Journal of Neurology (2022) 269:1427–1438 https://doi.org/10.1007/s00415-021-10706-3 ORIGINAL COMMUNICATION Risk of intracranial haemorrhage and ischaemic stroke after convexity subarachnoid haemorrhage in cerebral amyloid angiopathy: international individual patient data pooled analysis Isabel Charlotte Hostettler1 · Duncan Wilson1 · Catherine Arnold Fiebelkorn2 · Diane Aum3 · Sebastián Francisco Ameriso4 · Federico Eberbach4 · Markus Beitzke5 · Timothy Kleinig6 · Thanh Phan7,21 · Sarah Marchina8 · Romain Schneckenburger9 · Maria Carmona‑Iragui10 · Andreas Charidimou11 · Isabelle Mourand12 · Sara Parreira13 · Gareth Ambler14 · Hans Rolf Jäger15 · Shaloo Singhal7,21 · John Ly7,21 · Henry Ma7,21 · Emmanuel Touzé16 · Ruth Geraldes17,18 · Ana Catarina Fonseca13 · Teresa Melo13 · Pierre Labauge12 · Pierre‑Henry Lefèvre19 · Anand Viswanathan11 · Steven Mark Greenberg11 · Juan Fortea10 · Marion Apoil9 · Marion Boulanger9,16 · Fausto Viader9 · Sandeep Kumar8 · Velandai Srikanth7,21 · Ashan Khurram6 · Franz Fazekas5 · Veronica Bruno4 · Gregory Joseph Zipfel3 · Daniel Refai20 · Alejandro Rabinstein2 · Jonathan Graff‑Radford2 · David John Werring1 Received: 22 May 2021 / Revised: 6 July 2021 / Accepted: 9 July 2021 / Published online: 17 July 2021 © The Author(s) 2021 Abstract Objective To investigate the frequency, time-course and predictors of intracerebral haemorrhage (ICH), recurrent convexity subarachnoid haemorrhage (cSAH), and ischemic stroke after cSAH associated with cerebral amyloid angiopathy (CAA). Methods We performed a systematic review and international individual patient-data pooled analysis in patients with cSAH associated with probable or possible CAA diagnosed on baseline MRI using the modified Boston criteria. We used Cox proportional hazards models with a frailty term to account for between-cohort differences. Results We included 190 patients (mean age 74.5 years; 45.3% female) from 13 centers with 385 patient-years of follow- up (median 1.4 years). The risks of each outcome (per patient-year) were: ICH 13.2% (95% CI 9.9–17.4); recurrent cSAH 11.1% (95% CI 7.9–15.2); combined ICH, cSAH, or both 21.4% (95% CI 16.7–26.9), ischemic stroke 5.1% (95% CI 3.1–8) and death 8.3% (95% CI 5.6–11.8). In multivariable models, there is evidence that patients with probable CAA (compared to possible CAA) had a higher risk of ICH (HR 8.45, 95% CI 1.13–75.5, p = 0.02) and cSAH (HR 3.66, 95% CI 0.84–15.9, p = 0.08) but not ischemic stroke (HR 0.56, 95% CI 0.17–1.82, p = 0.33) or mortality (HR 0.54, 95% CI 0.16–1.78, p = 0.31). Conclusions Patients with cSAH associated with probable or possible CAA have high risk of future ICH and recurrent cSAH. Convexity SAH associated with probable (vs possible) CAA is associated with increased risk of ICH, and cSAH but not ischemic stroke. Our data provide precise risk estimates for key vascular events after cSAH associated with CAA which can inform management decisions. Keywords Non-traumatic convexity/convexal/cortical subarachnoid haemorrhage · Intracerebral haemorrhage · Ischemic stroke · Cerebral amyloid angiopathy · Stroke Introduction Convexity subarachnoid haemorrhage (cSAH) describes non-traumatic subarachnoid bleeding limited to the suba- rachnoid space over the convexities of the brain that does not extend into the parenchyma, sylvian fissures, ventricles, * David John Werring or basal cisterns [1, 2]. cSAH may present with transient d.werring@ucl.ac.uk focal neurological episodes (TFNE) of unilateral spread- Extended author information available on the last page of the article ing sensory or motor symptoms [3–5]. In older individuals 13 Vol.:(0123456789) 1428 Journal of Neurology (2022) 269:1427–1438 (over about 60 years), cSAH is often associated with imag- investigation and excluding other causes) according to the ing markers of cerebral amyloid angiopathy (CAA) includ- modified Boston criteria defined by MRI [23]; and available ing cortical superficial siderosis (cSS), cerebral microbleeds follow-up data. We excluded patients with any underlying (CMBs) or both [6]. CAA is characterized by amyloid depo- alternative cause of cSAH including ruptured aneurysm, sition within pial and cortical penetrating arterioles [7, 8] arterio-venous-malformation, tumors, reperfusion injuries and is an important cause of symptomatic intracerebral or hemorrhagic transformation of ischemic stroke, vasculi- haemorrhage (ICH) [9–11]. An aggregate data meta-analysis tis, other inflammatory diseases, or reversible cerebral vaso- from small cohort studies of patients with cSAH and sus- constriction syndrome (RCVS, the most common cause of pected CAA found a high rate of subsequent symptomatic cSAH in younger people, defined as cSAH in patients with ICH [12] but did not investigate recurrent cSAH, ischemic reversible abnormalities in vessel caliber found on CTA, stroke, or mortality. An understanding of the prognosis for MRA or DSA [2, 24]. intracranial bleeding and ischemia is important to inform Anonymized data was transferred from participating management decisions, including starting or restarting cohorts to the Stroke Research Centre, UCL Queen Square antithrombotic drugs. Moreover, previous studies have not Institute of Neurology, using a pre-specified data collection established the time course of these key adverse vascular sheet. Authors contributed individual patient-data, includ- outcomes. ing patient characteristics, past medical history, medication We therefore performed an international collaborative history and, where available, baseline brain imaging markers individual patient data pooled analysis of cohort studies of and follow-up data. Data was checked for internal consist- patients with cSAH and known probable and possible CAA ency with respect̄ to range, and consistency with published status to increase statistical power and improve the preci- reports. Inconsistencies or missing data were reviewed, sion of event rate estimates for ICH, symptomatic ischemic and attempts were made to resolve any inconsistencies by stroke, recurrent cSAH and death. We investigated whether consensus. In case where missing values are present this is neuroimaging markers indicating probable CAA influences indicated in the tables. the risks of these events. Definition of outcomes Material and methods Symptomatic ICH was defined as acute or subacute onset of neurological symptoms (i.e., occurring within a few days Request for individual patient‑data before presentation) with radiological evidence of recent intracerebral haemorrhage (acute blood, perihematomal A repeated systematic literature review (conducted by 2 edema). Recurrent cSAH was defined as acute haemorrhage authors ICH and DW) using PubMed, EMBASE and refer- limited to the subarachnoid space over the convexities of the ence searches identified 21 potentially eligible publications brain, not extending into the parenchyma, sylvian fissures, for this individual patient-data pooled analysis, including ventricles, or basal cisterns [1]. Symptomatic ischemic our own cohort (see Figure e-1) [12]. The search was not stroke was defined as acute or subacute focal neurological restricted by language. As previously described the key symptoms attributed to cerebral infarction confirmed by words used were “convex* adj4 subarachnoid OR cortical brain imaging. Outcome events were ascertained locally by adj4 subarachnoid OR sulc* adj4 subarachnoid”. Of these, contributing teams. All events were truncated at 5 years. 