In 10% 30% of patients with spontaneous subarachnoid

Transcription

1 J Neurosurg 120:99 103, 2014 AANS, 2014 Repeat digital subtraction angiography after a negative baseline assessment in nonperimesencephalic subarachnoid hemorrhage: a pooled data meta-analysis A systematic review Nicolaas A. Bakker, M.D., Ph.D., 1 Rob J. M. Groen, M.D., Ph.D., 1 Mahrouz Foumani, M.D., 1 Maarten Uyttenboogaart, M.D., Ph.D., 2 Omid S. Eshghi, M.D., 3 Jan D. M. Metzemaekers, M.D., Ph.D., 1 Natasja Lammers, B.Sc., 1 Gert-Jan Luijckx, M.D., Ph.D., 2 and J. Marc C. Van Dijk, M.D., Ph.D. 1 Departments of 1 Neurosurgery, 2 Neurology, and 3 Radiology, University Medical Center Groningen, University of Groningen, The Netherlands Object. A repeat digital subtraction angiography (DSA) study of the cranial vasculature is routinely performed in patients with diffuse nonperimesencephalic subarachnoid hemorrhage (SAH) after negative baseline CT angiography (CTA) and DSA studies. However, DSA carries a low but substantial risk of neurological complications. Therefore, the authors evaluated the added value of repeat DSA in patients with initial angiographically negative diffuse nonperimesencephalic SAH. Methods. A systematic review of the contemporary literature was performed according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement. Studies from January 2000 onward were reviewed since imaging modalities have much improved over the last decade. A pooled analysis was conducted to identify the detection rate of repeat DSA. In addition, the diagnostic yield of repeat DSAs in a prospectively maintained single-center series of 1051 consecutive patients with SAH was added to the analysis. Results. An initial search of the literature yielded 179 studies, 8 of which met the selection criteria. Another 45 patients from the authors institution were included in the study, providing 368 patients eligible for the pooled analysis. In 37 patients (10.0%, 95% CI 7.4% 13.6%) an aneurysm was detected on repeat DSA. The timing of the repeat DSA varied from 1 to 6 weeks after the initial DSA. The use of 3D techniques was poorly described among these studies, and no direct comparisons between CTA and DSA were made. Conclusions. Repeat DSA is still warranted in patients with a diffuse nonperimesencephalic SAH and negative initial assessment. However, the exact timing of the repeat DSA is subject to debate. ( Key Words subarachnoid hemorrhage digital subtraction angiography perimesencephalic hemorrhage vascular disorders In 10% 30% of patients with spontaneous subarachnoid hemorrhage (SAH), the cause of bleeding is undetected on initial CT angiography (CTA) and catheter angiography (that is, digital subtraction angiography [DSA]) of the cranial vasculature in the acute phase. This patient group is challenging, as swift detection of the cause of bleeding is mandatory to avoid further complications, such as (often devastating) aneurysm rebleeding. Three subcategories of SAH patients with initial negative CTA and DSA findings can be identified: 1) patients with typical perimesencephalic SAH, 2) patients with SAH diagnosed based on CSF xanthochromia without SAH on CT, and 3) patients with a diffuse nonperimesencephalic (classic) SAH pattern. As regards Abbreviations used in this paper: CTA = CT angiography; DSA = digital subtraction angiography; MRA = MR angiography; PRISMA = Preferred Reporting Items for Systematic Reviews and Meta- Analyses; SAH = subarachnoid hemorrhage. the first category, many studies have demonstrated a very low additional diagnostic value for DSA, and evaluation via noninvasive CTA is recommended. 4,9,12,17 The same may hold true for the second category of patients with CSF xanthochromia. 11,15,18,19 The third category is the most challenging. Probable causes for initially negative findings in these patients may be a thrombosed cerebral aneurysm or severe vasospasm. Studies from the pre-cta era have shown that a repeat DSA study is valuable in these patients, with a diagnostic yield up to 46%. 10,11,18 It is therefore common practice to perform repeat DSA in such patients. Nevertheless, cranial DSA is an invasive procedure that carries a 0.5% 1.8% risk of neurological complications, with permanent deficit in 0.09% 0.5%. 3,8,22 In addition, nonneurological complications occur in 0.6% of patients, including groin hematoma, peripheral thromboembolism, transient hypotension, and arteriovenous fistulas. 3 Over the last decade, noninvasive imaging techniques 99

