Distinguishing clinical and radiological features of non-traumatic convexal subarachnoid hemorrhage

ORIGINAL ARTICLE Distinguishing clinical and radiological features of non-traumatic convexal subarachnoid hemorrhage J. Graff-Radford, J. E. Fugate, J. Klaas, K. D. Flemming, R. D. Brown and A. A. Rabinstein Department of Neurology, Mayo Clinic, Rochester, MN, USA Keywords: convexal, convexity, sulcal subarachnoid hemorrhage Received 10 August 2015 Accepted 4 November 2015 European Journal of Neurology 2016, 0: 1 8 doi:10.1111/ene.12926 Introduction Correspondence: A. A. Rabinstein, MD, Mayo Clinic, 200 1st Street SW, Rochester, MN 55905, USA (tel.: (507) 507 284 4741; fax: (507) 538 6012; e-mail: rabinstein.alejandro@mayo.edu). Background and purpose: The full spectrum of causes of convexal subarachnoid hemorrhage (csah) requires further investigation. Therefore, our objective was to describe the spectrum of clinical and imaging features of patients with non-traumatic csah. Methods: A retrospective observational study of consecutive patients with non-traumatic csah was performed at a tertiary referral center. The underlying cause of csah was characterized and clinical and imaging features that predict a specific etiology were identified. The frequency of future csah or intracerebral hemorrhage (ICH) was determined. Results: In all, 88 patients [median age 64 years (range 25 85)] with non-traumatic csah were identified. The most common causes were reversible cerebral vasoconstriction syndrome (RCVS) (26, 29.5%), cerebral amyloid angiopathy (CAA) (23, 26.1%), indeterminate (14, 15.9%) and endocarditis (9, 10.2%). CAA patients commonly presented at an older age than RCVS patients (75 years versus 51 years, P < 0.0001). Thirteen patients (14.7%) had recurrent csah, and 12 patients (13.6%) had a subsequent ICH. However, the risk was high amongst those with CAA compared to those caused by RCVS, with recurrent csah in 39.1% and subsequent lobar ICH in 43.5% of CAA cases. Conclusions: Our study demonstrates the clinical diversity of csah. Older age, sensorimotor dysfunction and stereotyped spells suggest CAA as the underlying cause. Younger age and thunderclap headache predict RCVS. Yet, various other causes also need to be considered in the differential diagnosis. Approximately 85% of subarachnoid hemorrhage (SAH) is caused by a ruptured brain aneurysm with the majority of the remaining non-aneurysmal cases composed of perimesencephalic SAH [1]. Non-traumatic convexal SAHs (csahs) make up approximately 5% 6% of SAH cases [2]. Recently, several case series have been published highlighting that cerebral amyloid angiopathy (CAA) and reversible cerebral vasoconstriction syndrome (RCVS) are the most common causes of csah [2 4]. These studies have suggested that in patients older than 60 years CAA is the most common cause, often presenting with transient ischaemic attack (TIA) like events; and in patients less than 60 years old RCVS is the most common cause [2 4]. Still, given the small number of subjects investigated in these studies, the overall spectrum of the causes, imaging features and clinical outcomes of csah remains poorly characterized. The objective of the current study was to describe the causes and the clinical and imaging features of nontraumatic csah. Participants Methods The Mayo Clinic Medical Records Linkage system, which allows the identification of virtually all cases of a specific diagnostic entity seen over a specified EUROPEAN JOURNAL OF NEUROLOGY 1