14 groups agreed to contribute individual patient-data of their already published data; if groups included overlapping Risk of bias and study quality samples, we included the latest and larger sample making sure no patient was included twice; we thus included data We reported our study according to the STROBE guide- from 13 cohorts [3, 4, 12–22]. A pre-specified protocol and lines and did the pooled analysis according to the PRISMA analysis plan was agreed upon among all collaborating cent- guidelines [25, 26]. We assessed all studies for risk of bias ers to ensure uniform definitions for diagnosis and outcome and quality using the Cochrane Collaboration tool, which parameters. Clinical and outcome data were collected pro- demonstrated a low risk of bias (see Table e-1). spectively and retrospectively at each center as part of ongo- ing clinical registries. The main inclusion criteria were: non- Radiological data traumatic symptomatic cSAH (haemorrhage judged acute on CT, MRI or both, limited to the subarachnoid space over Probable CAA (versus possible CAA) was diagnosed the convexities of the brain, not extending into the paren- according to the modified Boston criteria by trained observ- chyma, sylvian fissure, ventricles, or basal cisterns [1]); ers at each center. Convexity SAH was not included as a cSAH attributed to probable or possible CAA (after local component of the modified Boston criteria [12, 23]. White 13 Journal of Neurology (2022) 269:1427–1438 1429 matter hyperintensities (leukoaraiosis) were measured using Table 1  Baseline characteristics of the full cohort the van Swieten scale and divided into “severe” (any score of All N = 190 2) or non-severe (no score of 2) [27]. Cerebral microbleeds were rated on blood-sensitive sequences (T2* weighted Age, mean (SD) 74.5 (8.7) or susceptibility weighted images, SWI) [28]. Cortical Female sex 86 (45.3) superficial siderosis (cSS) was rated on T2*-GRE and SWI Current smoker, N (%) 17/169a (10.1) sequences [29, 30]. Current drinker, N (%) 22/149 (14.8) PMH HTN, N (%) 117 (61.6) Statistical analysis Hypercholesterolemia, N (%) 85 (44.7) DM, N (%) 26 (13.7) Univariable analysis AF, N (%) 18/186 (9.9) OAC use, N (%) 15/185 (8.1) We estimated the rates of each outcome event using Kaplan Antiplatelet use, N (%) 63/185 (34.1) Meier survival curves. We investigated the association Statin use, N (%) 66/185 (35.7) between the risk of cSAH and each outcome separately using Anti HTN use, N (%) 104/185 (56.2) the Cox proportional hazards model with a frailty term to Previous ICH, N (%) 24 (12.6) account for differences between study cohorts. The frailty Previous IS, N (%) 21 (11.1) term allows the (absolute) risk to be different in the different Symptoms 188/190 studies, i.e., to adjust for unmeasured study-level covari- Negative, N (%) 90 (47.9) ates. We checked the proportional hazards assumption for all Positive, N (%) 58 (30.9) analyses by visual inspection of the log–log plot of survival Both, N (%) 27 (14.4) (log cumulative hazard versus log time). If the lines were not Headache, N (%) 11 (5.9) parallel, we tested the proportional hazard assumption using Other symptoms, N (%) 2 (1.1) Schoenfeld residuals. Spreading symptoms, N (%) 57/189 (30.2) Future events Multivariable analysis ICH, N (%) 51 (26.8) Ischemic stroke, N (%) 19 (10) We performed a pooled multivariable regression analysis for Recurrent cSAH, N (%) 39 (20.5) each outcome using a frailty term to account for differences Restarted on Antiplatelets 54/171 (31.6) between study cohorts. The multivariable model included Restarted on OAC 13/169 (7.7) the prespecified variables probable CAA and age, as well Death, N (%) 31 (16.3) as variables that had a p value below 0.2 in the univari- Probable CAA​ 153 (80.5) able analysis for each outcome. As a sensitivity analysis we Possible CAA​ 37 (19.5) conducted a competing risk analysis for all of the outcome cSS (%) 188/190(99) events. Due to a strong degree of overlap (collinearity), 0 18 (9.6) we did not include probable CAA and cortical superficial 1 (focal) 55 (29.3) siderosis in the same model. 2 (disseminated) 115 (61.2) Statistical analysis was performed using STATA 15 CMB binary 125/188 (66.5) (StataCorp. 2017. Stata Statistical Software: Release 15. AF atrial fibrillation, CAA​ cerebral amyloid angiopathy, CMB cer- College Station, TX: StataCorp LLC). ebral microbleeds, cSAH cortical subarachnoid hemorrhage cSS cor- tical superficial siderosis, DM diabetes mellitus, FU follow-up, HTN hypertension, ICH intracerebral hemorrhage, OAC oral anticoagula- tion, PMH past medical history, preICH previous intracerebral hem- Results orrhage, preIS previous ischemic stroke, SD standard deviation a In case of missing values in the predictors the number is displayed as We identified 190 eligible patients with cSAH in 13 a fraction and percentage is based on complete cases cohorts published or recruited between 2001 and 2018 with 385 patient-years of follow-up (median follow-up time 1.4 years). The mean age was 74.5 (SD 8.8) and 86 (45.3%) data. 153 patients (80.5%) fulfilled the modified Boston cri- were female. See Table 1 for the overall baseline charac- teria for probable CAA and 37 (19.5%) for possible CAA. teristics and supplementary Table e-2 for individual cohort 13 1430 Journal of Neurology (2022) 269:1427–1438 The characteristics of patients by CAA status are shown in Risk of recurrent cSAH during follow‑up supplementary Table e-3. We observed 39 recurrent cSAH over 351 patient-years, an absolute event rate per patient-year of 11.1% (95% CI