硬脑膜动静脉瘘介入诊断及治疗

硬脑膜动静脉瘘介入诊治的新进展1. 分型和临床表现鉴于DAVFs痿口位置和引流血管构筑的不同，其临床表现也有很大的差异[1] 。

这些临床表现主要包括突眼、颅内血管杂音、颅神经功能障碍、耳鸣、出血、颅内压升高、视神经乳头水肿，严重时甚至有发生充血性心衰的可能。

关于DAVFs的分型有几种，主要的两类分型分别以瘘口位置和引流静脉作为依据。

目前使用最多的分型是Cognard等人和Borden等人制定的分型。

这些分型的共同点是都以DAVFs的引流静脉作为区分的重点，单纯与静脉窦连接、血流为顺行的DAVFs其临床表现相对较轻微，而血流为逆流并有皮层静脉参与的DAVFs则表现出较严重的症状如颅内出血、颅神经功能障碍、颅内压升高等。

以1995 年Cognard等人制定的Cognard分型为例[2] : I型，DAVFs位于静脉窦内，血流为顺行，症状主要为搏动性耳鸣和颅内血管杂音；n 型，DAVFs位于静脉窦内，血流为逆流入静脉窦或皮层静脉，症状以颅内出血、颅神经功能障碍等为主；m型，皮层静脉直接引流，无静脉扩张，临床症状以颅内出血为主；W型，皮层静脉直接引流，有静脉扩张，具有占位效应，临床症状以中枢神经系统症状、颅高压为主，其发生颅内出血的可能性也最大；^型，血液引流入脊髓的髓周静脉，导致椎管内静脉压升高，脊髓缺血，临床表现为锥体束征阳性。