2 N. A. Bakker et al. (CT and MRI) have markedly improved. In many institutions, CTA has replaced DSA as the first-line imaging modality in patients with SAH because of its high diagnostic yield. 20,21 In cases of diagnostic uncertainty, however, DSA is still routinely performed, with improved detection given the introduction of 3D rotational angiography and high-resolution digital x-ray detectors. Our aim in the present systematic review and singlecenter study was to determine whether delayed repeat DSA in patients with a diffuse nonperimesencephalic bleeding pattern and a negative initial CTA and DSA study still has added diagnostic value outweighing its known complication risks. Methods Literature Search Three electronic databases were searched: MEDLINE (PubMed), CENTRAL (Cochrane Central Register of Controlled Trials), and Embase. We considered studies in all languages and searched the databases for studies published from January 2000 to May In addition, two authors (N.A.B. and J.M.C.V.D.) scrutinized the reference lists of all selected studies for additional studies. The MeSH (medical subject headings) terms were: SAH[All Fields] OR Subarachnoid[All Fields]) AND (repeat[all Fields] OR delayed[all Fields]) AND (( catheters [MeSH Terms] OR catheters [All Fields] OR catheter [All Fields]) OR DSA[All Fields]). Institutional review board approval was not required for this retrospective review. Studies assessing 5 or more patients were considered eligible for analysis, whereas those on patients with perimesencephalic SAH or CSF xanthochromia (without blood on CT) were excluded. Moreover, studies evaluating patients before January 1, 2000, were excluded, as imaging techniques have dramatically improved over the last few years. In cases of disagreement between the two authors regarding which studies should be included in the analysis, consensus was reached by discussion. Figure 1 shows the flowchart according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement. Two independent reviewers (N.A.B. and J.M.C.V.D.) assessed methodological quality by using specific study design related forms designed by the Dutch Cochrane Collaboration. The two reviewers independently assessed the included articles. Data Extraction The following data (if available) were extracted from all studies: 1) number of patients with a diffuse nonperimesencephalic bleeding pattern and negative initial CTA and DSA findings in whom repeat DSA was performed; 2) diagnostic yield of repeat DSA in terms of detecting the underlying vascular pathology probably responsible for SAH; 3) type of DSA performed (4- or 6-vessel, use of 3D rotational capabilities); 4) study timeframe; 5) time between initial and repeat DSA; 6) retrospective visibility of vascular pathology on the initial study; 7) type of imaging assessment (for example, retrospectively blinded, consensus meeting, and so forth); 8) consequences in terms of treatment initiation; and 9) type of study (for example, randomized controlled trial, observational cohort, and so forth). University Medical Center Groningen Series A consecutive cohort of 1051 patients with SAH in the period from January 2005 to December 2011, after the introduction of a 64-multidetector CT scanner and 3D DSA facility at the University Medical Center Groningen, the Netherlands, was assessed. All patient data had been prospectively collected in a database and were retrospectively analyzed. A subgroup of 45 patients (4%) with diffuse nonperimesencephalic (classic) SAH had a negative initial cranial CTA and DSA study. Patients with typical perimesencephalic SAH (n = 76) and those without detectable blood on CT (SAH diagnosis made based on CSF xanthochromia, n = 52) were excluded. No patients were lost to follow-up. The other 878 patients with classic aneurysmal SAH were excluded because the vascular pathology was already demonstrated on the initial CTA or DSA. Treatment Protocol. A multidisciplinary protocol was followed for all of the patients with SAH. All underwent immediate CTA on admission. If an aneurysm or vascular malformation was detected, treatment was administered as appropriate. In cases of negative CTA in the presence of a classic SAH or in cases of diagnostic doubts, patients underwent DSA within 24 hours. If DSA was negative as well, a repeat DSA study was routinely performed after 14 days. In addition, MRI and MR angiography (MRA) of the neuraxis were performed. A second repeat DSA was not routinely performed. A joint multidisciplinary team of interventional