Risk of outcome events 7.9–15.2). The median time to recurrent cSAH was 1.3 years (IQR 3.3). The event rate was 11.9% (95% CI 8.3–16.5) for Over the full period of follow-up, the frequency of events probable CAA and 7.1% (95% CI 1.9–18.3) for possible was as follows: 51 (26.8%) had an ICH, 39 (20.5%) a recur- CAA (p = 0.2; Table 3). See Fig. 2B for the Kaplan Meier rent cSAH, and 19 (10%) an ischemic stroke; 31 (16.3%) (KM) survival estimates according to probable CAA diag- died. Figure 1 demonstrates the event rate over time adjust- nosis. Patients with recurrent cSAH were more often male, ing for censoring. Most outcome events occurred in par- and more often had probable CAA (Table 3). In the multi- ticipants with probable CAA; the proportions of patients variable Cox regression model, adjusted for the prespeci- fulfilling these criteria were: 50/51 (98%) patients with ICH; fied variables probable CAA, age, sex and previous ischemic 35/39 (89.7%) patients with recurrent cSAH, and 15/19 stroke and anticoagulation (Table 3), there was a higher risk (79%) with ischemic stroke. of recurrent cSAH for probable CAA patients, but this was not significant at the 5% level (HR 3.66, 95% CI 0.84–15.9, p = 0.08). OAC use was not associated with the risk of recur- Risk of symptomatic ICH during follow‑up rent cSAH (HR 1.65, 95% CI 0.51–5.35). We observed 51 ICH over 385 patient-years, an absolute Risk of any intracranial haemorrhage during follow‑up event rate of 13.2% (95% CI 9.9–17.4) per patient-year; the median time to ICH was 1.4 years (IQR 3.4), while 21.6% The risk of any intracranial haemorrhage (recurrent ICH, occurred in the first month (Fig. 1). The ICH rate for patients cSAH, or both) per patient-year was 21.4% (95% CI with probable CAA was 15.2% (95% CI 11.3–20) per 16.7–26.9); 17 patients had both recurrent ICH and cSAH. patient-year, compared to 1.8% (95% CI 0.1–9.9) for those without probable CAA (p = 0.023; Table 2); see Fig. 2A Risk of ischemic stroke on follow‑up for the Kaplan Meier (KM) survival estimates according to probable CAA diagnosis. Multivariable Cox regression, We observed 19 symptomatic ischemic stroke events over including the pre-specified variables probable CAA and 373 patients-years, an absolute event rate per patient-year of age as well as anticoagulation, confirmed that patients with 5.1% (95% CI 3.1–8). The median time to ischemic stroke probable CAA had a higher risk of ICH than those without was 1.3 years (IQR 3.3). The event rate for patients with (HR 8.45, 95% CI 1.13–75.5, p = 0.02; Table 2). Despite the probable CAA was 4.7% (95% CI 2.6–7.7) and for patients small HR, we did not find any evidence of association of with possible CAA 7.5% (95% CI 2–19.2; p = 0.32; Table 4, OAC with ICH (HR 0.21; 95% CI 0.04–2.06). Fig. 2C). Patients with ischemic stroke on follow-up were Fig. 1  Kaplan Maier survival estimates for all outcome events in the whole cohort 13 Journal of Neurology (2022) 269:1427–1438 1431 Table 2  Predictors of ICH during follow-up for the full cohort No ICH on FU ICH on FU Univariable analysis Multivariable analysis N = 139 N = 51 HR 95% CI p value HR 95% CI p value Age, mean (SD), N (%) 74.4 (9.2) 75 (7.5) 1 0.96–1.04 0.96 1.01 0.97–1.05 0.76 Female sex, N (%) 62 (44.6) 24 (47.1) 1.14 0.65–1.99 0.66 Current smoker 12/122a (9.8) 5/47 (10.6) 1.69 0.64–4.5 0.29 Current drinker 16/109 (14.7) 6/40 (15) 1.85 0.73–4.69 0.2 PMH HTN, N (%) 83 (59.7) 34 (66.7) 1.08 0.6–1.96 0.8 Hypercholesterolemia, N (%) 59 (42.5) 26 (51) 0.99 0.55–1.76 0.96 DM, N (%) 17 (12.2) 9 (17.8) 1.1 0.51–2.36 0.82 AF, N (%) 15/135 (11.1) 3 (5.9) 0.57 0.17–1.9 0.36 OAC use, N (%) 14/135 (10.4) 1/50 (2) 0.25 0.03–1.85 0.17 0.21 0.04–2.06 0.21 Antiplatelet use, N (%) 46/135 (34.1) 17/50 (34) 0.84 0.44–1.59 0.59 Statin use, N (%) 48/135 (35.6) 18/50 (36) 0.84 0.46–1.51 0.55 Anti HTN use, N (%) 74/135 (54.8) 30/50 (60) 1.09 0.61–1.96 0.77 Previous ICH, N (%) 16 (11.5) 8 (15.7) 1.03 0.46–2.29 0.95 Previous IS, N (%) 17 (12.3) 4 (7.8) 0.6 0.21–1.7 0.34 Restart antiplatelet, N (%) 38/122 (31.2) 16 (32.7) 0.87 0.46–1.66 0.68 Restart OAC, N (%) 9/120 (7.5) 4 (8.2) 0.86 0.3–2.49 0.79 Probable CAA, N (%) 103 (74.1) 50 (98) 10.26 1.37–76.77 0.023 8.45 1.13–75.5 0.02 cSS, N (%) 123/138 (89.1) 47/50 (94) 1.37 0.41–4.63 0.61 CMB, N (%) 85/138 (61.6) 40/49 (81.6) 1.32 0.6–2.9 0.49 WMH, N (%) 87/128 (68) 33/48 (68.8) 0.88 0.44–1.77 0.73 Significant p values are marked in bold AF atrial fibrillation, CAA​ cerebral amyloid angiopathy, CI confidence interval, CMB cerebral microbleeds, cSS cortical superficial siderosis, DM diabetes mellitus, DWI diffused white matter, FU follow-up, HR Hazard Ratio, HTN hypertension, ICH intracerebral hemorrhage, OAC oral anticoagulation, PMH past medical history, preICH previous intracerebral hemorrhage, preIS previous ischemic stroke, SD standard deviation, WMH white matter hyperdensity a In case of missing values in the predictors the number is displayed as a fraction and percentage is based on complete cases more often male, more frequently had a history of hyper- variables age and hypertension (Table 5), probable tension, were on anticoagulation, and had probable CAA CAA was not associated with death (HR 0.54, 95% CI (Table 4). In our multivariable model, adjusted with the pre- 0.16–1.78, p = 0.31). In a sensitivity analysis, adding specified variables age and anticoagulation (Table 4), prob- smoking status to the multivariable model did not change able CAA was not associated with a higher risk of ischemic the findings (results not shown). stroke (HR 0.56, 95% CI 0.17–1.82, p = 0.33). When comparing patients with ICH versus ischemic stroke on follow-up, patients suffering an ICH had a higher Risk of death during follow‑up mortality rate than those who had an ischemic stroke (45.5% vs 20.5%). The proportional hazard assumption was not We observed 31 deaths over 373 patient-years, an overall violated for any of the outcomes. As a sensitivity analysis, absolute event rate of 8.3% (95%CI 5.6–11.8) per patient- we conducted a competing risk analysis for all of the out- year. The median time to death was 1.2 years (IQR 3.3). come events, which demonstrated similar results (results not There was no statistically significant difference in mortal- shown). ity between patients with probable CAA [7.3% for patients with probable CAA (95% CI 4.6–10.9)] and those with- out [14.2% for patients without probable CAA (95% CI Influence of symptoms at presentation on outcome events 6.1–28); p = 0.23; Table 5, Fig. 