Frequency and Characteristics of Spinal Bleeding Sources in Nontraumatic Angiogram-Negative Subarachnoid Hemorrhage

Nontraumatic angiogram-negative subarachnoid hemorrhage (AN-SAH) is an acute bleeding disease of the brain and spine with worldwide regional variations [1‒4], accounting for 5–15% of SAH [5‒8]. 65% of AN-SAH cases are characterized by prepontine or perimesencephalic (PM) hemorrhage [3, 9]. In most cases of AN-SAH, digital subtraction angiography (DSA) and cerebral magnetic resonance imaging (MRI) are unable to detect the source of bleeding [8, 10]. However, standard diagnostic approaches focus on the neurocranium and craniocervical junction [11, 12]. Accordingly, data on spinal bleeding sources remain scarce and the value of repeated or additional imaging is controversial [13‒16]. The present study aims to evaluate characteristics of spinal bleeding sources in cases of AN-SAH, including PM and non-PM (NPM) as well as CT-negative SAH, to guide indications for further diagnostics and imaging.

SAH was diagnosed by computed tomography (CT) or lumbar puncture (LP). If a LP had to be performed, additionally to xanthochromia the concentration of ferritin in the cerebrospinal fluid was analyzed with a cut-off of >15 ng/mL to diagnose SAH [17]. Referring to our institution’s standardized SAH protocol developed in accordance with international guidelines, all patients suffering from SAH underwent DSA to rule out intracranial bleeding sources [12, 18]. If the initial cranial DSA remained negative, cases were classified as AN-SAH. Medical charts of all patients with AN-SAH treated at the author’s institution between January 2012 and January 2022 were reviewed for this retrospective single-center cohort study. Patients with aneurysmatic or traumatic SAH were excluded. PM SAH was defined according to Rinkel et al. [4] with center of the hemorrhage located immediately anterior to the midbrain, with or without extension of blood to the anterior part of the ambient cistern or to the basal part of the sylvian fissure, no complete filling of the anterior interhemispheric fissure and no extension to the lateral sylvian fissure, except for minute amounts of blood, and absence of frank intraventricular hemorrhage. NPM SAH was defined as the remaining patterns of SAH. We analyzed baseline clinical data such as age, sex, antithrombotic therapy, hypertension, neurological/WFNS status on admission [19], and radiological data such as the Fisher grade [20] of all patients.

According to our institutional protocol, all patients with AN-SAH should have received a standardized clinical brain MRI including either the craniocervical junction or the complete spinal axis to detect a potential bleeding source. However, some patients have brought the images from other institutions or had refused further MRI due to claustrophobia. MR imaging was performed on clinical whole-body 1.5 T and 3.0 T systems (Achieva, Philips Healthcare, Best, The Netherlands; for 1.5 T: gradient system: 45 mT/m maximum amplitude, 200 T/m/s maximum slew rate; for 3.0 T: 80 mT/m maximum amplitude, 200 T/m/s maximum slew rate; equipped with dual-source RF transmission technology). Morphological MR imaging of the spine was acquired according to the routine clinical MRI protocol used at our institution which included at least a sagittal T1-weighted spin-echo (450–750/6–12, repetition time, TR, msec/echo time, TE, msec) and T2-weighted turbo spin-echo sequence (3,000–5,000/80–120, TR/TE) as well as a sagittal T2 spectral attenuated inversion recovery-weighted turbo spin-echo sequence (3,000–5,000/50–120, TR/TE), the craniocervical junction was additionally imaged with a dedicated high-resolution heavily T2-weighted driven equilibrium (DRIVE) spin-echo sequence acquired in axial plane (1,500–3,000/100–250, TR/TE). In particular, this study included patients who received spinal MRI in our institute that were still available during data acquisition. All patients were monitored for at least 21 days with repeated routine clinical CT imaging to rule out any delayed complications after bleeding such as re-bleeding, cerebral vasospasm, vasospasm-related cerebral isolated or general infarction, and hydrocephalus.

Data were analyzed using SPSS Statistics (Version 27, IBM Corp. Armok, NY, USA). Parametric data are presented as mean values ±standard deviation and were analyzed by two-tailed t test. Categorical variables are given as absolute and relative frequencies and were analyzed in contingency tables by using the Fisher’s exact test. Considering the small sample size, relative risk (RR) and 95% confidence interval (CI) were calculated when p values <0.05.

140 AN-SAH patients were treated at the authors’ neurosurgical department between January 2012 and January 2022. 52.1% were diagnosed with PM SAH, 35.0% with NPM SAH, and 12.9% with CT-negative SAH. The mean age was 56.2 (±14.1) years and the female-male ratio was balanced.

Most patients (85.7%) were in a good clinical condition on admission (WFNS score 1–2). Radiological Fisher grades 3–4 were found in 84.3% of cases. In 87.1% of cases, AN-SAH was detected by CT and in 12.9% of cases with LP. In addition, 85.0% of patients underwent craniocervical and/or holospinal MRI. Some patients did not show up to the scheduled follow up investigations or refused to perform control imaging; 76 patients (54.3%) underwent secondary DSA and 25 patients (17.9%) showed up with a secondary MRI.

Antithrombotic therapy (oral anticoagulation or platelet aggregation inhibitors) was registered in 17.1%, hypertension in 42.9% and nicotine abuse in 20.0%. After discharge, none of the 140 patients was re-admitted to our institution due to rebleeding. Table 1 shows the baseline characteristics of the AN-SAH patients.