2.介入治疗的进展DAVFs的治疗应重视个体化治疗，充分考虑到疾病的病史、血管构筑情况和临床症状的严重程度。

其治疗原则为尽可能充分、彻底地闭塞瘘口，同时不影响正常的静脉回流[3]。

DAVFs 的治疗方法包括传统外科手术、放射外科、介入治疗以及综合治疗。

区别于前两种治疗方法，介入治疗可使栓塞材料直接到达病灶血管，闭塞瘘口，减少了一系列并发症。

随着技术和栓塞材料的不断进步，介入治疗逐渐成为治疗DAVFs的首选方法，尤其是在一些复杂的、高风险的DAVFs中，外科手术仅用于介入治疗无法实施的病患。

Lin Bo Zhao,MDDae Chul Suh,MD,PhD Dong-Geun Lee,MD Sang Joon Kim,MD, PhDJae Kyun Kim,MD Seungbong Han,PhD Deok Hee Lee,MD,PhD Jong Sung Kim,MD, PhDCorrespondence toDr.Suh:dcsuh@amc.seoul.kr Association of pial venous reflux with hemorrhage or edema in dural arteriovenous fistulaABSTRACTObjective:We investigated whether pial venous reflux(PVR)is associated with hemorrhage or edema in dural arteriovenous fistula(DAVF).Methods:We evaluated the association of hemorrhage or edema with the occurrence of PVR or cortical venous reflux(CVR)in222patients with DAVF.We determined whether angiographic findings of PVR or CVR(more than Borden I or Cognard IIa)were associated with symptoms, lesion location,or brain lesion(hemorrhage or edema).We evaluated the lesion progression or the follow-up results after obliteration of the DAVF.Results:Hemorrhage or edema developed in18%(40/222)of the patients with DAVF and55% (40/72)of the patients with PVR.There were2patterns of PVR associated with hemorrhage or edema:(1)PVR in any particular CVR territory(75%),and(2)direct PVR not via CVR(25%).The presence of brain lesion increased the odds of presence of PVR by4.09times compared to the group without brain lesion(95%confidence interval51.570–11.394,p50.004).Brain edema caused by PVR was reversible after obliteration of the fistula and may have progressed to hem-orrhage without proper patient management performed within several weeks after the initial presentation.Conclusions:Our results show that PVR is more closely associated with the hemorrhage or edema than CVR in patients with DAVF.PVR can occur not only as a part of CVR but also directly in cer-tain types of DAVF.Neurology®2014;82:1897–1904GLOSSARYCVR5cortical venous reflux;DAVF5dural arteriovenous fistula;GK5gamma knife;mRS5modified Rankin Scale;PVR5 pial venous reflux;SAH5subarachnoid hemorrhage.Current classifications of dural arteriovenous fistula(DAVF)are focused primarily on the pres-ence of cortical venous reflux(CVR)related to cerebral venous hypertension leading to cerebral infarction or hemorrhage.1–3CVR is known to be related to the so-called aggressive type of DAVF because29%to46%of patients with CVR may develop cerebral hemorrhage.4–6How-ever,it has not been precisely determined why hemorrhage or edema in certain brain areas is related to CVR.Pial venous reflux(PVR),a part of CVR,has not been clearly identified or differentiated from CVR.7However,there have been only a few descriptions of the relationship of PVR to CVR and the location of their anatomical junction.The aim of this study was to investigate the relationship of PVR vs CVR to hemorrhage or edema.To achieve this,we assessed serial angiographic and cross-sectional imaging findings of CVR and PVR associated with hemorrhage or edema.We then present a concept regarding how PVR occurring in patients with CVR is related to hemorrhage or edema.METHODS We reviewed prospectively collected records of222consecutive patients diagnosed with DAVF at a single medical insti-tution(Asan Medical Center,Seoul,Korea)between July1998and October2012.We analyzed the patients’angiographic findings andFrom the Department of Radiology and Research Institute of Radiology(L.B.Z.,D.C.S.,D.-G.L.,S.J.K.,D.H.L.)and Department of Neurology (J.S.K.),University of Ulsan,College of Medicine,and Department of Epidemiology and Biostatistics(S.H.),Asan Medical Center,Seoul,Korea; Department of Radiology(L.B.Z.),First Affiliated Hospital of Nanjing Medical University,Nanjing,China;and Department of Radiology(J.K.K.), Chung-Ang University,College of Medicine,Seoul,Korea.Go to for full disclosures.Funding information and disclosures deemed relevant by the authors,if any,are provided at the end of the article.©2014American Academy of Neurology1897medical records to assess the patient demographics,the presence of brain lesions(hemorrhage or edema),shunt localization,and the presence of CVR or PVR.We excluded pial-type brain arteriovenous malformations with a dural supply.Selective angiography of the internal carotid artery,external carotid artery,and vertebral arteries was obtained using high-resolution, biplane,digital subtraction angiography(AXIOM Artis zee biplane angiography system;Siemens AG Medical Solutions, Erlangen,Germany).The clinical symptoms were separated into2groups,i.e., benign and aggressive.8The benign group consisted of an inci-dental diagnosis,nonspecific headaches,cranial nerve deficits, chemosis/proptosis,bruit or pulsatile tinnitus,mass lesions,and cardiac insufficiency.The aggressive group included seizures, intracranial hemorrhage,motor or sensory deficits,visual field defects,aphasia,global neurologic deficits(dementia,delayed psychomotor development,macrocrania),and other nonhemor-rhagic neurologic deficits such as incontinence.We did not mea-sure the venous pressure either directly or indirectly.Standard protocol approvals,registrations,and patient consents.The institutional review board