2D]. Patients who died were older, and more often had hypertension, and were In a final step, we adjusted for symptoms on presentation on antihypertensive medication more often (Table 5). In (positive, negative or spreading symptoms) in the univari- our multivariable model, adjusted for the pre-specified able and multivariable model; clinical presentation did 13 1432 Journal of Neurology (2022) 269:1427–1438 Fig. 2  Kaplan Maier survival estimates based on outcome and probable CAA status: A ICH; B recurrent cSAH; C ischemic stroke; and D death not influence the risks of stroke or death during follow-up on antiplatelets and 13/169 (7.7%) on oral anticoagulants. Nei- (results not shown). ther antiplatelet nor anticoagulant use after cSAH were associ- ated with any of the outcome events. Influence of age on outcome events In a post hoc analysis we included an age term dichotomized Discussion at age 70 in the Cox models (based on a previous publication which used this cutoff) [15]; there were no significant differ- Our pooled individual patient-data analysis confirms that ences between age > 70 and < 70 in any of our outcomes of patients who had a cSAH due to suspected CAA have a interest (data not shown). high rate of future intracranial haemorrhage (both ICH and cSAH), with an overall intracranial haemorrhage Use of antithrombotic drugs during follow‑up risk (ICH, cSAH, or both) of 21.4% per patient-year. The rates were highest within the first month after cSAH, dur- Data on starting or restarting antiplatelets and oral anticoagu- ing which 21.6% of ICH and 12.8% of cSAH occurred. lants were available for 171/190 (90%) and 169/190 (89%) The rate of ischemic stroke was much lower at 5.1% per patients, respectively; 54/171 (31.6%) were started or restarted patient-year, while mortality was 8.3% per patient-year. 13 Journal of Neurology (2022) 269:1427–1438 1433 Table 3  Predictors of recurrent cSAH during follow-up No recSAH FU recSAH FU Univariable analysis Multivariable analysis N = 151 N = 39 HR 95% CI p value HR 95% CI p value Age, mean (SD) 74.7 (9.2) 73.9 (7.1) 0.97 0.94–1.01 0.19 0.98 0.94–1.02 0.24 Female sex, N (%) 74 (49) 12 (30.8) 0.47 0.23–0.95 0.04 0.42 0.19–0.9 0.03 Smoker 15/134a (11.2) 2/35 (5.7) 0.57 0.13–2.47 0.45 Drinker 19/116 (16.4) 3/3 (9.1) 0.6 0.18–2.04 0.41 PMH HTN, N (%) 92 (60.9) 25 (64.1) 0.94 0.48–1.86 0.86 Hypercholesterolemia, N (%) 67 (44.4) 18 (46.2) 1.3 0.67–2.52 0.44 DM, N (%) 18 (11.9) 8 (20.5) 1.67 0.74–3.78 0.22 AF, N (%) 13/149 (8.7) 5/37 (13.5) 1.66 0.61–453 0.32 OAC use, N (%) 11/148 (7.4) 4/37 (10.8) 1.87 0.62–5.61 0.26 1.65 0.51–5.35 0.41 Antiplatelet use, N (%) 50/48 (33.8) 13/37 (35.1) 0.81 0.39–1.65 0.56 Statin use, N (%) 52/148 (35.1) 14/37 (37.8) 1.18 0.59–2.36 0.63 Anti HTN use, N (%) 84/148 (56.8) 20/37 (54.1) 0.79 0.39–1.56 0.49 Previous ICH, N (%) 20 (13.3) 4 (10.3) 0.81 0.28–2.38 0.71 Previous IS, N (%) 19/150 (12.7) 2 (5.1) 0.29 0.07–1.25 0.1 0.26 0.06–1.14 0.08 Restart antiplatelet, N (%) 43/134 (32.1) 11/37 (29.7) 0.89 0.42–1.91 0.89 Restart OAC, N (%) 12/133 (7.5) 3/36 (8.3) 0.92 0.27–3.15 0.92 Probable CAA, N (%) 118 (78.2) 35 (89.7) 2.07 0.69–6.23 0.2 3.66 0.84–15.9 0.08 cSS, N (%) 132/149 (88.6) 38 (97.4) 5.05 0.67–38.07 0.12 CMB, N (%) 96/148 (64.9) 29 (74.4) 1.2 0.54–2.65 0.65 WMH, N (%) 94/137 (68.6) 26 (66.7) 0.67 0.31–1.44 0.3 AF atrial fibrillation, CAA​ cerebral amyloid angiopathy, CI confidence interval, CMB cerebral microbleeds, recSAH recurrent cortical subarach- noid hemorrhage, cSS cortical superficial siderosis, DM diabetes mellitus, DWI diffused white matter, FU follow-up, HR Hazard Ratio, HTN hypertension, OAC oral anticoagulation, PMH past medical history, preICH previous intracerebral hemorrhage, preIS previous ischemic stroke, SD standard deviation, WMH white matter hyperdensity a In case of missing values in the predictors the number is displayed as a fraction and percentage is based on complete cases In adjusted analyses, we found that the rate of future ICH 15 (39%) experienced recurrent cSAHs and 14 (37%) suf- was around 8.5 times higher in patients fulfilling the modi- fered lobar ICHs; notably, of 22 new ICHs, 17 occurred fied Boston criteria for probable CAA compared to those at sites of previous cSAHs or cSS [14]. Moreover, in four without probable CAA. patients, imaging demonstrated early meningeal enhance- Our study expands on previous smaller studies report- ment, suggesting active vascular leakage associated with ing the rates of ICH and recurrent cSAH in patients pre- leptomeningeal CAA, and rapid expansion of cSAH into the senting with cSAH and emphasizes that presentation with parenchyma causing an ICH [19]. These data, together with cSAH can be a sentinel event warning of impending ICH. our study, are consistent with the idea that cSAH—presum- Due to larger patient numbers and individual patient-data, ably associated with severe leptomeningeal CAA—could in we provide more precise estimates of the high future ICH some instances be an initial step in ICH formation. risk in cSAH patients fulfilling the modified Boston criteria Our study also provides new information on the future for probable and possible CAA [31, 32]; we found that the rate of ischemic stroke after cSAH [12], confirming that the ICH rate for patients with probable CAA was 15.2% (95% risk of ischemic stroke is much lower than that of recurrent CI 11.3–20) per patient-year, compared to 1.8% (95% CI ICH or cSAH. 0.001–9.9) for those without (p = 0.023). Patients fulfilling Such information on hemorrhagic and ischemic cerebral the modified Boston criteria for probable CAA had a higher events is important for informing management decisions, rate of recurrent cSAH on follow-up, but this finding was not particularly regarding antithrombotic therapy. Indeed, we statistically significant. found that 39.5% of our cohort were taking antithrombotic The high rates of ICH and recurrent cSAH are also drugs at the time of the cSAH event, suggesting that whether consistent with another previous study of 38 patients with to restart such drugs is a common clinical dilemma after cSAH, over a mean of 24 months of follow-up, during which cSAH. In our study, only 31.6% were started or restarted on 13 1434 Journal of Neurology (2022) 269:1427–1438 Table 4  Predictors of ischemic stroke during follow-up No IS on FU IS on FU Univariable analysis Multivariable analysis N = 171 N = 19 HR 95% CI p value HR 95% CI p value Age, mean (SD) 74.5 (9) 74.6 (6.5) 1.01 0.95–1.07 0.8 1 0.94–1.07 0.9 Female sex, N (%) 79 (46.2) 7 (36.8) 0.72 0.28–1.85 0.5 Smoker 15/150a (10) 2 (10.5) 1.25 0.28–5.56 0.77 Drinker 19/131 (14.5) 3/18 (16.7) 1.83 0.5–6.7 0.36 PMH HTN, N (%) 103 (60.2) 14 (73.7) 1.71 0.61–4.81 0.31 Hypercholesterolemia, N (%) 78 (45.6) 7 (36.8) 0.71 0.27–1.88 0.49 DM, N (%) 22 (12.9) 4 (21.1) 1.71 0.55–5.34 0.36 AF, N (%) 15/2167 (9) 3 (15.8) 2 0.55–7.33 0.3 OAC use, N (%) 11/166 (6.6) 4 (21.1) 3.63 1.13–11.68 0.03 3.44 1.07–11.06 0.04 Antiplatelet use, N (%) 56/166 (33.7) 7 (36.8) 1.06 0.4–2.77 0.91 Statin use, N (%) 60/166 (36.1) 6 (31.6) 0.86 0.32–2.32 0.77 Anti HTN use, N (%) 93/166 (56) 11 (5791) 1.02 0.4–2.6 0.97 Previous ICH, N (%) 23 (13.5) 1 (5.3) 0.36 0.05–2.71 0.32 Previous IS, N (%) 18/170 (10.6) 3 (15.8) 1.75 0.49–6.21 0.39 Restart antiplatelet, N (%) 48/152 (31.6) 6 (31.6) 1.14 0.42–3.13 0.79 Restart OAC, N (%) 12/150 (8) 1 (5.3) 0.53 0.07–4.07 0.54 Probable CAA, N (%) 138 (80.7) 15 (78.9) 0.55 0.16–1.81 0.32 0.56 0.17–1.82 0.33 cSS, N (%) 155/169 (91.7) 15 (79) 0.39 0.12–1.24 0.11 CMB, N (%) 110/168 (65.5) 15 (79) 1.15 0.36–3.67 0.82 WMH, N (%) 107/157 (68.2) 13 (68.4) 0.73 0.26–2.06 0.56 AF atrial fibrillation, CAA​cerebral amyloid angiopathy, CI confidence interval, CMB cerebral microbleeds, cSS cortical superficial siderosis, DM diabetes mellitus, DWI diffused white matter, FU follow-up, HR Hazard Ratio, HTN hypertension, OAC oral anticoagulation, PMH past medical history, preICH previous intracerebral hemorrhage, preIS previous ischemic stroke, SD standard deviation, WMH white matter hyperdensity a In case of missing values in the predictors the number is displayed as a fraction and percentage is based on complete cases antiplatelets and 7.7% on anticoagulation, suggesting there status—might favor avoiding antithrombotics for patients is anxiety about future intracranial bleeding risk in clinical with cSAH who meet the modified Boston criteria for prob- practice. Although antithrombotic therapy (antiplatelets or able CAA. There are no other observational data or rand- anticoagulants) could reduce the risk of ischemic events, omized controlled trials addressing the risks of (re)starting they might increase the risk of future ICH or cSAH. Our antithrombotic drugs after cSAH. Data on whether anti- study did not find evidence that restarting antithrombotic platelet drugs affect recurrent ICH risk after a symptomatic therapy after cSAH increases the rate of intracranial bleed- ICH are conflicting: while two small single center studies ing, but this observation is likely to be affected by bias and gave inconsistent data on whether the rate of recurrent ICH confounding including physician decisions to restart these was increased by the use of aspirin [34, 35], the RESTART agents. randomized trial in ICH survivors found no evidence of an Despite the much higher risks of ICH and cSAH in people increase in the risk of recurrent intracerebral haemorrhage with probable vs possible CAA, we did not find evidence of with antiplatelet therapy for patients on antithrombotic ther- a higher mortality rate in this group, although the rate in our apy for the prevention of occlusive vascular disease when cohort was comparable to previous reports [33]. they developed ICH [36]. However, this trial included small We found that patients with cSAH attributed to probable numbers of patients with disseminated superficial sidero- or possible CAA have a much higher annual risk of future sis. We found a high risk of ICH after cSAH despite most intracranial haemorrhage (ICH or cSAH) than ischemic patients not receiving antithrombotic therapy. Thus, in those stroke [13.2% (95% CI 9.9–17.4) and 11.1% (95%CI judged to have a low vaso-occlusive risk starting or restart- 7.9–15.2), respectively, compared with a 5.1% (95% CI ing them might be best avoided. Our finding that about one 3.1–8)]. The substantially higher ICH risk in probable in five ICH occur in the first month raises particular con- compared to possible CAA (15.2% vs 1.8% per year)— cern for antithrombotic therapy during this early period with a similar ischemic stroke risk regardless of CAA following cSAH. Although randomized data are needed to 13 Journal of Neurology (2022) 269:1427–1438 1435 Table 5  Predictors of mortality during follow-up Survived Death Univariable analysis Multivariable analysis N = 159 N = 31 HR 95% CI p value HR 95% CI p value Age, mean (SD) 74 (9) 77.5 (6.7) 1.04 0.99–1.1 0.14 1.04 0.98–1.11 0.21 Female sex, N (%) 74 (46.5) 12 (38.7) 0.89 0.43–1.87 0.76 Smoker 12/139a (8.6) 5/30 (16.7) 2.47 0.88–6.9 0.09 Drinker 18/125 (14.4) 4/24 (16.7) 1.68 0.54–5.25 0.37 PMH HTN, N (%) 93 (58.5) 24 (77.4) 1.82 0.77–4.27 0.17 2.27 1.24–4.17 0.008 Hypercholesterolemia, N (%) 70 (44) 15 (48.4) 1.06 0.51–2.24 0.87 DM, N (%) 20 (12.6) 6 (19.4) 1.3 0.51–3.31 0.58 AF, N (%) 13/155 (8.4) 5 (16.1) 1.54 0.56–4.25 0.41 OAC use, N (%) 12/155 (7.7) 3/30 (10) 1.26 0.36–4.42 0.71 Antiplatelet use, N (%) 55/155 (35.5) 8/30 (26.7) 0.66 0.28–1.55 0.34 Statin use, N (%) 56/155 (36.1) 10/30 (33.3) 0.8 0.36–1.75 0.57 Anti HTN, N (%) 83/155 (53.6) 21/30 (70) 1.57 0.7–3.53 0.28 Previous ICH, N (%) 20 (12.6) 4 (12.9) 0.96 0.32–2.85 0.94 sPrevious IS, N (%) 16 (10.1) 5 (16.1) 1.17 0.43–3.15 0.76 Restart antiplatelet, N (%) 46/140 (32.9) 8 (25.8) 0.76 0.32–1.77 0.52 Restart OAC, N (%) 9/138 (6.5) 4 (12.9) 1.2 0.4–3.56 0.75 Probable CAA, N (%) 130 (81.8) 23 (74.2) 0.57 0.23–1.42 0.23 0.54 0.16–1.78 0.31 cSS, N (%) 146/158 (92.4) 24/30 (80) 0.31 0.12–0.81 0.02 CMB, N (%) 104/158 (65.8) 21/29 (72.4) 0.94 0.38–2.33 0.63 WMH, N (%) 100/146 (68.5) 20/30 (66.7) 0.6 0.25–1.43 0.25 Significant p values is marked in bold AF atrial fibrillation, CAA​cerebral amyloid angiopathy, CI confidence interval, CMB cerebral microbleeds, cSS cortical superficial siderosis, DM diabetes mellitus, DWI diffused white matter, FU follow-up, HR Hazard Ratio, HTN hypertension, OAC oral anticoagulation, PMH past medical history, preICH previous intracerebral hemorrhage, preIS previous ischemic stroke, SD standard deviation, WMH white matter hyperdensity a In case of missing values in the predictors the number is displayed as a fraction and percentage is based on complete cases inform antithrombotic decisions after cSAH, the rarity of different centers could not be tested. However, uniform cSAH and lack of clinician equipoise might make such tri- definitions of CAA were applied by trained raters as per als challenging. our predefined study protocol. Also, our rate of spreading Our pooled analysis approach has several strengths. We symptoms