2 J. GRAFF-RADFORD ET AL. period, was used to identify all patients evaluated with non-traumatic csah between 1 January 1996 and 1 October 2014. Only patients with non-traumatic csah were included. Clinical and follow-up data were only obtained through the medical record. Diagnostic criteria The modified Boston criteria including lobar microbleeds were used to classify patients as probable or possible CAA [5,6]. The presence of csah or superficial siderosis was not included in the diagnostic criteria of CAA. Since follow-up imaging was available in a subset of patients, microbleeds seen after the csah were included. The diagnosis of definite RCVS was based on published criteria [7] and includes (i) acute onset of severe headache (often thunderclap); (ii) monophasic course without new symptoms after 1 month; (iii) vasoconstriction on angiography [magnetic resonance angiography (MRA), computed tomography angiography (CTA) or catheter]; (iv) normal or near normal cerebrospinal fluid; (v) resolution or substantial improvement in vasoconstriction within 12 weeks of onset. Patients were included as probable RCVS if they met all the RCVS criteria but did not have follow-up cerebral angiography [8]. The remaining diagnoses were recorded following review of the comprehensive medical record. Etiology of csah was independently confirmed by two reviewers. Neuroimaging data Head computed tomography (CT), brain magnetic resonance imaging (MRI) scans and vessel imaging studies (MRA, CTA and direct catheter angiograms) were reviewed to abstract neuroimaging data. Neuroimaging data abstracted for all patients included the presence of microbleeds, evidence of superficial siderosis on follow-up MRI, presence of diffusion-weighted imaging abnormalities and presence of prior cerebral infarction. Vessel imaging was reviewed for evidence of vasoconstriction, occlusion and significant stenosis. Superficial siderosis was considered present on followup imaging (i.e. after acute presentation) if there was evidence of chronic hemosiderin deposition or if hemosiderin deposition was present in another location at the time of the index csah. Statistics The characteristics of clinical groups were described with median, range, counts and proportions. For comparison of groups, Wilcoxon rank sum tests were performed for continuous variables and v 2 tests for categorical variables. Only the two most common causes of csah were compared because of the limited sample size of the other causes. Standard protocol approvals, registrations and patient consents All patients consented to the use of their Mayo Clinic comprehensive medical record for the purpose of research, and the study was approved by the Mayo Clinic Institutional Review Board. Results Clinical and neuroimaging features A total of 88 patients had non-traumatic csah. The demographics can be found in Table 1. The most common presenting clinical feature was headache (n = 37, 42.0%) followed by sensorimotor dysfunction (n = 24, 27.3%) and confusion/altered mental status (n = 14, 15.9%). Figure 1 summarizes the patterns of restricted diffusion seen. All 88 patients underwent neuroimaging. Seventynine (89.8%) had MRI imaging. Nine (10.2%) had only CT at presentation. Seventy-six (86.4%) had vessel imaging with one or more of the following: MRA, CTA or catheter angiogram. Table 1 Demographics of convexal subarachnoid hemorrhage (csah)patients Age at csah occurrence, median (range) 64 (25 85) Female, no. (%) 53 (60) Previous ICH, no. (%) 1 (1.1) Antiplatelet use at presentation, no. (%) 39 (44.3) Anticoagulation use at presentation, no. (%) 9 (10.2) Vasoactive medication use at presentation, no. (%) 22 (25) Outcome Future ICH, no. (%) 12 (13.6) Recurrent csah, no. (%) 13 (14.7) Neuroimaging MRI, no. (%) 79 (89.8) CT only (%) 9 (10.2) Hemosiderin-sensitive MRI at presentation, no. (%) 70 (79.5) Involvement of one sulcus, no. (%) 38 (43.1) Unilateral, no. (%) 64 (72.7) Right sided (amongst unilateral), no. (%) 40 (45.4) Acute ischaemic infarction concurrent, no. (%) 22 (25) Prior cerebral infarction, no. (%) 25 (28.4) DWI abnormality a, cortical, no. (%) 15 (18.9) Microbleed b, no. (%) 29 (41.4) ICH, intracerebral hemorrhage; MRI, magnetic resonance imaging; CT, computed tomography; DWI, diffusion-weighted imaging. Acute ischaemic infarction concurrent refers to an acute infarct on MRI or CT not next to csah. DWI abnormality, cortical refers to restricted diffusion adjacent to the csah. a Only 79 with MRI counted; b out of 70 with hemosiderin-sensitive sequences.