approved the study, and written informed consent was obtained from each patient.MRI/CT findings.Patients who presented with hemorrhage or edema seen on MRI/CT obtained before treatment were analyzed and classified into3subgroups:(1)hemorrhage,defined as paren-chymal or subarachnoid hemorrhage(SAH)with little or no edema;(2)edema,defined as parenchymal edema with no evi-dence of hemorrhage;or(3)edema combined with hemorrhage, defined as both edema and hemorrhage,with the edema being disproportionate to the amount that would be expected surround-ing a parenchymal hemorrhage.We correlated symptoms,lesion locations,and angiographic types.The patients who presented with acute neurologic deficits underwent an imaging study accord-ing to our acute stroke protocol,and which therefore included fluid-attenuated inversion recovery imaging(n528),diffusion-weighted imaging(n524),apparent diffusion coefficient imaging (n520),and perfusion imaging(n55).One patient who presented with a brainstem sign showed MRI findings mimicking brainstem tumor and thus underwent magnetic resonance spectroscopy.Because the application of MRI/CT studies varied according to the attending physician or patient’s presenting symptom,analysis of imaging studies was based on the neuroradiologist’s report and was additionally reviewed by consensus of2experienced neuroradiologists(S.J.K.and D.C.S.). Angiographic typing.Angiographically,benign and aggressive lesions were defined according to the absence or presence of CVR9and were also grouped using the classification systems of Borden2and Cognard.1Borden I(sinus drainage only),Cognard I (antegrade sinus drainage without CVR),and Cognard IIa(retro-grade sinus drainage without CVR)were considered as“benign”DAVFs,whereas all of the higher grades that have cortical and spinal drainage with or without sinus drainage were grouped as“aggressive”DAVFs.10,11The main locations of DAVFs were categorized as the cavernous sinus,the transverse-sigmoid sinus,the superior sagittal sinus,the ethmoidal roof,and the petrous area.We also identified a new type of DAVF lesion in the parietotemporal convexity and defined it as parietotemporal convexity DAVF.CVR vs PVR.The presence of cortical and pial venous drainage was determined.Veins were defined as“cortical”when they coursed along the cortical surface draining into the venous sinus and as“pial”when the fine and tortuous veins were within the brain or on the brain surface,as seen on cerebral angiography (figure1E)and/or MRI.Presence of PVR was also decided by comparison with cortical veins in the venous phase of the ipsilat-eral internal carotid arteriogram(figure2,G and J).Compared with cortical veins,which are regarded as the main leptomeningeal veins draining into sinus,the fine pial veins or the intracortical veins beneath the pial membrane were regarded as having a corkscrew-like appearance or intraparenchymal course,12which cannot be seen on a routine normal angiogram(figures1E,2I,and3J).We did not apply any size criteria for the differentiation because the pial vein is much smaller and more peripherally located than the cortical vein. In patients who underwent serial imaging studies,their presenting symptom pattern was compared with the development of a brain lesion according to the time interval.Follow-up.Follow-up data for the222study patients were col-lected from the time of their admission until the end of2012.A complete history was obtained from each patient,and a neurologic examination was performed by independent neurologists who were not involved in the interventional procedure.If a patient was not followed up or the patient’s status was not exactly mentioned in an outpatient clinic,an experienced nurse telephoned the patients to evaluate the possibility of any clinically relevant event.Functional outcome was assessed with the modified Rankin Scale(mRS).13 Median clinical follow-up of all patients was15months (range1–178months),and final outcome was evaluated using the mRS,as shown in the table.The40patients with hemorrhage or edema were followed for a median of12months(range1–155 months).Treatment included embolization in23,gamma knife (GK)irradiation in6,surgical resection in5,and no treatment in 6patients.14Good(mRS score#2)vs poor(mRS score.2) outcome was compared for each treatment modality. Statistical methods.Cross-tabulations using patient sex,age, angio-type(benign vs aggressive),clinical symptoms group (benign vs aggressive),lesion location,brain lesion(hemorrhage or edema),and the presence of CVR or PVR were performed. Statistical significance was calculated for each group using the Fisher exact test and t test for categorical variables and continuous variables,respectively.We conducted univariate and multivariable analyses using variables that were significant in frequency or mean comparison between the presence and absence of PVR compared with the presence of CVR without PVR.Because lesion location was the cavernous sinus in62%of patients and the frequency of some category levels for the other lesion locations was small,we regrouped lesion location into a smaller number of categories(the cavernous sinus vs others).15Similarly,we regrouped135patients with the