was surprisingly low which we think to be due to have included the largest number of patients with cSAH the retrospective nature of the study and reflects an under- associated with probable or possible CAA studied to date, estimation. This needs to be verified in an independent large with a long follow-up period, allowing us to provide more cohort. precise estimates of the rates of future ICH, cSAH, as well as ischemic stroke after a previous cSAH, including data on the longer-term time course of these events. The Conclusions inclusion of detailed individual patient data additionally allowed us to evaluate multivariable models. Reassuringly, We confirm that patients with cSAH are at high risk of future in our pooled analysis, the risk of bias was low, so any loss ICH and recurrent cSAH (13.2% and 11.1% per patient-year, to follow-up is most likely due to random (non-informa- respectively), with a high early risk. By contrast, the risk of tive) censoring. ischemic stroke is much lower (5.1% per patient-year). Prob- However, our study also has limitations. This was a ret- able vs possible CAA is a predictor of higher ICH but not rospective study, although we only included studies where ischemic stroke risk. Our data provide precise risk estimates investigators had systematically and prospectively collected of key vascular outcomes after cSAH that can help inform data on the follow-up events of interest as defined in the prognosis and management decisions after cSAH. protocol. Additionally, imaging was not reviewed centrally. This is a potential source of bias as agreement between the 13 1436 Journal of Neurology (2022) 269:1427–1438 Supplementary Information The online version contains supplemen- Availability of data and material Anonymized data will be shared on tary material available at https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 00415-0​ 21-1​ 0706-3. request from any qualified investigator, subject to approval of the par- ticipating collaborators. Authors’ contributions ICH received funding from the Alzheimer Research UK and Dunhill Medical Trust Foundation. DW received Code availability Not applicable. funding from the Stroke Foundation/British Heart Foundation. CAF reports no disclosures. DA reports no disclosures. SFA reports no dis- closures. FE reports no disclosures. MB reports no disclosures. TK Declarations reports no disclosures. TP has received Honorarium as speaker from Genzyme, Boehringer Ingelheim, Pfizer, BMS, Bayer. Dr. P is also on Conflicts of interest The authors declare that they have no conflict of the Advisory Board for Genzyme on Fabry disease. SM reports no dis- interest. closures. RS reports no disclosures. MCI reports grants from Instituto de Salud Carlos III, grants from the Alzheimer’s Association and the Ethical approval All included studies were approved in accordance Global Brain Health Institute, grants from Jérome Lejeune Foundation, with local regulatory and ethical approvals. Only fully anonymized grants from Societat Catalana de Neurologia, and non-financial support data were shared. Where necessary, a data use agreement was signed. from Fundació Catalana Síndrome de Down during the conduct of the study. AC reports no disclosures. IM reports no disclosures. SP reports Consent to participate Not applicable. no disclosures. GA reports no disclosures. HRJ reports no disclosures. SS reports no disclosures. JL reports no disclosures. HM reports no Consent for publication Not applicable. disclosures. ET reports no disclosures. RG reports no disclosures. ACF reports no disclosures. TM reports no disclosures. PL reports no dis- Open Access This article is licensed under a Creative Commons Attri- closures. PHL reports no disclosures. AV reports no disclosures. SMG bution 4.0 International License, which permits use, sharing, adapta- is a consultant for Hoffmann-LaRoche and GlaxoSmithKline. Dr. G tion, distribution and reproduction in any medium or format, as long is on the data monitoring committee of Biogen and the Data Safety as you give appropriate credit to the original author(s) and the source, Monitoring Board at Washington University/IQVIA. He is on the Edi- provide a link to the Creative Commons licence, and indicate if changes torial Board of following journals: Stroke, Cerebrovascular Disease, were made. The images or other third party material in this article are Neurology and American Journal of Alzheimer’s Disease and Other included in the article’s Creative Commons licence, unless indicated dementias. He is additionally a Section Editor for Stroke. Dr. Green- otherwise in a credit line to the material. If material is not included in berg receives royalties from UpToDate and is funded by following the article’s Creative Commons licence and your intended use is not NIH grants: #R01NS096730, #R01AG026484, #U24 NS107154, #U24 permitted by statutory regulation or exceeds the permitted use, you will NS100591. JF reports grants from Instituto de Salud Carlos III, grants need to obtain permission directly from the copyright holder. To view a from Fundació Marató TV3, grants from NIH, grants from Depar- copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. tament de Salut de la Generalitat de Catalunya, grants from Jérome Lejeune Foundation, grants from Fundació Víctor Grífols i Lucas, and non-financial support from Fundació Catalana Síndrome de Down dur- ing the conduct of the study; personal fees from AC Inmune Clinical Advisory Board, personal fees from Novartis Adjudication Commit- tee, personal fees from Lundbeck Advisory Board, personal fees from References Esteve Conference fees, personal fees from NovoNordisk Conference fees, personal fees from Roche Conference fees and advisory board, 1. Beitzke M, Gattringer T, Enzinger C, Wagner G, Niederkorn K, personal fees from Fujirebio Conference fees, personal fees from Bio- Fazekas F (2011) Clinical presentation, etiology, and long-term gen Conference fees, and personal fees from Merck Adjudication Com- prognosis in patients with nontraumatic convexal subarachnoid mittee outside the submitted work; in addition, Dr. Fortea has a patent haemorrhage. Stroke; J Cereb Circ 42(11):3055–3060. https://d​ oi.​ for Markers of synaptopathy in neurodegenerative disease licensed org/​10.​1161/​STROK​EAHA.​111.​621847 EP18382175.0. MA reports no disclosures. MB reports no disclosures. 2. Hostettler IC, Werring DJ (2018) Acute convexity subarachnoid FV reports no disclosures. SK reports no disclosures. VS reports no haemorrhage: what the neurosurgeon needs to know. World Neu- disclosures. AK reports no disclosures. FF reports no disclosures. VB rosurg. https://​doi.