Neurologic complications
0 likes40 views
This study prospectively analyzed 2,899 cerebral angiography procedures to identify risk factors for neurologic complications. The main findings were: - There was a 1.3% overall neurologic complication rate, with 0.7% being transient, 0.2% reversible, and 0.5% permanent. - Neurologic complications were more common in patients 55 or older (1.8% rate) and those with cardiovascular disease (2.3% rate). - Longer fluoroscopy times of 10 minutes or more were also associated with higher neurologic complication rates (1.9% rate). - Fellows alone performed most procedures (1.878) and had a 1.3%
Neurologic complications
- 1. Robert A. Willinsky, MD, FRCPC Steve M. Taylor, BA Karel terBrugge, MD, FRCPC Richard I. Farb, MD, FRCPC George Tomlinson, PhD Walter Montanera, MD, FRCPC Index terms: Cerebral angiography, complications, 17.124, 17.44 Cerebral angiography, technology, 17.124 Published online before print 10.1148/radiol.2272012071 Radiology 2003; 227:522–528 Abbreviations: CVD ϭ cardiovascular disease DSA ϭ digital subtraction angiography 1 From the Department of Medical Im- aging, Toronto Western Hospital, Uni- versity Health Network, Fell Pavilion 3-210, 399 Bathurst St, Toronto, On- tario, Canada M5T 2S8. Received De- cember 19, 2001; revision requested February 27, 2002; final revision re- ceived August 7; accepted August 21. Supported by the Canadian Institutes of Health Research Burroughs Well- come Fund Research Award and the University of Toronto Summer Re- search Scholarship Program. Address correspondence to R.A.W. (e-mail: robert.willinsky@uhn.on.ca). Author contributions: Guarantor of integrity of entire study, R.A.W.; study concepts and design, R.A.W.; literature research, R.A.W., S.M.T.; clinical studies, R.A.W., R.I.F., K.t.B., W.M.; data acquisition, R.A.W., K.t.B., R.I.F., W.M.; data analysis/inter- pretation, R.A.W., S.M.T., G.T.; statis- tical analysis, G.T.; manuscript prepa- ration, R.A.W., S.M.T.; manuscript editing, R.A.W., S.M.T., K.t.B., R.I.F.; manuscript revision/review, R.A.W., S.M.T., K.t.B., R.I.F., G.T.; manuscript definition of intellectual content and fi- nal version approval, R.A.W., K.t.B., R.I.F. © RSNA, 2003 Neurologic Complications of Cerebral Angiography: Prospective Analysis of 2,899 Procedures and Review of the Literature1 PURPOSE: To prospectively identify risk factors for neurologic complications related to cerebral angiography. MATERIALS AND METHODS: A total of 2,899 consecutive cerebral digital subtraction angiograms obtained with nonionic contrast material were prospectively evaluated. Neurologic complications were categorized as transient (Ͻ24 hours), reversible (24 hours to 7 days), and permanent (Ͼ7 days). The neurologic complication rate was corre- lated with patient age, type of indication for catheter angiography, medical history, fluoro- scopic time, number and size of catheters, type and number of vessels injected, operator experience, and the quartile in which the study was performed. The correlations were statistically analyzed with Fisher exact tests and a multiple logistic regression model. RESULTS: There were 39 (1.3%) neurologic complications in 2,899 procedures; 20 were transient (0.7%), five (0.2%) were reversible, and 14 (0.5%) were permanent. Neurologic complications were significantly more common in patients 55 years of age or older (25 of 1,361; 1.8%) (P ϭ .035), in patients with cardiovascular disease (CVD) (20 of 862; 2.3%) (P ϭ .004), and when fluoroscopic times were 10 minutes or longer (24 of 1,238; 1.9%) (P ϭ .022). The neurologic complication rate was higher in procedures performed by fellows alone (24 of 1,878; 1.3%) compared with that when staff alone performed the procedures (three of 598; 0.5%), but the difference was not significant (P ϭ .172). Neurologic complications were lower in the fourth quartile of the study (six of 171; 0.9%) compared with the first quartile (16 of 776; 2.1%), which was likely due to fewer patients being examined for carotid stenosis or ischemic stroke and fewer patients with CVD (P ϭ .085). CONCLUSION: Age-related vascular disease accounted for the failure to lower the neurologic complication rate of cerebral angiography despite technical advances. © RSNA, 2003 Noninvasive imaging of the craniocervical vessels has dramatically improved during the past decade (1–4). Despite advances, cerebral angiography continues to be used for the examination of patients with cerebrovascular diseases. In the past decade, safer contrast agents have been used and there have been important technical advances including smaller catheters, hydrophylic guide wires, and digital imaging systems (5–7). The purpose of this prospective study was to identify the risk factors for neurologic complications that are related to cerebral angiography. MATERIALS AND METHODS Patients Between January 1996 and June 2001, 2,900 consecutive diagnostic cerebral angiograms obtained at one institution were studied prospectively, with institutional review board Neuroradiology 522 Radiology