presence of CVR into presence or absence groups of the brain lesion.The univariate and multivariable logistic regression model proposed by Firth was fitted for the binary outcome variable.This method can handle the separation problem occurring when some categorical levels have zero counts of brain lesion as in the patients with CVR but without PVR.16All statistical analyses were performed using SPSS18software(SPSS Inc., Chicago,IL)and R software(R Foundation for Statistical Computing,Vienna,Austria;).The R package “logistf”was used to fit the bias-reduced logistic regression model.17 Significance was determined at p,0.05.We retrospectively computed the statistical power under some assumptions.Group sample sizes of42in group1with brain lesion and84in group2without brain lesion achieve90%power to detect a difference between the group proportions of0.3.The pro-portion in group1is assumed to be0.3under the null hypothesis and0.6under the alternative hypothesis.The proportion in group 2is0.3.The test statistic used was the2-sided z test with pooled variance.The significance level of the test was targeted at0.05.1898Neurology82May27,2014RESULTS Baseline characteristics.The baseline clini-cal and angiographic features of the 222patients with intracranial DAVF are summarized in the table.There were 134women (60%)and 88men (40%)with a mean age at admission of 57years (range 14–85years).The most common DAVF location was the cavernous sinus region (137patients [62%])followed by the transverse and sigmoid sinus regions (38pa-tients [17%]).Major presenting symptoms or signs in patients who presented with hemorrhage or edema were altered consciousness (n 511),orbital/ocular symptoms related to the brain/brainstem (n 55),neurologic deficit (n 511),seizure (n 56),severe headache (n 56),and dizziness (n 51).Seizure developed in patients with edema in the parietal lobe due to the DAVF of the parietotemporal convexity (n 55)or superior sagittal sinus (n 51)(figure 2).MRI/CT findings.Forty patients presented with hem-orrhage or edema,as seen on MRI/CT.Twenty-nine patients revealed hemorrhage associated with (n 523)or without (n 56)surrounding edema (figure 1).Eleven patients only had edema without evidence of hemorrhage (figure 2).Among the 29patients with hemorrhage,27presented with parenchymal hemorrhage,one presented with intracerebral hemorrhage followed by massive SAH,and one presented with massive SAH with acute hydrocephalus.The location of the hemorrhage was lobar (n 518),the cerebellar hemisphere (n 56),and the brainstem (n 54).One patient who presented with a brainstem sign showed MRI findings mimicking brainstem tumor,but magnetic resonance spectroscopy did not reveal any evidence of tumor or ischemia (figure 3).CVR vs PVR.The distribution of CVR and PVR in222patients is presented in the table.There was no CVR in 39%,CVR only in 28%,CVR and PVR in 28%,and PVR only in 5%.Presence of PVR in 72patients (32%)was associated with hemorrhage or edema (p ,0.001).Among 40patients (18%)who developed hemorrhage or edema,30patients (75%)revealed the presence of CVR and PVR (figure 1).Ten patients (25%)revealed direct filling of PVR not via CVR (figure 2).Univariate analysis revealed that the presence of brain lesion (hemorrhage or edema)increased the odds of the presence of PVR by 5.68times compared with the group without brain lesion (95%confidence interval 52.571–13.369,p ,0.001).Compared with the cavernous sinus location,other locations increased the odds of presence of PVR by 2.80times.Focal cerebral edema with development of subsequent hemorrhage in a 66-year-old woman who presented with neurologic deficit.A magnetic resonance fluid-attenuated inversion recovery image (A)shows localized high signal intensity in the right parietal subcortical area.(B)CT imaging obtained 10days later showed focal hemorrhage surrounded by edema in the same area of the right parietal lobe.Anteroposterior (C)and lateral (D)views of the right external carotid arteriogram show a dural arteriovenous fistula in the superior sagittal sinus (SSS)supplied by the middle meningeal and superficial temporal arteries.There is occlusion of the SSS (white arrows in D and E).Note the diffuse corkscrew-like tortuous,fine pial venous engorge-ment (black arrows)in the late venous phase (E).Her neurologic deficit remained after obliteration of the fistula by intrao-perative coil embolization.Neurology 82May 27,20141899Moreover,aggressive symptoms increased the odds of presence of PVR by 3.23times compared with the benign symptom group.In the multivariable analysis,after adjusting for location and aggressive symptoms,presence of brain lesion was still significant and increased the odds of presence of PVR by 4.09times compared with the group without brain lesion (95%confidence interval 51.570–11.394,p 50.004).Follow-up results.During median 15months offollow-up,there was no difference in good vs poor outcome,likely because the patients with aggressive angio-type underwent active treatment while the others did not (table).During the follow-up period,there were 3patients who developed hemorrhage (n 52)in patients with PVR and hydrocephalus (n 51)in a patient with CVR.A successful