NON-TRAUMATIC CONVEXAL SUBARACHNOID HEMORRHAGE 3 Figure 1 Restricted diffusion-weighted imaging patterns in convexal subarachnoid hemorrhage (csah). Top row: Patient with csah with acute blood seen in the sulcus on magnetic resonance imaging (MRI) fluid-attenuated inversion recovery (FLAIR) sequence. Cortically based linear restricted diffusion is detected adjacent to the hemorrhage. Bottom row: Patient with csah with acute blood seen on MRI FLAIR sequence with restricted diffusion remote from the site of blood in addition to cortically based restricted diffusion at the site of the hemorrhage. Diagnostic groups Table 2 summarizes the clinical and imaging features of each diagnostic group. Cerebral amyloid angiopathy Sensorimotor dysfunction consisted of unilateral migrating numbness/tingling occurring over minutes and/or heaviness/clumsiness of the arm and face. The sensorimotor events were often initially diagnosed as TIAs or seizures. Speech and language dysfunction would occasionally accompany the events. Eighteen patients (78.3%) presented with stereotyped episodes (recurrent with the same features). Visual change, confusion or generalized tonic clonic seizure was the presenting symptom for one patient (4.3%) each. One (4.3%) had an intracerebral hemorrhage (ICH) prior to the csah. Five patients (21.7%) had evidence of superficial siderosis present in another location at the time of the index csah. The median time to recurrent csah (n = 9) was 7 months (range 0 65). The median time to ICH (n = 10) amongst CAA patients was 29 months (range 0 65 months). In two cases, the ICH occurred in the region of the csah. Four patients with subsequent ICH were on aspirin 81 mg after the initial csah. Three of these patients had a history of symptomatic coronary artery disease. No patients were on anticoagulation following csah. Figure 2 highlights several neuroimaging features of csah-associated CAA. Reversible cerebral vasoconstriction syndrome Twenty patients (77%) met criteria for RCVS. Six (23%) were categorized as probable RCVS (they did not have repeat angiographic imaging). Four had coexisting posterior reversible encephalopathy syndrome (PRES) and one case was post-partum. The presenting symptom consisted of headache in 23 cases [88.5% (21 thunderclap)], seizures in two patients and hemiparesis in one patient. One patient had an occipital lobe ICH after presentation approximately 1 week after onset of symptoms. All six patients with probable RCVS had clinical follow-up with resolution or significant improvement of symptoms. Posterior reversible encephalopathy syndrome Two patients presented with generalized tonic clonic seizures and two with confusion. The MRI scans were consistent with PRES in addition to having coexisting csah. Endocarditis All patients with bacterial endocarditis had positive blood cultures and were found to have a vegetation by trans-esophageal echocardiogram. One patient had non-bacterial thrombotic endocarditis (which was diagnosed at autopsy). Five patients presented with confusion and one each with headache and stereotyped episodes of sensorimotor dysfunction. The patient with non-bacterial thrombotic endocarditis presented as a transfer to the medical intensive care unit for respiratory failure; prior to transfer, the patient had complained of peri-oral numbness which was followed by the development of encephalopathy. The csah was discovered incidentally in two patients who underwent CT imaging for surgical planning purposes. These patients had already been diagnosed with

4 J. GRAFF-RADFORD ET AL. Table 2 Clinical and radiological characteristics by diagnostic subtype Etiology CAA RCVS PRES Endocarditis Venous thrombosis Large vessel disease Vasculitis Other d Indeterminate Number 23 26 4 9 4 4 2 2 14 Age at bleed, 75 (62 85) 51 (25 77) 68.5 (33 72) 64 (47 77) 53.5 (48 66) 70 (50 76) 59 (38 80) 69 (67 71) 66.5 (29 80) median (range) Female, no. (%) 13 (56.5) 21 (80.7) 2 (50) 0 (0) 2 (50) 4 (100) 2 (100) 2 (100) 7 (50) Most common NA Headache 6 (42.9) presenting symptom, no. (%) Thunderclap headache, no. (%) Sensorimotor 20 (87) Headache 23 (88) Seizure 2 (50) Confusion 5 (55.6) Headache 3 (75) Hemiparesis 1 (25) Headache 1 (50) 0 (0) 21 (80.7) 0 (0) 0 (0) 1 (25) 0 (0) 1 (50) 1 (50) 4 (28.6) Stereotyped events 18 (78.3) 0 (0) 0 (0) 1 (11) 0 (0) 0 (0) 0 (0) 0 (0) 2 (14.3) Vasoactive drugs 1 (4.3) 17 (65.4) 1 (25) 0 (0) 0 (0) 0 (0) 1 (50) 0 2 (14.3) Recurrent csah 9 (39.1) 2 (7.7) 0 (0) 1 (11.1) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) ICH after csah 10 (43.4) 1 (3.8) 0 (0) 1 (11.1) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) Seizure