- 2. approval and patient informed consent. All studies were performed on basis of accepted clinical indications for treat- ment. One procedure was excluded since it was performed to confirm brain death. Diagnostic cerebral angiograms obtained as part of neurointerventional endovas- cular procedures were not included in this study. Three diagnostic procedures were complicated by thromboembolism, which prompted immediate direct in- traarterial thrombolysis with a micro- catheter. The three procedures were in- cluded in this report. Technique Patients were restricted from having solid foods 6 hours before the procedure but were allowed clear fluids. A question- naire that included past medical health was filled out for all outpatients. The out- patient procedures were performed in the morning, and the patients were observed for a minimum of 4 hours until dis- charge. A nurse monitored the vital signs and neurologic status of the outpatients until discharge. A family member or a friend accompanied all outpatients over- night. Outpatients who developed com- plications within 4 hours of the proce- dure were admitted. In all patients, an intravenous catheter was placed prior to angiography. Electro- cardiography, pulse oximetry, and vital signs were monitored throughout the procedure. Femoral arterial punctures were performed in all but one procedure, which was performed by using a retro- grade brachial approach. A 5-F sheath (Radiofocus Introducer II; Terumo, To- kyo, Japan) was used in all femoral punc- tures. The sheath was constantly flushed with heparinized saline (6,000 IU of hep- arin in 1,000 mL of normal saline). A similar heparinized saline solution was used for intermittent flushing of the catheter. In 51 procedures, a bolus of 2,000–2,500 IU of heparin was adminis- tered at the beginning of the procedure and was not repeated or reversed. Manual compression at the puncture site for 8–15 minutes was performed at the com- pletion of all procedures. An angled 0.35-inch radiofocus guide wire (Terumo) was used in 2,704 cases; a 0.35-inch heparin-coated guide wire (Cook, Bloomington, Ind), in 167 cases; and both guide wires, in 28 cases. A one- way stopcock (Cook) was used with the selective catheter. Intermittent flushing techniques were used to prevent clots from developing in the catheter. Special attention was devoted to avoiding any dead-space within the catheter when the guide wire was used. Manipulation with the guide wire was generally less than 60 seconds between catheter flushing. Nonionic contrast media (Omnipaque 300; Nycomed, Oslo, Norway) were used in all cases. All injections in the angiog- raphy suite were performed with a power injector (Mark V; Medrad, Indianola, Pa). The intraoperative procedures were per- formed with hand injections. Standard injection rates and volumes were as fol- lows: 4–6 mL/sec for 8–12 mL for com- mon carotid artery, 4–5 mL/sec for 8–10 mL for internal carotid artery, 2–3 mL/sec for 5 mL for external carotid artery, 3–4 mL/sec for 7–9 mL for vertebral artery, 6–8 mL/sec for 14 mL for subclavian ar- tery, and 15–20 mL/sec for 30–40 mL for aortic arch. The distribution of the cath- eterized arteries was as follows: 1,649 (56.9%) right internal carotid arteries, 1,663 (57.3%) left internal carotid arter- ies, 1,502 (51.8%) left vertebral arteries, 957 (33.0%) right vertebral arteries, 945 (32.6%) right common carotid arteries, 968 (33.4%) left common carotid arter- ies, 607 (20.9%) right external carotid ar- teries, 611 (21.1%) left external carotid arteries, 237 (8.2%) right subclavian ar- teries, and 255 (8.8%) left subclavian ar- teries. In 220 procedures (7.6%), selective injections of additional vessels, including occipital, ascending pharyngeal, facial, and lingual arteries, were performed. From January 1996 to December 1998, 1,533 procedures were performed with single-plane digital subtraction angiogra- phy (DSA) (LUA; GE Medical Systems, Milwaukee, Wis). Between January 1999 and June 2001, 1,245 cases were per- formed with biplane DSA (LCN; GE Med- ical Systems) and 71, with single-plane DSA. Fifty intraoperative angiograms were obtained with a DSA unit (OEC 9800; GE Medical Systems). Neuroradiol- ogy fellows and staff performed the pro- cedures. Fellows were allowed to perform the procedures on their own once it was clear that they had reached a safe level of using the technique. A staff neuroradi- ologist (R.A.W., R.I.F., K.t.B., W.M.) su- pervised the procedures performed by the fellows alone. Procedures performed by both staff and fellows were performed early during the fellow’s training or when the fellow was having difficulty with a part of the procedure. Data Collection At the beginning of the procedure, a technologist prepared the data forms and filled in the patient demographics. At the completion of the procedure, the tech- nologist filled in the contrast material volume and the fluoroscopic time. The angiographer filled out the clinical infor- mation. The data forms included the date of the procedure, name of the referring physician, inpatient versus outpatient sta- tus, reason for the procedure (hemorrhage, aneurysm, arteriovenous malformation, carotid stenosis, ischemic stroke, sodium amytal test results, tumor, trauma), med- ical history (not