

Benchimol

Smeeth

Guttmann

, et al. The REporting of studies Conducted using Observational Routinely-collected health data (RECORD) statement. PLoS Med 2015; 12: e1001885.

15.

Meinel

Branca

De Marchis

, et al. Prior anticoagulation in patients with ischemic stroke and atrial fibrillation. Ann Neurol 2021; 89: 42–53.

16.

Vynckier

Kaesmacher

Wardlaw

, et al. Phenotypes of chronic covert brain infarction in patients with first-ever ischemic stroke: a cohort study. Stroke 2022; 53: 558–568.

17.

Fazekas

Chawluk

Alavi

, et al. MR signal abnormalities at 1.5 T in Alzheimer's dementia and normal aging. AJR Am J Roentgenol 1987; 149: 351–356.

18.

Gregoire

Chaudhary

Brown

, et al. The Microbleed Anatomical Rating Scale (MARS): reliability of a tool to map brain microbleeds. Neurology 2009; 73: 1759–1766.

19.

Charidimou

Boulouis

Pasi

, et al. MRI-visible perivascular spaces in cerebral amyloid angiopathy and hypertensive arteriopathy. Neurology 2017; 88: 1157–1164.

20.

Charidimou

Boulouis

Roongpiboonsopit

, et al. Cortical superficial siderosis multifocality in cerebral amyloid angiopathy: a prospective study. Neurology 2017; 89: 2128–2135.

21.

Pasi

Boulouis

Fotiadis

, et al. Distribution of lacunes in cerebral amyloid angiopathy and hypertensive small vessel disease. Neurology 2017; 88: 2162–2168.

22.

Staals

Makin

Doubal

, et al. Stroke subtype, vascular risk factors, and total MRI brain small-vessel disease burden. Neurology 2014; 83: 1228–1234.

23.

Chatfield

. TABLE1_MC: Stata module to create “table 1” of baseline characteristics for a manuscript. 2020. https://EconPapers.repec.org/RePEc:boc:bocode:s458351

24.

Casolla

Moulin

Kyheng

, et al. Five-year risk of major ischemic and hemorrhagic events after intracerebral hemorrhage. Stroke 2019; 50: 1100–1107.

25.

Biffi

Murphy

Kubiszewski

, et al. APOE genotype, hypertension severity and outcomes after intracerebral haemorrhage. Brain Commun 2019; 1: fcz018.

26.

Biffi

Sonni

Anderson

, et al. Variants at APOE influence risk of deep and lobar intracerebral hemorrhage. Ann Neurol 2010; 68: 934–943.

27.

Pantoni

. Cerebral small vessel disease: from pathogenesis and clinical characteristics to therapeutic challenges. Lancet Neurol 2010; 9: 689–701.

28.

Seiffge

Wilson

Ambler

, et al. Small vessel disease burden and intracerebral haemorrhage in patients taking oral anticoagulants. J Neurol Neurosurg Psychiatry 2021; 92: 805–814.

29.

Lioutas

Norton

, et al. Cerebral small vessel disease burden and functional and radiographic outcomes in intracerebral hemorrhage. J Neurol 2018; 265: 2803–2814.

30.

Boulouis

Charidimou

Auriel

, et al. Intracranial atherosclerosis and cerebral small vessel disease in intracerebral hemorrhage patients. J Neurol Sci 2016; 369: 324–329.

31.

Smith

Nandigam

Chen

, et al. MRI markers of small vessel disease in lobar and deep hemispheric intracerebral hemorrhage. Stroke 2010; 41: 1933–1938.

32.

Nandigam

Viswanathan

Delgado

, et al. MR imaging detection of cerebral microbleeds: effect of susceptibility-weighted imaging, section thickness, and field strength. AJNR Am J Neuroradiol 2009; 30: 338–343.

33.

RESTART Collaboration. Effects of antiplatelet therapy after stroke due to intracerebral haemorrhage (RESTART): a randomised, open-label trial. Lancet 2019; 393: 2613–2623.

34.

Tomari

Levi

Holliday

, et al. One-year risk of stroke after transient ischemic attack or minor stroke in Hunter New England, Australia (INSIST study). Front Neurol 2021; 12: 791193.

35.

Seiffge

Goeldlin

Tatlisumak

, et al. Meta-analysis of haematoma volume, haematoma expansion and mortality in intracerebral haemorrhage associated with oral anticoagulant use. J Neurol 2019; 266: 3126–3135.

36.

Hemphill

Bonovich

Besmertis

, et al. The ICH score. Stroke 2001; 32: 891–897.

37.

Dusenbury

Tsivgoulis

Chang

, et al. Validation of the National Institutes of Health Stroke Scale in intracerebral hemorrhage. Stroke 2023; 3: e000834.

38.

Delcourt

Sato

Zhang

, et al. Intracerebral hemorrhage location and outcome among INTERACT2 participants. Neurology 2017; 88: 1408–1414.

39.

Charidimou

Boulouis

Frosch

, et al. The Boston criteria version 2.0 for cerebral amyloid angiopathy: a multicentre, retrospective, MRI-neuropathology diagnostic accuracy study. Lancet Neurol 2022; 21: 714–725.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.02 MB

0.04 MB