treatment outcome (mRS score #2)was obtained in 33of the 40patients:20by emboli-zation,5by GK irradiation,4after surgery,and 4who received no treatment.18There was a poor treatment outcome (mRS score .2,n 57)in 3patients who underwent embolization,one who had surgery,one after GK irradiation,and in 2patients with no treat-ment.Of 29patients who presented with hemorrhage,26were treated using endovascular techniques (n 521,transarterial or transvenous or both),DAVF-resection surgery (n 53),or by GK irradiation (n 52)as the first treatment option.Five patients under-went combined therapy because of incomplete removal of the fistula using GK treatment (n 51)or surgery (n 51)followed by embolization or embolization fol-lowed by surgery (n 52)or GK treatment (n 51).The patients who had presented with hemorrhage showed complete resolution of the hemorrhage on follow-up imaging,and no patient revealed recurrent hemorrhage during the follow-up period.Of the 11patients who presented with brain edema only,8underwent endovascular treatment and 3were lost to follow-up after either GK treatment (n 52)or no treatment (n 51).The edema resolved completely in 7of these patients,as seen on MRIs obtained 2months after treatment (figures 2and 3);edema decreased in one patient after a follow-up period of 1month.Progression of brain lesions.Eight patients developedbrain lesions during the follow-up period.Brain edema (n 52,both 3months after their initial diagnosis)or hemorrhage (n 52,1and 8months after the initial diagnosis)appeared after initial MRI studies were normal.In the patients whopresentedLocalized cerebral edema in a 62-year-old man who presented with right-side myoclonus and tonic seizure.A gradient echo image obtained at the time of the seizure (A)shows a focal edema and dark signal along the left frontoparietal cortex.MRI obtained 3months later shows aggravated edema on T2-weighted image (B)and on T1-weighted image (C).Gadolinium-enhanced image shows slight enhancement along the cortical margin (D).T2-weighted image obtained 4years after embolization (E)shows normalized brain parenchyma without any other neurologic deficit.Note only a faint iron deposition along the cortical margin.Anteroposterior (F)and lateral (G)views of the external carotid angiogram show the superficial temporal artery supplying a dural arteriovenous fis-tula over the left parietal convexity via the emissary artery (thick,short arrows)into the pial veins (thick,long arrows).Note collateral filling of a remote pial vein (thin,long arrow in G –K)via the intracortical veins (thin,short arrows in G –K).Selective anteroposterior (H)and lateral (I)angiograms obtained at the emissary artery show pial venous reflux filling to the intracortical veins (thin,short arrows in G –K)as well as intraparenchymal collateral to the other pial vein (thin,long arrows in G –K).The venous phase of the internal carotid arteriogram (J),in contrast to pial venous reflux,shows no visible abnormality in the cortical venous drainage.Schematic drawing (K)shows a shunt filling the pial vein (thick,long arrow)and the other pial vein filling (thin,long arrow)via the intraparenchymal veins (thin,short arrows).Note cortical veins in the subarachnoid space (pink-colored space).The reason the fistular shunt flow remains in the pial venous system is suggested by thrombosed disconnection (asterisk)of pial-cortical venous drainage.1900Neurology 82May 27,2014with brain edema,there was aggravation of the edema (n 52,1and 3months after initial MRI)(figure 2),development of hemorrhage (n 51,10days after initial MRI)(figure 1),or even improvement with some residual encephalomalacia (n 51,1.5years after initial MRI).The locations of these lesions were the parietotemporal convexity (n 54),cavernous sinus (n 52),superior sagittal sinus (n 51),and transverse-sigmoid sinus (n 51).DISCUSSION Our study suggests that PVR is moreclosely related to hemorrhage or edema.PVR was associated with hemorrhage or edema in 75%of pa-tients and was related to a certain brain area,whereas CVR occurred in a wide vascular territory.Our study also revealed that PVR was found without filling of the cortical vein in a certain type of DAVF in 25%of patients with hemorrhage or edema.Therefore,presence of PVR should be identified in addition to the CVR,which is currently known as a risk factor for hemorrhage in DAVF.19–21In addition,DAVF diagnosis needs to be considered when there is corkscrew-like pial venous engorgement in patients who reveal brain lesion associated with seizure or neurologic deficit.In contrast to arterial ischemic infarction,many parenchymal abnormalities secondary to venous con-gestion are reversible.22If venous hypertension can be relieved before cell death or intracranial hemor-rhage,the parenchymal changes may partially or com-pletely resolve.However,if venous pressure continues to increase,with a consequent reduction in arterial perfusion pressure,cell death may ensue.Our study revealed that edema after embolization was progres-sively completely resolved,although the edema re-mained during the initial short-term follow-up period.Three levels of cortical veins have been described,i.e.,the main leptomeningeal veins,the