significant, coronary ar- tery disease, peripheral vascular disease, hypertension, renal disease, diabetes melli- tus), catheter size and type, type and num- ber of vessels injected, type of anesthesia (local, neuroleptic, general), heparin bo- lus (dose), complications within 4 hours (none, neurologic, nonneurologic), com- plications within 24 hours (none, neuro- logic, nonneurologic), adverse events, and operator experience. Outpatients were con- tacted by phone on the next working day. Inpatients were evaluated the next day by the attending staff and the next work- ing day by the angiographer. A neurologic complication was defined as any new neurologic sign or symptom or worsening of a preexisting neurologic deficit that occurred during the proce- dure or within 24 hours. Neurologic com- plications were classified as transient if they resolved within 24 hours, reversible if they lasted more than 24 hours and resolved within 7 days, and permanent if they persisted more than 7 days. A non- neurologic complication was defined as any sign or symptom occurring either lo- cally at the puncture site or systemically within 24 hours of the procedure. An ad- verse event was defined as a procedure- related angiographic finding without neurologic signs or symptoms that may or may not have required treatment (eg, arterial dissection). The exact nature of the complication was clarified on the data form. Hematomas at the puncture site were recorded as small if they were less than 5 cm in diameter and large if they were 5 cm or larger in diameter. Adverse events and their management were documented on the data sheet. A chart review approved by the institu- tional review board was performed (R.A.W., S.M.T.) for patients with neuro- logic and nonneurologic complications that persisted more than 24 hours. Statistical Analysis For the analysis (R.A.W., S.M.T., G.T.) of preexisting comorbidity, coronary ar- tery disease, hypertension, and periph- eral vascular disease were grouped as car- Volume 227 ⅐ Number 2 Neurologic Complications of Cerebral Angiography ⅐ 523 Radiology
- 3. diovascular disease (CVD). All patients with multiple preexisting comorbidities had at least one of the CVDs; the diseases were grouped as cardiovascular. For the analysis of the indication for catheter an- giography, carotid stenosis and ischemic stroke were grouped together. The Fisher exact test was used to test associations between the neurologic complication rate and patient age, indication for cath- eter angiography, medical history, fluo- roscopic time, type of vessels injected, number of vessels injected, catheter size, number of catheters, operator experi- ence, and the quartile in which the pro- cedure was performed (R.A.W., S.M.T., G.T.). The relationships between the complication rate and the risk factors that were significant in these bivariate analyses were analyzed (G.T.) with a mul- tiple logistic regression model to deter- mine which factors were independently associated with complications. Interac- tions between individual cardiovascular risk factors could not be performed with regression models, since we could not ad- here to the widely recognized principles of regression analysis that recommend as a guideline a minimum of 10 outcomes per parameter. RESULTS A total of 1,368 (47.1%) procedures were performed in female patients and 1,531 (52.8%) were performed in male patients. The age range was 5–91 years, with a mean age of 52.5 years (Table 1). A total of 1,344 procedures (46.3%) were per- formed as outpatient and 1,555 (53.7%) were performed as inpatient; 2,777 (95.8%) were performed with local anes- thesia; 61 (2.1%), with neuroleptic anes- thesia; and 61 (2.1%), with general anes- thesia. The medical history and indications for catheter angiography are outlined in Tables 2 and 3. The mean fluoroscopic time was 11.5 minutes (median time, 8.6 minutes). The mean contrast material volume was 114.7 mL. A bolus of 2,000–2,500 IU of heparin was administered in 51 (1.8%) patients. A total of 2,589 (89.3%) proce- dures were performed with 5-F catheters; 279 (9.6%), with 4-F catheters; five (0.2%), with 6-F catheters; one (0.03%), with a 7-F catheter; and 25 (0.9%), each with more than one size catheter. A mul- tipurpose hockey-stick curved catheter (Bern 1 Performa; Merit Medical, South Jordan, Utah) was used in 2,414 (83.3%) procedures; a Sidewinder catheter (Sim 2; Merit Medical), in 228 (7.9%); a Mani cerebral catheter (Cook), in 29 (1.0%); and more than one type of catheter, in 228 (7.9%). A total of 1,888 patients had one angiogram, 344 had two, 78 had three, 16 had four, and five had five. One to three vessels were studied in 1,621 (55.9%) procedures; four to six vessels, in 1,138 (39.2%); and seven or more vessels, in 140 (4.8%). Neurologic Complications There were 39 (1.3%) neurologic com- plications in 2,899 procedures; 20 were transient (0.7%), five (0.2%) were revers- ible, and 14 (0.5%) were permanent. Seven (0.25%) of 2,899 permanent com- plications were considered major or dis- abling strokes. In the outpatient group, all neurologic complications were evi- dent within 4 hours of the procedure. Two of the transient