A spinal bleeding source was detected in four cases (2.85%) through MRI of craniocervical junction or spinal axis: an intraspinal aneurysm at T3 (shown in Fig. 1a), a perimedullary angioma at T11, an arteriovenous malformation at C3/4 (shown in Fig. 1b), and a sub-ependymoma at the IV ventricle and craniocervical junction. 75% of these patients complained initially about local back pain or had neurological deficits. Spinal bleeding sources were more likely in younger patients (mean difference 14.85 years [95% CI: 0.85–28.85; p = 0.038]) and patients suffering from spinal symptoms like local spinal pain or neurological deficits (RR: 3.9706 [95% CI: 0.7272–21.6792; p < 0.001]).

AN-SAH is common and accounts for up to 15% of SAH cases [8, 13, 21, 22]. In PM AN-SAH, the center of bleeding is prepontine or in the PM cisterns [23, 24]. The NPM AN-SAH pattern is less specific and may be caused by a vertebrobasilar aneurysm rupture or arteriovenous malformation; but a bleeding source has only been identified in up to 5% of cases [25, 26]. The data about craniocervical and thoracic bleeding sources are scarce. Therefore, the present study aimed to assess frequency and characteristics of spinal bleeding sources.

We report a spinal bleeding source in 2.86% of patients with AN-SAH. Kashefiolasl et al. [27] reported an MRI detection rate and incidence of lumbar bleeding sources of 1.05% among 190 patients with AN-SAH. Some authors have reported a spinal tumor as the bleeding source [27, 28]. The most common symptoms in such cases are back pain and headache [29]. A traumatic spinal bleeding source has also been reported [30]. In our cohort, the bleeding was caused by an arteriovenous malformation in two cases and an aneurysm and a tumor in one case each, and was associated with neurological deficits or local spinal pain. Most studies about spinal bleeding sources did not differentiate between SAH and AN-SAH, so detailed analyzation of these data was limited [31‒33].

Our data are consistent with the literature regarding demographic parameters and relative frequencies of PM and NPM AN-SAH. Of note, patients with an identified spinal bleeding source were younger than patients without a spinal bleeding source [3, 34, 35]. Patients with AN-SAH had a good neurological and clinical status on admission, which also consists with previous studies [6, 23, 36, 37]. Co-factors such as hypertension and nicotine consumption were not relevant in our cohort. These are sometimes described as associated factors for SAH or contributing to increased risk for intracerebral bleeding [37‒41], whereas other authors reported no such correlation [42, 43]. The correlation between antithrombotic therapy and AN-SAH was also not relevant in our study but is subject of controversial discussion [44‒47].

Most of our AN-SAH patients received second DSA after a few weeks, which remained negative in all cases. While some authors recommended second DSA [15, 16, 48, 49], others advise against it due to the risk of complications [1, 24, 50, 51]. Other authors consider repetition of CT angiography and MRI instead [25, 52]. However, the risk of re-bleeding after AN-SAH is described as very low and no case occurred in our cohort [10].

Besides imaging, there may be predictors for bleeding sources. There are multiple known genetic syndromes that involve aneurysm or arteriovenous malformation development such as hereditary hemorrhagic telangiectasia and Loeys-Dietz syndrome [53, 54]. Most of these syndromes describe vascular pathologies in the brain, but they may also occur at the spine [55]. These studies describe an association to the transforming growth factor-β signaling pathway that may predispose for developing vascular pathologies in the brain and spine. Unfortunately, we were not able to analyze this association due to the retrospective design of this study. However, it could be a possible opportunity for further investigations and diagnostics in uncertain cases.

Admittedly, performing such extensive diagnostics is not feasible in resource-limited settings. Nevertheless, we believe that spinal bleeding sources should be considered in AN-SAH, especially in presence of the described symptoms, as optimal treatment can only be administered after adequate detection.

This study has several limitations. AN-SAH is a rare disease, limiting the sample size of our cohort, especially of patients with a spinal bleeding source. The results of the statistical analyses have to be interpreted with caution. This limitation can only be overcome with a longer observational period (resulting in more different diagnostic workup) or multicentric approach. Another limitation is the retrospective character of this study. Therefore, repeated analysis of the images or traceability of the indications for performing different images (e.g., MRI of the craniocervical junction or the complete spinal axis) was not achievable. So furthermore, that MRI imaging was not performed uniformly in all patients, represents another limitation, too.

AN-SAH caused by spinal pathology is rare. This study indicates that craniocervical and holospinal MRI should be considered in patients suffering from AN-SAH, especially for younger patients with neurological deficits or lumbar pain.

This study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of the Medical Faculty and the University Hospital of Bonn (code: 256/22). Informed consent was waived due to the retrospective study design.

The authors have no conflicts of interest to declare.

No funding was received to assist with the preparation of this manuscript.

E.G. and M.B. conceived the study. T.L., S.B., and H.A. collected the data. T.L. and S.B. analyzed the results and drafted the manuscript. J.W., F.C.S., M.H., A.R., E.G., H.V., and M.B. reviewed and edited the manuscript. All authors have read and agreed to the published version of the manuscript.

All data generated or analyzed during this study are included in this published article. Further inquiries can be directed to the corresponding author.