a 1 (4.3) 3 (11.5) 2 (50) 1 (11.1) 0 (0) 1 (25) 0 (0) 0 (0) 2 (14.3) Epileptiform 2/16 0/3 0/2 0/1 0/0 0/2 0/1 0/0 1/4 abnormality on EEG/total number with EEG Neuroimaging Microbleeds, no. (%) b 23/23 (100) 0 (0) 0 (0) 2/4 (50) 0 (0) 1/4 (25) 1/2 (50) 1/2 (50) 1/9 (11.1) Unilateral, no. (%) 21 (91.3) 15 (57.7) 3 (75) 5 (55.6) 1 (25) 4 (100) 1 (50) 2 (100) 12 (85.7) Involvement of one sulcus, no. (%) Acute ischaemic infarction concurrent, no. (%) Old infarction, no. (%) DWI cortical, no. (%) Superficial siderosis at follow-up, no. (%) c 12 (52.2) 10 (38.5) 2 (50) 3 (33.3) 1 (25) 2 (50) 1 (50) 0 (0) 7 (50) 7 (30.4) 4 (15.4) 1 (25) 4 (44.4) 1 (25) 4 (4) 0 (0) 0 (0) 4 (28.6) 13 (56.5) 0 (0) 0 (0) 3 (33.3) 0 (0) 4 (100) 0 (0) 1 (0) 4 (28.6) 9/23 (39.1) 1/25 (4) 0 (0) 2/4 (50) 0 (0) 2/4 (50) 0 (0) 0 (0) 1/11 (9.1) 19/20 (95) 2/7 (28.5) 0 (0) 0 (0) 0 (0) 2/4 (50) 0 (0) 0 (0) 2/6 (33.3) CAA, cerebral amyloid angiopathy; RCVS, reversible cerebral vasoconstriction syndrome; PRES, posterior reversible encephalopathy syndrome; csah, convexal subarachnoid hemorrhage; ICH, intracerebral hemorrhage; EEG, electroencephalography; DWI, diffusion-weighted imaging; MRI, magnetic resonance imaging; NA, not applicable. a Seizures occurring at any time during the evaluation; b out of those with hemosiderin-sensitive sequences; c out of those with follow-up MRI scans with hemosiderin-sensitive sequences; d other (thrombosed developmental venous anomaly, hyperperfusion syndrome).

NON-TRAUMATIC CONVEXAL SUBARACHNOID HEMORRHAGE 5 (a) (b) (c) (d) (e) (f) (g) (h) Figure 2 Evolution of cerebral amyloid angiopathy. Top row: 15 months before convexal subarachnoid hemorrhage (csah) an 80- year-old underwent brain magnetic resonance imaging (MRI) for dementia. Susceptibility-weighted imaging demonstrates microbleeds in the left temporal lobe (a) and cerebellum (b) and evidence of superficial siderosis (c). (d) MRI fluid-attenuated inversion recovery sequence demonstrates no evidence of acute csah. Bottom row: 15 months after initial MRI the patient returns with right hemibody migrating paresthesias. Computed tomography head scan demonstrates acute csah (e). Two months later, the acute csah has resolved (f). Approximately 10 months after csah, the patient presents with a lobar intracerebral hemorrhage (g, h). endocarditis and surgery was being planned for valve repair or replacement. One patient developed an ICH approximately 2 weeks after having a csah. One patient developed a recurrent csah 3 months after the initial bleed. Venous thrombosis Three were dural sinus thrombosis and one was a cortical vein thrombosis. Three presented with headache, and one presented with a generalized tonic clonic seizure. Large artery occlusive disease Four patients had large artery disease deemed contributory to the csah. These consisted of one left carotid occlusion, one right middle cerebral artery occlusion, one right carotid dissection with luminal compromise and one high-grade right carotid artery stenosis. Presenting symptoms included neck pain, headache, confusion and hemiparesis. Vasculitis Two patients had central nervous system vasculitis as a cause of csah. One was biopsy-proven amyloid-brelated angiitis (ABRA), and one had a systemic vasculitis involving the skin and central nervous system. She had a temporal artery biopsy consistent with vasculitis. The systemic vasculitis case presented with headache (thunderclap), and the ABRA patient presented with confusion. Both patients had angiographic findings consistent with vasculitis. Other One patient was a 67-year-old female who presented with headache and on imaging was found to have a thrombosed developmental venous anomaly detected on the susceptibility-weighted imaging sequence. The other was a 71-year-old female with cerebral hyperperfusion syndrome after carotid surgery. Indeterminate In 14 cases (15.9%), a clinical diagnosis could not be made with certainty, despite independent review by two neurologists. Six patients presented with headache (four were thunderclap). Twelve patients had MRI imaging and two had CT imaging. One patient could not get an MRI because of an implantable cardioverter defibrillator in place and the other was in the intensive care unit and too medically ill