and one of the per- manent complications were related to thromboembolism, which was evident on the angiogram during the procedure. All three were treated with intraarterial thrombolysis. Of 20 transient complica- tions, six had the typical findings and course of transient global amnesia. Neurologic complications were signifi- cantly more common in procedures per- formed in patients who were 55 years of age or older (25 of 1,361; 1.8%) compared with those performed in patients younger than 55 years of age (14 of 1,538, 0.9%) (P ϭ .035) (Table 1). From a logistic regres- sion against age, the estimated odds ratio for neurologic complications was 1.22 per 10 years of age (95% CI: 1.00, 1.52; P ϭ .05). Neurologic complications were signif- icantly more common in patients with CVD (20 of 862; 2.3%) compared with those without these risk factors (19 of 2,037; 0.9%) (P ϭ .004) (Table 2). When the fluoroscopic times were 10 minutes or longer, there were significantly more neu- rologic complications (24 of 1,238; 1.9%) compared with those when fluoroscopic times were shorter than 10 minutes (15 of 1,661; 0.9%) (P ϭ .004). Multiple logistic regression was used to analyze which of the factors (age, CVD, and fluoroscopic time) were indepen- dently related to the neurologic compli- cation rate. There were eight (2 ϫ 2 ϫ 2) regression models that could be formed by including or excluding each of these three factors. Each of these models was fitted, and the best model was chosen on the basis of likelihood ratio tests. The best model was one that included only fluoroscopic time and CVD. In that model, the estimated relative risk for neurologic complications associated with CVD was 2.32 (95% CI: 1.22, 4.43; P ϭ .010). The estimated relative risk for neu- rologic complications associated with fluoroscopic times longer than 10 min- utes was 1.97 (95% CI: 1.00, 3.87; P ϭ .039). Addition of age to this model did not significantly improve its fit (P ϭ .40). The fellows alone had a higher neuro- logic complication rate (24 of 1,878; 1.3%) compared with that of the staff alone (three of 598; 0.5%), but this did not reach significance (P ϭ .172) (Table 4). When the fellow and staff together performed the procedures, there was a significantly higher neurologic complica- tion rate (12 of 423; 2.8%) compared with that when a fellow or staff alone performed the procedures (27 of 2,476; 1.0%) (P ϭ .009). There were differences in the neuro- logic complication rates when they were analyzed for an association with age, in- dication for catheter angiography, type and number of vessels studied, number of catheters, catheter size, and the quar- tile in which the study was performed, TABLE 1 Age as a Risk Factor for Neurologic Complication Patient Age (y) No. of Procedures Neurologic Complication Transient Reversible Permanent Total 5–14 16 0 0 0 0 15–24 121 1 (0.8) 0 1 (0.8) 2 (1.6) 25–34 254 2 (0.8) 0 0 2 (0.8) 35–44 531 4 (0.8) 0 0 4 (0.8) 45–54 616 3 (0.5) 2 (0.3) 1 (0.2) 6 (1.0) 55–64 607 6 (1.0) 1 (0.2) 3 (0.5) 10 (1.6) 65–74 547 4 (0.7) 1 (0.2) 7 (1.3) 12 (2.2) 75–84 189 0 1 (0.5) 2 (1.1) 3 (1.6) 85–91 18 0 0 0 0 Total 2,899 20 (0.7) 5 (0.2) 14 (0.5) 39 (1.3) Note.—Data in parentheses are percentages. 524 ⅐ Radiology ⅐ May 2003 Willinsky et al Radiology
- 4. but these were not significant (Table 5). The neurologic complication rate was similar in male (17 of 1,367; 1.2%) and female (22 of 1,532; 1.4%) patients (P ϭ .747). Neurologic complications were more common when the indication for catheter angiography was carotid steno- sis or ischemic stroke (17 of 965; 1.8%) compared with that of other indications (22 of 1,934; 1.1%) (P ϭ .174) (Table 3). When four or more vessels were injected, the neurologic complication rate was 1.4% (18 of 1,278) compared with 1.2% (21 of 1,621) when fewer than four ves- sels were injected (P ϭ .871). When the posterior circulation was studied (verte- bral artery or subclavian artery injec- tions), the neurologic complication rate was higher (27 of 1,798; 1.5%) compared with that when the posterior circulation was not studied (12 of 1,101; 1.1%) (P ϭ .408). When the right vertebral artery was injected, the neurologic complication rate was higher (17 of 1,082; 1.6%) compared with that when the right vertebral artery was not injected (22 of 1,817; 1.2%) (P ϭ .41). When more than one catheter was used, there was a higher neurologic com- plication rate (five of 245; 2.0%) compared with that when one catheter was used (34 of 2,654; 1.3%) (P ϭ .375). When 4-F cath- eters were used, there was a lower neuro- logic complication rate (two of 279; 0.7%) compared with that when 5-F catheters were used (37 of 2,589; 1.4%) (P ϭ .581). A lower neurologic complication rate TABLE 5 Summary of the Statistical Analysis of Risk Factors Risk Factor P Value Significant (P Յ .05) CVD .0045 Fluoroscopic time Ն 10 minutes .0216 Age Ն 55 years .0354 Nonsignificant (P Ͼ .05) Operator experience .1708 Carotid stenosis or ischemic stroke .1741 No. of catheters .3748 Injection of posterior fossa .4084 Injection of right vertebral artery .4143 Catheter size .5814 Sex .7474 No. of vessels .8714 TABLE 2 Medical History as a Risk Factor for Neurologic Complication Medical History No. of Procedures Mean Age (y) Mean No. of Catheters Mean Fluoroscopic Time (min) Neurologic Complication Transient Reversible Permanent Total None 1,990 (68.6) 48.0 1.05 10.7 14 (0.7) 2 (0.1) 3 (0.2) 19 (1.0) CVD 862 (29.7) 62.7 1.22 13.8 6 (0.7) 3 (0.3) 11 (1.3) 20 (2.3) Renal disease 11 (0.4) 47.4 1.00 12.9 0 0 0 0 Diabetes mellitus 36 (1.2) 59.5 1.06 11.4 0 0 0 0 Note.