​org/​10.​1016/j.​wneu.​2018.​12.​062 reports no disclosures. GJZ reports no disclosures. DR reports no dis- 3. Ni J, Auriel E, Jindal J, Ayres A, Schwab KM, Martinez-Ramirez closures. AR is a member of the external committee for adverse event S, Gurol EM, Greenberg SM, Viswanathan A (2015) The char- adjudication for PREVAIL trial/CAP2 registry. Dr. R is on the edito- acteristics of superficial siderosis and convexity subarachnoid rial board Neurology, Neurocritical Care, Stroke and CONTINUUM. haemorrhage and clinical relevance in suspected cerebral amyloid He receives royalties for: Practical Neuroimaging in Stroke—Else- angiopathy. Cerebrovasc Dis 39(5–6):278–286. https://d​ oi.o​ rg/1​ 0.​ vier—2009 What To Do? Neurocritical Care—Oxford—2016. JG-R 1159/​00038​1223 is funded by an NIH grant: #K76 AG057015-01. DJW received funding 4. Kumar S, Goddeau RP Jr, Selim MH, Thomas A, Schlaug G, from the Stroke Foundation/British Heart Foundation. Dr. W reports Alhazzani A, Searls DE, Caplan LR (2010) Atraumatic convexal personal fees from Bayer, Alnylam and Portola. subarachnoid haemorrhage: clinical presentation, imaging pat- terns, and etiologies. Neurology 74(11):893–899. https://​doi.​org/​ 10.​1212/​WNL.​0b013​e3181​d55efa Funding This study was not funded and did not receive industry spon- 5. Stanton JED, Chandratheva A, Wilson D, Hostettler IC, Islam S, soring. ICH received funding from the Alzheimer Research UK and Werring DJ (2020) Clinical features distinguish cerebral amyloid Dunhill Medical Trust Foundation. DJW and DW receive funding from angiopathy-associated convexity subarachnoid haemorrhage from the Stroke Foundation/British Heart Foundation. This work was under- suspected TIA. J Neurol 267(1):133–137. https://d​ oi.o​ rg/1​ 0.1​ 007/​ taken at UCLH/UCL which receives a proportion of funding from the s00415-​019-​09558-9 Department of Health’s National Institute for Health Research (NIHR) 6. Raposo N, Viguier A, Cuvinciuc V, Calviere L, Cognard C, Bon- Biomedical Research Centres funding scheme. neville F, Larrue V (2011) Cortical subarachnoid haemorrhage in the elderly: a recurrent event probably related to cerebral amyloid 13 Journal of Neurology (2022) 269:1427–1438 1437 angiopathy. Eur J Neurol 18(4):597–603. https://d​ oi.o​ rg/1​ 0.1​ 111/j.​ clinical presentation, imaging patterns and etiologic findings in 1468-​1331.​2010.​03214.x 23 patients. Revue neurologique 169(1):59–66. https://d​ oi.o​ rg/1​ 0.​ 7. Vinters HV (1987) Cerebral amyloid angiopathy. A critical review. 1016/j.​neurol.​2012.​02.​012 Stroke; J Cereb Circ 18(2):311–324 22. Refai D, Botros JA, Strom RG, Derdeyn CP, Sharma A, Zipfel 8. Yamada M (2000) Cerebral amyloid angiopathy: an overview. GJ (2008) Spontaneous isolated convexity subarachnoid haemor- Neuropathology 20(1):8–22 rhage: presentation, radiological findings, differential diagnosis, 9. Charidimou A, Gang Q, Werring DJ (2012) Sporadic cerebral and clinical course. J Neurosurg 109(6):1034–1041. https://​doi.​ amyloid angiopathy revisited: recent insights into pathophysiology org/​10.​3171/​JNS.​2008.​109.​12.​1034 and clinical spectrum. J Neurol Neurosurg Psychiatry 83(2):124– 23. Linn J, Halpin A, Demaerel P, Ruhland J, Giese AD, Dichgans 137. https://​doi.​org/​10.​1136/​jnnp-​2011-​301308 M, van Buchem MA, Bruckmann H, Greenberg SM (2010) Preva- 10. Viswanathan A, Greenberg SM (2011) Cerebral amyloid angiopa- lence of superficial siderosis in patients with cerebral amyloid thy in the elderly. Ann Neurol 70(6):871–880. https://​doi.​org/​10.​ angiopathy. Neurology 74(17):1346–1350. https://​doi.​org/​10.​ 1002/​ana.​22516 1212/​WNL.​0b013​e3181​dad605 11. Wilson D, Hostettler IC, Ambler G, Banerjee G, Jager HR, Wer- 24. Ducros A, Boukobza M, Porcher R, Sarov M, Valade D, Bousser ring DJ (2017) Convexity subarachnoid haemorrhage has a high MG (2007) The clinical and radiological spectrum of reversible risk of intracerebral haemorrhage in suspected cerebral amyloid cerebral vasoconstriction syndrome. A prospective series of 67 angiopathy. J Neurol 264(4):664–673. https://​doi.​org/​10.​1007/​ patients. Brain: J Neurology 130(Pt 12):3091–3101. https://​doi.​ s00415-​017-​8398-y org/​10.​1093/​brain/​awm256 12. Wilson D, Hostettler IC, Ambler G, Banerjee G, Jager HR, Wer- 25. Moher D, Liberati A, Tetzlaff J, Altman DG, Group P (2009) Pre- ring DJ (2017) Convexity subarachnoid haemorrhage has a high ferred reporting items for systematic reviews and meta-analyses: risk of intracerebral haemorrhage in suspected cerebral amyloid the PRISMA statement. BMJ 339:b2535. https://​doi.​org/​10.​1136/​ angiopathy. J Neurol. https://​doi.​org/​10.​1007/​s00415-​017-​8398-y bmj.​b2535 13. Apoil M, Cogez J, Dubuc L, Bataille M, de la Sayette V, Touze 26. von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, E, Viader F (2013) Focal cortical subarachnoid haemorrhage Vandenbroucke JP, Initiative S (2007) The strengthening the revealed by recurrent paresthesias: a clinico-radiological syn- reporting of observational studies in epidemiology (STROBE) drome strongly associated with cerebral amyloid angiopathy. statement: guidelines for reporting observational studies. Lancet Cerebrovasc Dis 36(2):139–144. https://​doi.​org/​10.​1159/​00035​ 370(9596):1453–1457. https://​doi.​org/​10.​1016/​S0140-​6736(07)​ 3676 61602-X 14. Beitzke M, Enzinger C, Wunsch G, Asslaber M, Gattringer T, 27. van Swieten JC, Hijdra A, Koudstaal PJ, van Gijn J (1990) Grad- Fazekas F (2015) Contribution of convexal subarachnoid haem- ing white matter lesions on CT and MRI: a simple scale. J Neurol orrhage to disease progression in cerebral amyloid angiopathy. Neurosurg Psychiatry 53(12):1080–1083 Stroke; J Cereb Circ 46(6):1533–1540. https://​doi.​org/​10.​1161/​ 28. Gregoire SM, Chaudhary UJ, Brown MM, Yousry TA, Kallis C, STROK​EAHA.​115.​008778 Jager HR, Werring DJ (2009) The microbleed anatomical rating 15. Bruno VA, Lereis VP, Hawkes M, Ameriso SF (2013) Non- scale (MARS): reliability of a tool to map brain microbleeds. Neu- traumatic subarachnoid haemorrhage of the convexity. Curr rology 73(21):1759–1766. https://​doi.​org/​10.​1212/​WNL.