fine pial net-work,and the intracortical veins.12The main lepto-meningeal veins are located in the pia matter on the surface of the cortex.Pial veins form a dense superfi-cial network 23and they have been found to pass over sulci without entering them.24If the DAVF hemor-rhage is caused by rupture of the main leptomenin-geal veins,it should present more frequently with SAH than with hemorrhage.25This anatomical aspect also demonstrates that it is the intracortical veins or pial veins rather than the main leptomeningeal veins that rupture secondary to venous hypertension in pa-tients with cerebral DAVF.20The diameters oftheA 49-year-old man presented with diplopia,dizziness,and mild dysarthria.A T2-weighted image (A)shows high signal intensities in the left cerebellar pedun-cle and pons.There are no definite signal changes on diffusion-weighted imaging (B)and slightly increased signal intensity on apparent diffusion coefficient image (C).Susceptibility-weighted imaging (D)shows the dilated petrosal vein in the left cerebellopontine angle (long,white arrow)and intracortical venous engorgement (short,white arrow).A perfusion imaging study shows increased mean transit time (E),decreased cerebral blood flow (F),and decreased cere-bral blood volume (G)at the areas of round cursors on the high signal intensities on panel A.Perfusion curve (H)shows decreased perfusion status in the brain edema area (red line)compared with the contralateral normal side (blue line).Single-voxel spectroscopy (I)obtained at the left middle cerebellar peduncle with marked gadolinium enhancement shows decreased choline (Cho)and creatine peak and relatively preserved N -acetyl aspartate (NAA)peak with decreased Cho/NAA ratio suggesting a nontumorous condition.Preembolization (J)and postembolization (K)external carotid angiograms show disappear-ance of a fistular shunt and the engorged petrosal vein (long arrow in J).Note fine corkscrew-like pial veins (short arrows).Fluid-attenuated inversion recovery image obtained 8months after embolization (L)shows normalized brain parenchyma without any other neurologic deficit.Neurology 82May 27,20141901intracortical veins measured by photon microscopy studies were ,80m m,26which could not be detected on 3-tesla MRI or on cerebral digital subtraction angi-ography.Until now,there have been no literature reports describing the diameters of intracranial veins in patients with DAVF.The presence of susceptibility-weighted imaging hyperintensity within the venous structure could be a useful indicator of retrograde leptomeningeal venous drainage in patients with DAVF,although it did not distinguish the 3levels of the cortical vein.27There must be a subarachnoid-pial junction as seen in the cortical veno-dural junction that has a particular role in the cor-tical venous drainage of the brain.This subarachnoid-pial junction was also well demonstrated in a recent embryo-logic study in which it was seen that vessels penetrated the glial membrane into the brain substance.28Because our study is limited by its retrospective design even though we prospectively collected the data at the time of enrollment,we could not provide pro-spective sample size calculation.In addition,we did not obtain the imaging follow-up results especially for the benign symptom group or CVR without brain lesion.Although we described the association of PVR with cerebral hemorrhage or edema,we could not provide a distinctive anatomical demarcation of the cortico-pial venous junction because there was a smooth continuation of the venous drainage at the junctional zone at the pial membrane to the cortical vein,which is located in the subarachnoid space.High-resolution MRI obtained at more than 7tesla may demonstrate the anatomical relationship in the future study.We assumed that there must bearachnoid-duro-pialAbbreviations:CVR 5cortical venous reflux;FU 5follow-up;mRS 5modified Rankin Scale;P-T 5parietotemporal;PVR 5pial venous reflux;SSS 5superior sagittal sinus;Sx 5symptom;T-S 5transverse sigmoid.Data are n (%)unless otherwise indicated.15presence;25absence.aBorden type I or Cognard I 1IIa.bBorden type II and III or Cognard type more than IIb.cThree patients with osseous dural arteriovenous fistula,one patient with multiple dural arteriovenous fistulae.dMultivariable analysis,after adjusting for location and aggressive symptom group,revealed that the presence of brain lesion is still significant and increases the odds of presence of PVR by 4.09times compared with the group without brain lesion (95%confidence interval 51.570–11.394,p 50.004).1902Neurology 82May 27,2014adhesion or occlusion of the cortico-pial junction pre-cluding cortical venous drainage and diverting angioge-netic shunt flow into the pial veins.29However,further anatomical or pathologic studies will be required to support this hypothesis.Our study suggests that PVR is more closely related to brain lesions,such as hemorrhage or edema,than CVR in patients with DAVF.PVR can occur not only as a part of CVR but also directly in certain types of DAVF.Hemorrhage may be secondary to venous edema or the rupture of small cortical veins,especially intracortical veins,because PVR,as a newly proposed