—Data in parentheses are percentages. TABLE 3 Indication for Carotid Angiography as a Risk Factor for Neurologic Complication Indication No. of Procedures Mean Age (y) Mean No. of Catheters Mean Fluoroscopic Time (min) Neurologic Complication Transient Reversible Permanent Total Aneurysm 794 (27.4) 50.3 1.06 9.7 9 (1.1) 0 1 (0.03) 10 (1.3) Arteriovenous malformation 541 (18.7) 45.8 1.06 14.2 2 (0.4) 0 3 (0.6) 5 (0.9) Carotid stenosis or ischemic stroke 965 (34.3) 59.0 1.16 10.4 7 (0.7) 4 (0.4) 6 (0.6) 17 (1.8) Hemorrhage 499 (17.2) 51.8 1.09 13.6 2 (0.4) 0 3 (0.6) 5 (1.0) Sodium amytal test results 41 (1.4) 39.1 1.07 5.0 0 0 0 0 Tumor 56 (1.9) 53.0 1.21 15.9 0 1 (1.8) 1 (1.8) 2 (3.6) Trauma 3 (0.1) 33.0 1.00 5.4 0 0 0 0 Overall 2,899 52.5 1.10 11.5 20 (0.7) 5 (0.2) 14 (0.5) 39 (1.3) Note.—Data in parentheses are percentages. TABLE 4 Operator Experience as a Risk Factor for Neurologic Complication Operator No. of Procedures Mean Fluoroscopic Time (min) Mean Age (y) No. of Carotid Stenosis or Ischemic Stroke Complications No. of CVD Complications Neurologic Complication Transient Reversible Permanent Total Staff 598 (20.7) 10.2 52.1 179 (29.9) 40 (6.7) 2 (0.3) 0 1 (0.2) 3 (0.5) Fellow 1,878 (64.7) 10.9 52.1 620 (33.0) 564 (30.0) 12 (0.6) 4 (0.2) 8 (0.4) 24 (1.3) Staff and fellow 423 (14.6) 16.1 54.8 166 (39.2) 170 (40.2) 6 (1.4) 1 (0.2) 5 (1.2) 12 (2.8) Note.—Data in parentheses are percentages. Volume 227 ⅐ Number 2 Neurologic Complications of Cerebral Angiography ⅐ 525 Radiology
- 5. was found in the fourth quartile of the study (six of 671; 0.9%) compared with that in the first quartile (16 of 776; 2.1%) (P ϭ .085) (Table 6). Asymptomatic Adverse Events There were two asymptomatic arterial dissections. One was in the right verte- bral artery and required no treatment. The other was in the left internal carotid artery and was treated with systemic he- parinization. Nonneurologic Complications There were three allergic cutaneous re- actions in the 2,899 procedures (0.1%), including urticaria in two procedures and hives in one. These developed within 1 hour of the procedure and resolved with- out treatment. There were 14 hematomas (0.4%) at the puncture site; seven were small and seven were large. A pseudoan- eurysm was associated with one of the large hematomas and was treated with 15 minutes of compression with ultrasono- graphic (US) guidance. One of the proce- dures was complicated by a myocardial infarction. The patient remained in the hospital for 19 days and fully recovered. None of the 2,899 patients developed re- nal failure as a consequence of angiogra- phy. DISCUSSION Summary of the Literature The combined transient and reversible neurologic complication rate of cerebral an- giography has been reported to be as low as 0.4% and as high as 12.2% (6–18). The re- ported permanent neurologic complication rate varies from 0% to 5.4% (6–18). Limit- ing the review to prospective studies of 1,000 or more procedures reveals a com- bined transient and reversible neurologic complication rate from 0.4% to 2.3% (mean, 1.3%), a permanent neurologic complication rate from 0.1% to 0.5% (mean, 0.3%), and a mean overall rate of 1.6% (7–9,12). The series reported here had a slightly lower combined transient and re- versible neurologic complication rate of 0.9% (25 of 2,899), a slightly higher perma- nent rate of 0.5% (14 of 2,899), and a slightly lower overall rate of 1.3% (39 of 2,899). The prospective series reported here, with 2,899 procedures performed at one in- stitution, is, to our knowledge, the largest study of its kind to date. Authors of most of the studies who inves- tigated the complication rates of cerebral angiography have included complications that occurred within 24 hours of the proce- dure. The study of Dion et al (8) is unique in that it included complications that oc- curred 24–72 hours after the procedure. Factors that significantly correlated with in- creased neurologic events between 24 and 72 hours included the volume of contrast material, patient age, and diabetes. In pa- tients with frequent transient ischemic at- tacks and carotid stenosis, it is likely that at least some of these events are part of the natural history of the disease. Pathophysiology A number of mechanisms have been pro- posed to account for the neurologic compli- cations of cerebral angiography. The most common cause implicated is thromboem- bolism from the catheters or guide wires. These thrombi most likely develop inside the catheter during the manipulation of the guide wire. This is most likely to occur if the guide wire is withdrawn into the catheter, allowing blood to stagnate within this dead- space. In our practice, we stress the impor- tance of avoiding this dead-space and keep- ing