6 J. GRAFF-RADFORD ET AL. to be transported for MRI. Eleven had MRA or CTA and six also had conventional angiograms. Three patients presented with sensorimotor events, which were stereotyped in two. Four patients presented with confusion. One patient presented with seizure. One patient presented with recurrent csah 11 months after the initial bleed. This patient had thunderclap headaches with both csah. Spinal fluid examination and vessel imaging including catheter angiogram were negative in both instances. One patient had multiple microbleeds but could not be classified as amyloid angiopathy because he was being treated for acute myeloid leukemia. Two cases would meet criteria for possible CAA due to the presence of superficial siderosis based on modified criteria. Differences by age The clinical and imaging features were separated by age greater than or equal to 60 years since this has been shown to be an important age for diagnosis [3]. See Table 3 for details. Comparisons between groups Patients with CAA (n = 23) were compared to those with RCVS (n = 26). CAA patients were older than RCVS patients at the time of presentation (75 years versus 51 years, P < 0.0001). There was a trend towards a higher proportion of women with csah caused by RCVS compared to csah caused by CAA (81% vs. 57%, P = 0.065). Those with CAA had a high rate of subsequent ICH (43.4%) and superficial siderosis (95%) compared to 3.8% and 28.5% respectively for those with RCVS. Table 3 Symptoms and imaging findings by age Presenting symptom Age less than 60 years (n = 36), no. (%) Age greater than 60 years (n = 52), no. (%) Sensorimotor event(s) 0 (0) 24 (46) Headache 31 (86) 6 (11.5) Seizure 1 (2.7) 5 (9.6) Confusion 3 (8.3) 11 (21.2) Hemiparesis 1 (2.7) 2 (3.8) Asymptomatic 0 (0) 2 (3.8) Visual change 0 (0) 1 (1.9) (difficulty reading) Neck pain 0 (0) 1 (1.9) Neuroimaging Involvement of one sulcus 12 (33) 26 (50) Unilateral 21 (58.3) 43 (82.7) Acute ischaemic 6 (16.7) 16 (30.8) infarction concurrent Old infarction 2 (5.6) 23 (44) Microbleeds 1 (2.8) 28 (53.9) Discussion In this study, the broad spectrum of causes of nontraumatic csah is demonstrated. The term convexal subarachnoid hemorrhage was used, similar to most of the literature, although others have used cortical or sulcal SAH [9,10]. The diagnostic groups included RCVS, CAA, endocarditis, cerebral vein thrombosis, PRES, large artery occlusive disease, vasculitis, thrombosed developmental venous anomaly, and hyperperfusion syndrome after carotid endarterectomy. Similar to most previous reports [3,4], RCVS and CAA are found to be the most common causes of csah. However, this series highlights the need to consider a broader differential. Unusual causes in this clinical series included a thrombosed developmental venous anomaly and a case of non-bacterial thrombotic endocarditis. Other causes noted in the literature, but not seen in the current series, include meningitis, dural arteriovenous fistula, abscess, coagulopathy and cavernoma [11]. Older age distinguished cases of CAA from RCVS as has been previously reported [3,4]. Whilst age is an important clue, the overlapping ranges suggest that the age cutoff is not diagnostic, as proved by one of our cases of RCVS that was diagnosed in a 77-yearold. Headache, often thunderclap, was the most common presenting feature of RCVS, similar to RCVS cases without csah. CAA patients presented with spells of transient sensorimotor dysfunction, which were often initially diagnosed as TIAs or focal seizures. This characteristic presentation has been previously described [9]. This misdiagnosis may lead to unnecessary introduction or escalation of antithrombotics increasing bleeding risk [2]. These spells commonly stereotyped consisted most often of unilateral spreading sensory paresthesias of the face and upper extremities most commonly with or without associated weakness. The spreading sensory symptoms, stereotyped nature and frequently negative electroencephalogram of the spells have led investigators to suggest cortical spreading depression as the underlying etiology [12]. Yet, ischaemia can also be responsible as suggested by the presence of restricted diffusion in some of these cases. One case of endocarditis mimicked CAA by presenting with stereotyped sensorimotor spells and microbleeds without headache. Whilst some have suggested that the absence of headache strongly suggests CAA in csah [11], patients with csah caused by endocarditis and large vessel arterial occlusions also may present without headache, again supporting the need to consider a broad differential. Not surprisingly, headache was the most common presentation