concept,may be more closely correlated with venous edema and hemorrhage.Further studies will be required to support the concept and to demonstrate the close relationship between PVR and the aggressive clinical behavior seen in patients with cerebral DAVF. AUTHOR CONTRIBUTIONSDr.Lin Bo Zhao:acquisition and analysis of data,literature review. Dr.Dae Chul Suh:study concept and design,study supervision,final revision.Dr.Dong-Geun Lee:analysis and interpretation of data. Dr.Sang Joon Kim:critical revision of the manuscript for important intellectual content.Dr.Jae Kyun Kim and Mr.Seungbong Han:statis-tical analysis.Dr.Deok Hee Lee:interpretation of data.Dr.Jong Sung Kim:critical revision of the manuscript.ACKNOWLEDGMENTThe authors acknowledge the assistance of Min-ju Kim,Department of Clinical Epidemiology and Biostatistics,Asan Medical Center,with statis-tical analysis.STUDY FUNDINGNo targeted funding reported.DISCLOSUREL.Zhao reports no disclosures relevant to the manuscript.D.Suh serves as an executive committee member of World Federation of Interventional and Therapeutic Neuroradiology and holds patents on an intravascular occlusion device.D.Lee,S.Kim,J.K.Kim,and S.Han report no disclo-sures relevant to the manuscript.D.Lee holds patents on a stroke treat-ment device and guidewire.J.S.Kim serves as an associate editor of the International Journal of Stroke,an editorial board member of Stroke,asso-ciate editor of Cerebrovascular Diseases,and an editorial board member of Neurocritical Care.Go to for full disclosures.Received August29,2013.Accepted in final form February24,2014.REFERENCES1.Cognard C,Gobin YP,Pierot L,et al.Cerebral duralarteriovenous fistulas:clinical and angiographic correlation with a revised classification of venous drainage.Radiology 1995;194:671–680.2.Borden JA,Wu JK,Shucart WA.A proposed classificationfor spinal and cranial dural arteriovenous fistulous malfor-mations and implications for treatment.J Neurosurg 1995;82:166–179.3.Houdart E,Gobin YP,Casasco A,Aymard A,Herbreteau D,Merland JJ.A proposed angiographic clas-sification of intracranial arteriovenous fistulae and malfor-mations.Neuroradiology1993;35:381–385.4.Lucas CdP,Caldas JG,Prandini MN.Do leptomeningealvenous drainage and dysplastic venous dilation predicthemorrhage in dural arteriovenous fistula?Surg Neurol 2006;66(suppl3):S2–S5.5.Singh V,Smith WS,Lawton MT,Halbach VV,Young WL.Risk factors for hemorrhagic presentation in patients with dural arteriovenous fistulae.Neurosurgery2008;62:628–635.6.Piippo A,Laakso A,Seppa K,et al.Early and long-termexcess mortality in227patients with intracranial dural arteriovenous fistulas.J Neurosurg2013;119:164–171.7.Willinsky R,Terbrugge K,Montanera W,Mikulis D,Wallace MC.Venous congestion:an MR finding in dural arteriovenous malformations with cortical venous drain-age.Am J Neuroradiol1994;15:1501–1507.sjaunias P,Chiu M,ter Brugge K,Tolia A,Hurth M,Bernstein M.Neurological manifestations of intracranial dural arteriovenous malformations.J Neurosurg1986;64: 724–730.sjaunias P,TerBrugge K,Chiu M.Dural AVM.Neurosurgery1985;16:435–436.10.Davies MA,Saleh J,Ter Brugge K,Willinsky R,Wallace MC.The natural history and management of intracranial dural arteriovenous fistulae:part1:benign lesions.Interv Neuroradiol1997;3:295–302.11.Davies MA,Ter Brugge K,Willinsky R,Wallace MC.Thenatural history and management of intracranial dural arte-riovenous fistulae:part2:aggressive lesions.Interv Neuro-radiol1997;3:303–311.12.Duvernoy HM.Vascularization of the cerebral cortex[inFrench].Rev Neurol1999;155:684–687.13.Sulter G,Steen C,De Keyser e of the Barthel Indexand modified Rankin Scale in acute stroke trials.Stroke 1999;30:1538–1541.14.Chung SJ,Kim JS,Kim JC,et al.Intracranial dural arte-riovenous fistulas:analysis of60patients.Cerebrovasc Dis 2002;13:79–88.15.Suh DC,Lee JH,Kim SJ,et al.New concept in cavernoussinus dural arteriovenous fistula:correlation with present-ing symptom and venous drainage patterns.Stroke2005;36:1134–1139.16.Heinze G,Schemper M.A solution to the problem ofseparation in logistic regression.Stat Med2002;21: 2409–2419.17.Heinze G,Ploner M,Dunkler D,Southworth H.Logistf:Firth’s bias reduced logistic regression.R package version1.21[online].Available at:/package5logistf.Accessed December5,2013.18.Choi BS,Park JW,Kim JL,et al.Treatment strategy basedon multimodal management outcome of cavernous sinus dural arteriovenous fistula(CSDAVF).Neurointervention 2011;6:6–12.19.van Dijk JMC,terBrugge KG,Willinsky RA,Wallace MC.Clinical course of cranial dural arteriovenous fistulas with long-term persistent cortical venous reflux.Stroke2002;33:1233–1236.20.Daniels DJ,Vellimana AK,Zipfel GJ,Lanzino G.Intra-cranial hemorrhage from dural arteriovenous fistulas:clin-ical features and outcome.Neurosurg Focus2013;34:E15.21.Davies MA,TerBrugge K,Willinsky R,Coyne T,Saleh J,Wallace MC.The validity of classification for the clinical presentation of intracranial dural arteriovenous fistulas.J Neurosurg1996;85:830–837.22.Leach JL,Fortuna RB,Jones BV,Gaskill-Shipley MF.Imaging of cerebral venous thrombosis:current techni-ques,spectrum of findings,and diagnostic pitfalls.Radio-graphics2006;26(suppl1):S19–S41.Neurology82May27,20141903。