the guide wire manipulations to a minimum time. Continual flushing of the catheter with heparinized saline with use of a three-way stopcock has been advocated to reduce thromboembolic events (8). In our practice, we have used one-way stopcocks, and we keep a full contrast material column in the catheter between manipulations. Disruption of an atherosclerotic plaque by the catheter or guide wire has been com- monly implicated as a mechanism for stroke. Other causes include arterial dissections re- lated to the catheter or guide wire, platelet activation, changes in clotting factors, and neurotoxicity of contrast agents (19–22). Patient Factors The neurologic complication rate has been reported to increase with patient age (8,9,12,13,15). In our series, the risk of neu- rologic complications increases a relative 22% for each 10-year increase in age (Table 1) (P ϭ .05). We found a significantly higher neurologic complication rate in procedures performed in patients 55 years of age or older. Heiserman et al (12), in their prospec- tive series of 1,000 patients, reported no neurologic complications in 363 patients who were younger than 50 years of age. Dion et al (8) reported no neurologic com- plications in 204 patients who were younger than 30 years of age. In our study, there was only one permanent complica- tion in patients younger than 45 years of age. This complication occurred in a teen- ager during intraoperative angiography. The high neurologic complication rate (one of 50; 2.0%) in our study with use of intra- operative angiograms may be a reflection of the difficulty with working in the operating room. In previous reports, the permanent neurologic complication rate from intraop- erative cerebral angiography was also noted to be higher, ranging from 0.5% to 1.5% (23–25). Ischemic stroke has been reported to be a risk factor in cerebral angiography (8–14,16,26,27). In a meta-analysis by Cloft et al (28), the neurologic complication rate was lower in patients with a subarachnoid hemorrhage, aneurysm, or arteriovenous malformation compared with that in pa- tients with ischemic stroke. Faught et al (11), in a retrospective study of 147 patients to evaluate stroke, found neurologic com- plications in 12.2% of the procedures, and 5.2% were permanent. Results of Earnest et al (9) were similar, with transient neuro- logic complications in 10.8% of patients with frequent transient ischemic attacks TABLE 6 Quartile Study as a Risk Factor for Neurologic Complication Quartile No. of Procedures DSA Mean Age (y) No. of Carotid Stenosis or Ischemic Stroke Complications No. of CVD Complications Mean Contrast Material Volume (mL) Neurologic Complication Transient Reversible Permanent Total First 776 Single-plane 54.0 321 (41.3) 258 (33.2) 125.8 7 (0.9) 2 (0.3) 7 (0.9) 16 (2.1) Second 779 Single-plane 52.2 298 (38.3) 213 (27.3) 118.6 7 (0.9) 1 (0.1) 1 (0.1) 9 (1.2) Third 673 Biplane 51.9 197 (29.3) 228 (33.4) 108.7 5 (0.7) 0 3 (0.4) 8 (1.2) Fourth 671 Biplane 51.7 149 (22.2) 163 (24.3) 103.2 1 (0.1) 2 (0.3) 3 (0.4) 6 (0.9) Note.—Data in parentheses are percentages. 526 ⅐ Radiology ⅐ May 2003 Willinsky et al Radiology
- 6. and in 9.8% of patients with recent stroke. Earnest et al did not find a significant dif- ference between the reversible (3.6%) and permanent (0.6%) neurologic complication rate in patients with recent ischemic events compared with their overall rates of 2.3% and 0.3%, respectively. In our study, pa- tients with ischemic stroke or carotid steno- sis had a higher neurologic complication rate (1.8%), but it was not significant (P ϭ .174). Heiserman et al (12) also found that carotid stenosis correlated with an in- creased rate of neurologic complications when compared with their entire group but not in the subset that had ischemic stroke. In 297 patients examined with Doppler US of the carotid artery, Earnest et al (9) found no difference in the neurologic complica- tion rate between those with carotid steno- sis and those without. Dion et al (8) found that hypertension was a significant risk factor for neurologic events developing within 24 hours of the procedure. Earnest et al (9) found that ele- vated levels of creatinine were associated with an increased incidence of neurologic complications. In our study, there was a significantly higher complication rate in pa- tients with CVD (2.3%) (Table 2). Procedural Factors Long procedural times have been impli- cated to be a risk factor in cerebral angiog- raphy (15). Heiserman et al (12) found signi- ficant correlations between the neurologic complication rate and patient age, length of the procedure, volume of contrast material, carotid stenosis, and ischemic stroke. Within the ischemic stroke group, age was the only significant variable (12). In our series, fluo- roscopic times of 10 minutes or longer were associated with a significantly higher neu- rologic complication rate. The fluoroscopic times were not dependant on whether a staff or fellow performed the procedure (Ta- ble 4). A higher neurologic complication rate did correlate with longer fluoroscopic times in procedures performed by both the fellow and staff. Fluoroscopic time was the