neuroradiologists, vascular neurosurgeons, and vascular neurologists evaluated all imaging results. Computed Tomography Angiography. All CTA examinations were performed on a 64-multidetector CT machine (Somatom Sensation 64, Siemens Medical Systems) using a standard protocol. Source images were transferred to a remote computer workstation (Odelft Benelux Diagnostic Imaging) for viewing. Two-dimensional maximum intensity projection views and 3D surface-rendered and volume-rendered reconstructions were reformatted from the raw image data on a Vitrea computer workstation by an experienced neuroradiologist. Digital Subtraction Angiography. Cerebral catheter angiography was performed using the Seldinger technique, with a 4- or 5-Fr introducer in the common femoral artery. Standard 6-vessel angiography with intracranial views (frontal, lateral) was completed using 3D rotational angiography (Visipaque [iodixanol] 270 contrast material, GE Healthcare BV). Statistical Analysis Continuous variables were expressed as the means ± standard deviation or the median with a range, and categorical variables were expressed as counts and percentages. The 95% confidence intervals of the proportion were calculated using the Wilson procedure without continuity correction. All analyses were performed using SPSS software (version 20.0, SPSS Inc.). 100

3 Repeat digital subtraction angiography negative SAH Fig. 1. Flowchart according to the PRISMA statement. Pooled Analysis Results An initial search of the literature yielded 179 studies (Fig. 1), 8 of which met the inclusion criteria. 1,2,5,7,13,15,16,23 Results for these studies are summarized in Table 1. Three hundred sixty-eight patients, including those from the Groningen series, were eligible for the pooled analysis. In 37 patients (10.0%, 95% CI 7.4% 13.6%), a presumed causative vascular anomaly (an aneurysm in all cases) was detected using repeat DSA. Literature Review Four-vessel (4 studies) or 6-vessel (3 studies) DSA was routinely performed (one study did not report the type of DSA performed) together with 3D reconstructions in 5 studies. In none of the studies was repeat CTA directly compared with repeat DSA. One study documented a blinded reassessment of images, in which symptomatic aneurysms were retrospectively seen on the initial CTA or DSA in 3 of 4 patients. The timing of the repeat DSA varied from 1 week up to 6 weeks, but it was not reported in 4 of the 8 studies. Complications of DSA were not reported in most studies. Of note, the delayed detection of an aneurysm had clinical consequences in all patients. University Medical Center Groningen Series In our cohort of 45 patients who had undergone repeat DSA 14 days after the initial DSA, no vascular abnormalities were detected. Moreover, MRI and/or MRA of the neuraxis did not reveal any vascular abnormalities. All patients had long-term follow-up, with a median of 38 months (range 3 84 months). One patient had rebleeding 2 months after the initial hemorrhage. At reevaluation, a small middle cerebral artery aneurysm was detected and 101

4 N. A. Bakker et al. subsequently treated using neurosurgical clipping. In retrospect, the aneurysm could not be detected on either the initial or the repeat DSA after the primary hemorrhage. Two patients (4%) had complications attributable to the repeat DSA: one with reversible sensory symptoms in the left hand, and one with persistent neurological deficits and ischemic lesions revealed by MRI. Discussion The current pooled analysis of repeat DSA in 368 patients with diffuse nonperimesencephalic SAH and a negative baseline assessment showed a diagnostic yield of 10.0%. It is essential to exclude a potential cause of bleeding in these patients, as a rebleed could have devastating consequences. Therefore, it has always been common practice to perform repeat DSA. This habit was largely based on studies performed before 2000, in which detection rates up to 46% have been reported with repeat DSA. 11 After 2000, according to our pooled analysis, the detection rates lowered because of the better initial diagnostic yield of improved CTA and 3D DSA techniques. However, given the possible consequences of missing detectable vascular pathology, repeat imaging is still justified in diffuse nonperimesencephalic SAH. A limitation of our analysis is the heterogeneity of the DSA techniques used in the included studies as well as the variety in the timing of repeat DSA (if it was reported at all). In some centers, 4-vessel instead of 