NON-TRAUMATIC CONVEXAL SUBARACHNOID HEMORRHAGE 7 of patients younger than 60 years whilst those who presented with csah after age 60 most commonly had stereotyped sensorimotor spells. In addition to RCVS, severe headache was the most common presenting symptom of csah caused by cerebral vein thrombosis. Identifying the underlying cause of csah is important not only for identifying the need for immediate treatment but also for prognosis. Overall, approximately 64% of csah patients have an unfavorable outcome 3 years after presentation [4]. Most of this poor outcome is driven by the development of subsequent ICH. Whilst the overall rate of future ICH is relatively low (14%), it is higher in certain subgroups. For example, 43.4% of patients with csah caused by probable CAA develop lobar ICH within the following 5.5 years (median 2.5 years). In other series of patients with csah caused by CAA, the risk of future ICH has ranged from 27% to 43% over 10 30.7 months [13,14]. Recent series suggest that the risk of symptomatic hemorrhage in CAA patients with csah may be higher than in those with strictly lobar microbleeds [12]. In the current series, approximately 40% of CAA patients develop recurrent csah compared to 15% overall. Key differentiating imaging features emerged between groups. As expected, microbleeds were most common in csah caused by CAA but, importantly, also occurred in a significant number of endocarditis cases. Areas of restricted diffusion distant from the csah were seen most commonly in CAA, similar to previous reports [9]. These areas are similar to the areas of restricted diffusion that are seen in probable CAA patients remote from areas of hemorrhage [15]. CAA patients and those with large artery occlusive disease had csah that was most commonly unilateral. Of interest, in a substantial group of CAA patients there was restricted diffusion in the cortex adjacent to the SAH. There are several possible explanations for this finding. It may be that the stereotyped episodes are seizures caused by cortical irritation from blood products, and the seizure results in the cortically based restricted diffusion. Alternatively, the restricted diffusion may be secondary to significant CAA-related microvascular disease [14]. This appearance may be used as an additional diagnostic tool suggesting CAA as the underlying cause of csah even though it was rarely seen with other causes. Superficial siderosis has been associated with CAA [16]. Therefore, the high frequency of the development of superficial siderosis amongst CAA patients is not unexpected. Recently, it has been proposed that superficial siderosis be considered as part of the diagnostic criteria of CAA. Our findings support a strong association between the two and suggest that symptomatic csah may be the initial event in a subset of CAA patients with prominent superficial siderosis. Large artery occlusive disease causing csah has been reported previously [17,18]. In one retrospective series, csah was detected in 0.14% of acute stroke/ TIA cases. In our series, it represented approximately 4.5% of csah cases [19]. The site of the occlusion or severe stenosis occurred both intracranially and extracranially and was always ipsilateral to the csah. In 14 of the csah cases in this series (16%), the etiology remained indeterminate. This is similar to the number of indeterminate cases published in a smaller case series [2]. Several of these cases had features consistent with RCVS such as thunderclap headache or CAA such as stereotyped sensorimotor events, but with the absence of visualized microbleeds. This study has several limitations. Due to the retrospective nature and the long time span within which the data were collected, the patients were not evaluated in a standardized manner. The possibility that some patients with endocarditis had concomitant amyloid angiopathy cannot be excluded, especially in those for whom the csah was incidental. In addition, it was not possible to make statistical comparisons between all diagnostic subgroups because of the low number of cases in the majority of groups. Conclusion This study shows that the spectrum of causes of csah is extensive. A deliberate approach to the workup of csah allows identification of the underlying cause in the majority of cases, which is important because of the disparate outcomes amongst etiologies. The high rate of future ICHs, the development of superficial siderosis and recurrent csah amongst CAA patients deserve particular attention. None. Acknowledgements Disclosure of conflicts of interest The authors declare no financial or other conflicts of interest. References 1. van Gijn J, Rinkel GJ. Subarachnoid haemorrhage: diagnosis, causes and management. Brain 2001; 124: 249 278. 2. Khurram A, Kleinig T, Leyden J. Clinical associations and causes of convexity subarachnoid hemorrhage. Stroke 2014; 45: 1151 1153.