best measure of the difficulty of the proce- dure and reflected the time taken to cathe- terize the blood vessels. A multiple logistic regression model was used to determine if the significant risk fac- tors (CVD, age older than 55 years, and fluoroscopic times longer than 10 minutes) were independently associated with a higher neurologic complication rate. CVD and fluoroscopic times were found to be independent predictors, whereas age was not a significant predictor once the cardio- vascular status and fluoroscopic time were known. Higher volume of contrast material has been considered a risk factor in cerebral an- giography performed with ionic contrast material (8,15). In a prospective evaluation of 230 patients with symptomatic cerebro- vascular disease, McIvor et al (16) found a slightly lower neurologic complication rate with use of nonionic contrast material com- pared with that with ionic contrast mate- rial, but it was not statistically significant. In a retrospective study of 1,509 procedures in which ionic contrast material was used and of 1,000 procedures in which nonionic contrast material was used, Skalpe (29) re- ported a lower neurologic complication rate in procedures performed with nonionic contrast material. The neurologic complica- tion rate in our study with use of nonionic contrast material was similar to that in large prospective studies in which ionic contrast material was used (8,9). Wishart (30) reported a higher neurologic complication rate when cerebral angiogra- phy includes selective injections in the pos- terior circulation. Later reports did not sup- port this finding (8,9,15). Our study findings did not show a significantly higher neurologic complication rate when the pos- terior fossa was injected compared with that with only carotid artery injections. We found no difference in the neurologic com- plication rate when it was correlated with the number of vessels injected, which is similar to the results of Dion et al (8). The association between the number of catheters used, the catheter size, and fluo- roscopic times with the neurologic compli- cation rate is complex owing to the rela- tionship to atherosclerotic disease. Earnest et al (9) found a higher rate of neurologic complications when more than one cathe- ter was required. Dion et al (8) did not find a difference until more than three catheters were used. Heiserman et al (12) found no association with the number of catheters used. Our study findings did not show a significantly higher neurologic complica- tion rate and the number of catheters used. Earnest et al (9) found a lower neurologic complication rate when smaller catheters were used, similar to our study results, but the rate was not significant. DSA has replaced cut-film changers in most centers. Procedural time has been re- duced with DSA compared with cut-film an- giography. Grzyska et al (7), in a large series in which cut-film angiography was used, attributed their lower neurologic complica- tion rate (0.4% transient, 0.1% permanent) to the use of DSA and nonionic iso-osmotic contrast material. Later studies in which DSA and nonionic contrast material were used demonstrated similar neurologic com- plication rates to those of series that were performed with cut-film angiography and ionic contrast material (8,9,12–14). Our study had a similar neurologic complication rate as did these large prospective studies, despite the use of DSA and nonionic con- trast material. During the 51⁄2 years of our study, we found that the neurologic complication rate had decreased from the first quartile to the fourth quartile (Table 6). This improvement correlated with the reduction in the per- centage of patients who were being treated for carotid stenosis or ischemic stroke and the lower percentage of patients with CVD. The similar neurologic complication rate in the second and third quartiles may reflect the similar percentage of patients with CVD. We believe that the lower neurologic complication rate during the period when biplane DSA was used compared with that when single-plane DSA was used was not related to the technology, since it was not evident when second and third quartiles were compared. As expected, the mean con- trast material volume decreased during the period when biplane DSA was used. Heparinized saline is the standard flush so- lution used during cerebral angiography. In addition to heparinized saline, Dion et al (8) used a heparin bolus of 2,000 IU in selected cases (32.3% of procedures). Authors of other large prospective studies did not use a heparin bolus (7,9,12). In our study, we used a heparin bolus in 51 (1.7%) of the 2,899 procedures that were considered as high risk. This included some patients with frequent transient ischemic attacks or a re- cent stroke thought to be due to vasculitis. In comparing our study findings with those in the report by Dion et al, there was a similar overall neurologic complication rate, but their permanent complication rate was lower, which suggests there may be a benefit with the more liberal use of heparin. The disadvantage of heparin is hematomas at the puncture site, 6.9% in the series of Dion et al