岩上窦区硬脑膜动静脉瘘诊治一例王全;李静伟;孙力泳;陈圣攀;李桂林;张鸿祺【期刊名称】《中国脑血管病杂志》【年(卷),期】2015(000)010【总页数】3页(P534-536)【关键词】硬脑膜动静脉瘘;脑膜垂体干;岩静脉【作者】王全;李静伟;孙力泳;陈圣攀;李桂林;张鸿祺【作者单位】内蒙古自治区赤峰市宝山医院神经外科024076;100053北京，首都医科大学宣武医院神经外科;100053北京，首都医科大学宣武医院神经外科;100053北京，首都医科大学宣武医院神经外科;100053北京，首都医科大学宣武医院神经外科;100053北京，首都医科大学宣武医院神经外科【正文语种】中文患者女，54岁，主因“颈部疼痛1个月余，四肢无力伴二便障碍20 d，加重1周”于2015年5月18日入住首都医科大学宣武医院。

患者1个月前无明显诱因出现颈枕部持续性胀痛，未予以诊治，20 d前出现四肢无力伴二便困难，就诊于当地医院，诊断为“颈椎病”，接受药物保守治疗后症状无缓解。

1周前就诊于外院神经内科，考虑为“急性脊髓炎”而接受大剂量激素[注射用甲泼尼龙琥珀酸钠(商品名：甲强龙)1 g/d]冲击治疗2 d后，四肢肌力明显下降，并伴二便无法自行排出，为进一步诊治入住本院。

入院体格检查示：颈后部压痛，胸7以下浅深感觉减退，腹壁反射减退，双上肢肌力Ⅲ级，左下肢肌力Ⅱ- 级，右下肢肌力Ⅱ级。

双侧霍夫曼征阳性，双侧巴宾斯基征阳性。

术前头颈部MRI示延髓至胸3椎体水平脊髓明显增粗水肿，伴延颈髓腹侧髓周流空影(图1a)。

三维时间飞跃法(time of flight, TOF)MR血管成像(MRA)示迂曲扩张的脑膜垂体干小脑幕缘支起自右侧颈内动脉海绵窦段并沿小脑幕缘行至岩尖部(图1b)。

入院后于局部麻醉下行全脑DSA，右侧颈内动脉造影示瘘口位于岩尖部，由脑膜垂体干小脑幕缘支供血并向岩上静脉引流，引流静脉沿脑干小脑侧方向下走行，经寰枕交界区后汇入脊髓前、后静脉并向下引流(图2)。