​0b013​ Neurol Neurosci Rep 13(4):338. https:// ​ d oi. ​ o rg/ ​ 1 0. ​ 1 007/​ e3181​c34a7d s11910-​013-​0338-3 29. Charidimou A, Linn J, Vernooij MW, Opherk C, Akoudad S, 16. Geraldes R, Sousa PR, Fonseca AC, Falcao F, Canhao P, e Can- Baron JC, Greenberg SM, Jager HR, Werring DJ (2015) Cortical hao P (2014) Nontraumatic convexity subarachnoid haemorrhage: superficial siderosis: detection and clinical significance in cerebral different etiologies and outcomes. J Stroke Cerebrovasc Dis: Off amyloid angiopathy and related conditions. Brain: J Neurol 138(Pt J Natl Stroke Assoc 23(1):e23-30. https://​doi.​org/​10.​1016/j.​jstro​ 8):2126–2139. https://​doi.​org/​10.​1093/​brain/​awv162 kecer​ebrov​asdis.​2013.​08.​005 30. Haacke EM, Xu Y, Cheng YC, Reichenbach JR (2004) Suscepti- 17. Graff-Radford J, Fugate JE, Klaas J, Flemming KD, Brown RD, bility weighted imaging (SWI). Magn Reson Med 52(3):612–618. Rabinstein AA (2016) Distinguishing clinical and radiological https://​doi.​org/​10.​1002/​mrm.​20198 features of non-traumatic convexal subarachnoid haemorrhage. 31. Greenberg SM, Eng JA, Ning M, Smith EE, Rosand J (2004) Eur J Neurol 23(5):839–846. https://​doi.​org/​10.​1111/​ene.​12926 Haemorrhage burden predicts recurrent intracerebral haemorrhage 18. Khurram A, Kleinig T, Leyden J (2014) Clinical associations and after lobar haemorrhage. Stroke; J Cereb Circ 35(6):1415–1420. causes of convexity subarachnoid haemorrhage. Stroke; J Cereb https://​doi.​org/​10.​1161/​01.​STR.​00001​26807.​69758.​0e Circ 45(4):1151–1153. https://​doi.​org/​10.​1161/​STROK​EAHA.​ 32. van Etten ES, Auriel E, Haley KE, Ayres AM, Vashkevich A, 113.​004298 Schwab KM, Rosand J, Viswanathan A, Greenberg SM, Gurol 19. Ly JV, Singhal S, Rowe CC, Kempster P, Bower S, Phan TG ME (2014) Incidence of symptomatic haemorrhage in patients (2015) Convexity subarachnoid haemorrhage with PiB positive with lobar microbleeds. Stroke; J Cereb Circ 45(8):2280–2285. pet scans: clinical features and prognosis. J Neuroimaging: Off J https://​doi.​org/​10.​1161/​STROK​EAHA.​114.​005151 Am Soc Neuroimaging 25(3):420–429. https://​doi.​org/​10.​1111/​ 33. Calviere L, Viguier A, Patsoura S, Rousseau V, Albucher JF, Plan- jon.​12188 ton M, Pariente J, Cognard C, Olivot JM, Bonneville F, Raposo 20. Martinez-Lizana E, Carmona-Iragui M, Alcolea D, Gomez-Choco N (2019) Risk of intracerebral haemorrhage and mortality after M, Vilaplana E, Sanchez-Saudinos MB, Clarimon J, Hernandez- convexity subarachnoid haemorrhage in cerebral amyloid angi- Guillamon M, Munuera J, Gelpi E, Gomez-Anson B, de Juan- opathy. Stroke; J Cereb Circ 50(9):2562–2564. https://​doi.​org/​10.​ Delago M, Delgado-Mederos R, Montaner J, Ois A, Amaro S, 1161/​STROK​EAHA.​119.​026244 Blesa R, Marti-Fabregas J, Lleo A, Fortea J (2015) Cerebral 34. Biffi A, Halpin A, Towfighi A, Gilson A, Busl K, Rost N, Smith amyloid angiopathy-related atraumatic convexal subarachnoid EE, Greenberg MS, Rosand J, Viswanathan A (2010) Aspirin and haemorrhage: an ARIA before the tsunami. J Cereb Blood flow recurrent intracerebral haemorrhage in cerebral amyloid angiop- and metab: Off J Int Soc Cereb Blood Flow Metab 35(5):710–717. athy. Neurology 75(8):693–698. https://​doi.​org/​10.​1212/​WNL.​ https://​doi.​org/​10.​1038/​jcbfm.​2015.​25 0b013​e3181​eee40f 21. Mas J, Bouly S, Mourand I, Renard D, de Champfleur N, 35. Viswanathan A, Rakich SM, Engel C, Snider R, Rosand J, Green- Labauge P (2013) Focal convexal subarachnoid haemorrhage: berg SM, Smith EE (2006) Antiplatelet use after intracerebral 13 1438 Journal of Neurology (2022) 269:1427–1438 haemorrhage. Neurology 66(2):206–209. https://​doi.​org/​10.​1212/​ open-label trial. Lancet 393(10191):2613–2623. https://​doi.​org/​ 01.​wnl.​00001​94267.​09060.​77 10.​1016/​S0140-​6736(19)​30840-2 36. Collaboration R (2019) Effects of antiplatelet therapy after stroke due to intracerebral haemorrhage (RESTART): a randomised, Authors and Affiliations Isabel Charlotte Hostettler1 · Duncan Wilson1 · Catherine Arnold Fiebelkorn2 · Diane Aum3 · Sebastián Francisco Ameriso4 · Federico Eberbach4 · Markus Beitzke5 · Timothy Kleinig6 · Thanh Phan7,21 · Sarah Marchina8 · Romain Schneckenburger9 · Maria Carmona‑Iragui10 · Andreas Charidimou11 · Isabelle Mourand12 · Sara Parreira13 · Gareth Ambler14 · Hans Rolf Jäger15 · Shaloo Singhal7,21 · John Ly7,21 · Henry Ma7,21 · Emmanuel Touzé16 · Ruth Geraldes17,18 · Ana Catarina Fonseca13 · Teresa Melo13 · Pierre Labauge12 · Pierre‑Henry Lefèvre19 · Anand Viswanathan11 · Steven Mark Greenberg11 · Juan Fortea10 · Marion Apoil9 · Marion Boulanger9,16 · Fausto Viader9 · Sandeep Kumar8 · Velandai Srikanth7,21 · Ashan Khurram6 · Franz Fazekas5 · Veronica Bruno4 · Gregory Joseph Zipfel3 · Daniel Refai20 · Alejandro Rabinstein2 · Jonathan Graff‑Radford2 · David John Werring1 1 12 Stroke Research Centre, University College London, Department of Neurology, CHU de Montpellier, Hôpital National Hospital of Neurology and Neurosurgery, Institute Gui-de-Chauliac, Montpellier, France of Neurology, Queen Square, London WC1N, UK 13 Stroke Unit, Department of Neuroscience, Hospital de Santa 2 Department of Neurology, Mayo Clinic, Rochester, MN, Maria, University of Lisbon, Lisbon, Portugal USA 14 Department of Statistical Science, UCL, London WC1E 6BT, 3 Department of Neurological Surgery, Washington University UK School of Medicine, St. Louis, MO, USA 15 Neuroradiological Academic Unit, Department of Brain 4 Institute for Neurological Research, Fleni, Buenos Aires, Repair & Rehabilitation, University College London, Argentina Institute of Neurology, London, UK 5 16 Department of Neurology, Medical University of Graz, Graz, Normandy University, UNICAEN, INSERM U1237, Caen, Austria France 6 17 Department of Neurology, Royal Adelaide Hospital, Nuffield Department of Clinical Neurosciences, Oxford Adelaide, Australia University Hospitals, Oxford, UK 7 18 Department of Neurology, Monash Health and Stroke Neurology department, Frimley Health Foundation Trust, and Ageing Research Group, Melbourne, Australia Camberley, UK 8 19 Department of Neurology, Stroke Division, Beth Israel Department of Neuroradiology, CHU de Montpellier, Hôpital Deaconess Medical Center, Harvard Medical School, Boston, Gui-de-Chauliac, Montpellier, France MA, USA 20 Department of Neurosurgery, Emory University, Atlanta, 9 Department of Neurology, CHU Caen Normandie, Caen, GA, USA France 21 Department of Medicine, School of Clinical Sciences, 10 Memory Unit, Department of Neurology, Hospital de la Monash University, Melbourne, Australia Santa Creu I Sant Pau, Institut Investigació Biomèdica Sant Pau, Universitat Autònoma de Barcelona, Barcelona, Spain 11 J. Philip Kistler Stroke Research Center, Department of Neurology, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA 13