8 J. GRAFF-RADFORD ET AL. 3. Kumar S, Goddeau RP Jr, Selim MH, et al. Atraumatic convexal subarachnoid hemorrhage: clinical presentation, imaging patterns, and etiologies. Neurology 2010; 74: 893 899. 4. Beitzke M, Gattringer T, Enzinger C, Wagner G, Niederkorn K, Fazekas F. Clinical presentation, etiology, and long-term prognosis in patients with nontraumatic convexal subarachnoid hemorrhage. Stroke 2011; 42: 3055 3060. 5. Knudsen KA, Rosand J, Karluk D, Greenberg SM. Clinical diagnosis of cerebral amyloid angiopathy: validation of the Boston criteria. Neurology 2001; 56: 537 539. 6. van Rooden S, van der Grond J, van den Boom R, et al. Descriptive analysis of the Boston criteria applied to a Dutch-type cerebral amyloid angiopathy population. Stroke 2009; 40: 3022 3027. 7. Ducros A. Reversible cerebral vasoconstriction syndrome. Lancet Neurol 2012; 11: 906 917. 8. Singhal AB, Hajj-Ali RA, Topcuoglu MA, et al. Reversible cerebral vasoconstriction syndromes: analysis of 139 cases. Arch Neurol 2011; 68: 1005 1012. 9. Apoil M, Cogez J, Dubuc L, et al. Focal cortical subarachnoid hemorrhage revealed by recurrent paresthesias: a clinico-radiological syndrome strongly associated with cerebral amyloid angiopathy. Cerebrovasc Dis 2013; 36: 139 144. 10. Renou P, Tourdias T, Fleury O, Debruxelles S, Rouanet F, Sibon I. Atraumatic nonaneurysmal sulcal subarachnoid hemorrhages: a diagnostic workup based on a case series. Cerebrovasc Dis 2012; 34: 147 152. 11. Rico M, Benavente L, Para M, Santamarta E, Pascual J, Calleja S. Headache as a crucial symptom in the etiology of convexal subarachnoid hemorrhage. Headache 2014; 54: 545 550. 12. Ni J, Auriel E, Jindal J, et al. The characteristics of superficial siderosis and convexity subarachnoid hemorrhage and clinical relevance in suspected cerebral amyloid angiopathy. Cerebrovasc Dis 2015; 39: 278 286. 13. Ly JV, Singhal S, Rowe CC, Kempster P, Bower S, Phan TG. Convexity subarachnoid hemorrhage with PiB positive pet scans: clinical features and prognosis. J Neuroimaging 2015; 25: 420 429. 14. Martinez-Lizana E, Carmona-Iragui M, Alcolea D, et al. Cerebral amyloid angiopathy-related atraumatic convexal subarachnoid hemorrhage: an ARIA before the tsunami. J Cereb Blood Flow Metab 2015; 35: 710 717. 15. Kimberly WT, Gilson A, Rost NS, et al. Silent ischemic infarcts are associated with hemorrhage burden in cerebral amyloid angiopathy. Neurology 2009; 72: 1230 1235. 16. Linn J, Halpin A, Demaerel P, et al. Prevalence of superficial siderosis in patients with cerebral amyloid angiopathy. Neurology 2010; 74: 1346 1350. 17. Geraldes R, Santos C, Canhao P. Atraumatic localized convexity subarachnoid hemorrhage associated with acute carotid artery occlusion. Eur J Neurol 2011; 18: e28 e29. 18. Kleinig TJ, Kimber TE, Thompson PD. Convexity subarachnoid haemorrhage associated with bilateral internal carotid artery stenoses. J Neurol 2009; 256: 669 671. 19. Nakajima M, Inatomi Y, Yonehara T, Hirano T, Ando Y. Nontraumatic convexal subarachnoid hemorrhage concomitant with acute ischemic stroke. J Stroke Cerebrovasc Dis 2014; 23: 1564 1570.