6-vessel DSA was performed. Although their incidence is low, cranial dural arteriovenous fistulas might be missed with 4-vessel DSA. 14 We therefore propose always including projections of the external carotid arteries on the initial DSA. The timing of repeat DSA is also debatable. This was demonstrated in the study by Dalyai et al., 5 in which the first repeat DSA after 7 days revealed 10 aneurysms, but a second repeat DSA after 6 weeks revealed another 7 patients with a vascular abnormality. Nevertheless, it is difficult to draw any conclusions, as only a few studies have focused on the timing of repeat DSA. Given the vasospasm period and the chance of a thrombosed aneurysm, it is reasonable to perform a repeat DSA at least days after the ictus, thereby avoiding the need for a second repeat DSA. A negative initial DSA study is typically the result of a thrombosed aneurysm or severe vasospasm. In such cases, aneurysms might also be identified using repeat CTA. Although it has been reported that CTA has limitations with regard to small aneurysms located close to the skull base, a recent study on the diagnostic yield of repeat CTA compared with repeat MRA in diffuse nonperimesencephalic SAH revealed the superiority of CTA, with detection of small aneurysms (median 2.6 mm) in 9.3% of patients. 6 Furthermore, the retrospective study by Little et al. showed that 3 of 4 newly discovered aneurysms would have also been detected by CTA. 15 This finding is in line with those in a recent meta-analysis by Westerlaan et al., who demonstrated that high-resolution CTA might be considered an alternative to DSA as the primary imaging tool for detecting cerebral vascular abnormalities. 20,21 Nevertheless, studies directly comparing repeat CTA and TABLE 1: Literature summary of studies* No. of Cases w/ Treatment Initiation per Total Retrospective Visability of Vascular Pathology (no. of patients [%]) Imaging Assessment Timing of Repeat DSA No. of Repeat DSA Studies w/ Diagnostic Yield (%) No. of Patients Type of DSA Performed 3D? Study Timeframe Type of Study Authors & Year Andaluz & Zuccarello, 2008 retro vessel no 47 7 (14.9) NR NR NR 7/7 Delgado Almandoz et al., pros vessel yes 39 2 (5.1) 7 days panel, not blinded 1 (50) 2/2 Yu et al., 2012 retro NR NR 12 2 (16.7) NR NR NR NR Agid et al., 2010 retro vessel yes 28 4 (14) NR blinded 3 (75) NR Prestigiacomo et al., 2010 pros vessel yes 13 0 (0) NR NR NA NA Little et al., 2007 pros vessel yes 42 4/1 (12) 7 days; 6 wks NR NR NR Khan et al., 2013 retro vessel no 6 0 (0) 23 days panel, not blinded NA NA Dalyai et al., 2013 retro vessel yes /7 (12.5) 7 days; 6 wks not blinded NR 17/17 present study retro vessel yes days panel, not blinded NA NA * NA = not applicable; NR = not reported; pros = prospective; retro = retrospective. Two values represent the number of patients at the first and second repeat DSA. Multiple values indicate studies performed at different times. 102

5 Repeat digital subtraction angiography negative SAH repeat DSA are lacking. Further research on this topic is therefore warranted. Conclusions Although detection rates of repeat DSA have lowered over the last decade in patients with diffuse nonperimesencephalic SAH, the diagnostic yield of repeat DSA is still 10.0%. The timing and technique of repeat DSA is debatable. A direct prospective comparison between repeat CTA and repeat DSA is lacking. Disclosure The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. Author contributions to the study and manuscript preparation include the following. Conception and design: Bakker, Van Dijk. Acquisition of data: Bakker, Foumani, Lammers. Analysis and interpretation of data: Bakker. Drafting the article: Bakker, Van Dijk. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Bakker. Statistical analysis: Bakker. Study supervision: Groen. Performed DSA: Eshghi, Van Dijk. References 1. Agid R, Andersson T, Almqvist H, Willinsky RA, Lee SK, terbrugge KG, et al: Negative CT angiography findings in patients with spontaneous subarachnoid hemorrhage: When is digital subtraction angiography still needed? AJNR Am J Neuroradiol 31: , Andaluz N, Zuccarello M: Yield of further diagnostic workup of cryptogenic subarachnoid hemorrhage based on bleeding patterns on computed tomographic scans. Neurosurgery 62: , Cloft HJ, Joseph GJ, Dion JE: Risk of cerebral angiography in patients with subarachnoid hemorrhage, cerebral aneurysm, and arteriovenous malformation: a meta-analysis. Stroke 30: , Cruz JP, Sarma D, Noel de Tilly L: Perimesencephalic subarachnoid hemorrhage: when to stop imaging? Emerg Radiol 18: , Dalyai R, Chalouhi N, Theofanis T, Jabbour PM, Dumont AS, Gonzalez LF, et al: Subarachnoid hemorrhage with negative initial catheter angiography: a review of 254 cases evaluating patient clinical outcome and efficacy of short- and long-term repeat angiography. Neurosurgery 72: , Delgado Almandoz JE, Jagadeesan BD, Refai D, Moran CJ, Cross DT III, Chicoine MR, et al: Diagnostic yield of computed tomography angiography and magnetic resonance angiography in patients with catheter angiography-negative subarachnoid hemorrhage. Clinical article. J Neurosurg 117: , Delgado Almandoz JE, Jagadeesan BD, Refai D, Moran CJ, Cross DT III, Chicoine MR, et al: Diagnostic yield of repeat catheter angiography in patients with catheter and computed tomography angiography negative subarachnoid hemorrhage. Neurosurgery 70: , Grzyska U, Freitag J, Zeumer H: Selective cerebral intraarterial DSA. Complication rate and control of risk factors. Neuroradiology 32: , Huttner HB, Hartmann M, Köhrmann M, Neher M, Stippich C, Hähnel S, et al: Repeated digital substraction angiography after perimesencephalic subarachnoid hemorrhage? J Neuroradiol 33:87 89, Iwanaga H, Wakai S, Ochiai C, Narita J, Inoh S, Nagai M: Ruptured cerebral aneurysms missed by initial angiographic study. Neurosurgery 27:45 51, Jung JY, Kim YB, Lee JW, Huh SK, Lee KC: Spontaneous subarachnoid haemorrhage with negative initial angiography: a review of 143 cases. J Clin Neurosci 13: , Kershenovich A, Rappaport ZH, Maimon S: Brain computed tomography angiographic scans as the sole diagnostic examination for excluding aneurysms in patients with perimesencephalic subarachnoid hemorrhage. Neurosurgery 59: , Khan AA, Smith JD, Kirkman MA, Robertson FJ, Wong K, Dott C, et al: Angiogram negative subarachnoid haemorrhage: outcomes and the role of repeat angiography. Clin Neurol Neurosurg 115: , Lasjaunias P, Chiu M, ter Brugge K, Tolia A, Hurth M, Bernstein M: Neurological manifestations of intracranial dural arteriovenous malformations. J Neurosurg 64: , Little AS, Garrett M, Germain R, Farhataziz N, Albuquerque FC, McDougall CG, et al: Evaluation of patients with spontaneous subarachnoid hemorrhage and negative angiography. Neurosurgery 61: , Prestigiacomo CJ, Sabit A, He W, Jethwa P, Gandhi C, Russin J: Three dimensional CT angiography versus digital subtraction angiography in the detection of intracranial aneurysms in subarachnoid hemorrhage. J Neurointerv Surg 2: , Rinkel GJ, Wijdicks EF, Hasan D, Kienstra GE, Franke CL, Hageman LM, et al: Outcome in patients with subarachnoid haemorrhage and negative angiography according to pattern of haemorrhage on computed tomography. Lancet 338: , Rogg JM, Smeaton S, Doberstein C, Goldstein JH, Tung GA, Haas RA: Assessment of the value of MR imaging for examining patients with angiographically negative subarachnoid hemorrhage. AJR Am J Roentgenol 172: , Topcuoglu MA, Ogilvy CS, Carter BS, Buonanno FS, Koroshetz WJ, Singhal AB: Subarachnoid hemorrhage without evident cause on initial angiography studies: diagnostic yield of subsequent angiography and other neuroimaging tests. J Neurosurg 98: , Westerlaan HE, Gravendeel J, Fiore D, Metzemaekers JD, Groen RJ, Mooij JJ, et al: Multislice CT angiography in the selection of patients with ruptured intracranial aneurysms suitable for clipping or coiling. Neuroradiology 49: , Westerlaan HE, van Dijk JM, Jansen-van der Weide MC, de Groot JC, Groen RJ, Mooij JJ, et al: Intracranial aneurysms in patients with subarachnoid hemorrhage: CT angiography as a primary examination tool for diagnosis systematic review and meta-analysis. Radiology 258: , 2011 (Erratum in Radiology 260:612, 2011) 22. Willinsky RA, Taylor SM, TerBrugge K, Farb RI, Tomlinson G, Montanera W: Neurologic complications of cerebral angiography: prospective analysis of 2,899 procedures and review of the literature. Radiology 227: , Yu DW, Jung YJ, Choi BY, Chang CH: Subarachnoid hemorrhage with negative baseline digital subtraction angiography: is repeat digital subtraction angiography necessary? J Cerebrovasc Endovasc Neurosurg 14: , 2012 Manuscript submitted June 26, Accepted September 17, Please include this information when citing this paper: published online October 25, 2013; DOI: / JNS Address correspondence to: Nicolaas A. Bakker, M.D., Ph.D., Department of Neurosurgery AB71, University Medical Center Groningen, P.O. Box , 9700 RB Groningen, The Netherlands. n.a.bakker@umcg.nl. 103