compared with 0.5% in our series (8). Delayed hemorrhage at the puncture site is more common in patients who un- derwent systemic heparinization (31). Operator Experience The neurologic complication rate of cere- bral angiography has been reported to be lower in the hands of more experienced an- giographers (3,11,32). Mani et al (15) found a higher rate of neurologic complications in training hospitals (3.9%) compared with that in nontraining hospitals (0.9%). Mani et al also found that general radiologists or angiog- raphers had more than double the neurologic complication rate (1.8%) compared with that of fully trained neuroradiologists (0.7%). Findings of smaller prospective studies (16,32) also demonstrated signifi- cantly higher neurologic complication rates in trainees compared with those of consul- tant radiologists. Heiserman et al (12) found no difference in the neurologic complica- tion rate and the level of training. In our study, the neurologic complication rate in procedures performed by fellows was higher Volume 227 ⅐ Number 2 Neurologic Complications of Cerebral Angiography ⅐ 527 Radiology
- 7. than that in procedures performed by staff (Table 4). The difference can be explained in part by the higher percentage of patients with CVD in procedures performed by fel- lows. The mean fluoroscopic time and the mean patient age in the procedures per- formed by staff alone and fellows alone were similar. This finding differs from the report by Grzyska et al (7) who found that trainees use longer fluoroscopic times. In our study, procedures performed by both fellows and staff together had a signifi- cantly higher neurologic complication rate compared with that of staff or fellows alone (Table 4). In the majority of these procedures, the staff participated in the procedure once the fellow was having difficulty. The com- plexity of these cases is reflected by the longer fluoroscopic times. The procedures per- formed by fellows and staff together consisted of patients who were slightly older and with a higher percentage of CVD, carotid stenosis, and ischemic stroke. Technical advances in cerebral angiogra- phy have not overcome the patient-related risk factors associated with neurologic com- plications. The risk of neurologic complica- tions increases with age. CVD and fluoro- scopic times longer than 10 minutes were independent predictors of risk. These find- ings substantiate the argument that patients at a higher risk should undergo noninvasive imaging of the craniocervical vessels and that catheter angiography should be avoided. In patients without these risk factors, neurologic complications still occur and, therefore, the indications for catheter angiography should be limited. Our data support the shift to noninvasive imaging of the craniocervical vessels. Acknowledgments: We are thankful to Luc Harvey, RN, for his help in coordinating pa- tient follow-up. We thank Luc Harvey, RN, and Mary-Lou Montanera, RMT, for their work in data entry. We appreciate the efforts of our dedicated nursing team led by Herma Green- wood, RN, and our technologists in medical imaging led by Ian Findlay, RMT. We express thanks to the fellows who filled out the data forms and participated in the patient follow- up: Mark Hudon, MD, FRCPC, Foothills Hos- pital, Calgary, Alberta; Derek Emery, MD, FRCPC, University of Alberta, Edmonton; Manny Simantirakis, MD, FRCPC, Ottawa Civic Hospital, Ontario; Suzanne Laughlin, MD, FRCPC, Mt Sinai Hospital and University Health Network, Toronto, Ontario; Gary Re- dekop, MD, FRCSC, Vancouver General Hospi- tal, British Columbia; Dorothy Lazinski, MD, FRCPC, Mt Sinai Hospital and University Health Network, Toronto, Ontario; Ahmed Andeejani, MBBS, Kingston General Hospital, Ontario; Miriam Kim, MD, FRCSC, MeritCare Clinic Neuroscience, South Fargo, ND; William Miller, MD, FRCPC, Ottawa Civic Hospital, Ontario; Mayank Goyal, MBBS, MD, Ottawa Civic Hospital, Ontario; David Westman, MD, FRCPC, Medical College of Virginia, Rich- mond; Cheemun Lum, MD, FRCPC, Ottawa Civic Hospital, Ontario; Phil Porter, MD, FRCSC, Toronto Western Hospital, Ontario; Seon-Kyu Lee, MD, PhD, Toronto Western Hospital, Ontario; and Tommy Andersson, MD, PhD, Toronto Western Hospital, Ontario. References 1. Farb RI, McGregor C, Kim JK, et al. Intracranial arteriovenous malformations: real-time auto- triggered elliptic centric-ordered 3D gadolin- ium-enhanced MR angiography—initial as- sessment. Radiology 2001; 220:244–251. 2. White PM, Teasdale EM, Wardlaw JM, Easton V. Intracranial aneurysms: CT angiography and MR angiography for detection prospec- tive blinded comparison in a large patient cohort. Radiology 2001; 219:739–749. 3. Randoux B, Marro B, Koskas F, et al. Carotid artery stenosis: prospective comparison of CT, three-dimensional gadolinium-enhanced MR, and conventional angiography. Radiology 2001; 220:179–185. 4. Munera F, Soto JA, Palacio D, Velez SM, Me- dina E. Diagnosis of arterial injuries caused by penetrating trauma to the neck: compar- ison of helical CT angiography and conven- tional angiography. Radiology 2000; 216: 356–362. 5. Skalpe IO. Complications in cerebral angiog- raphy with iohexol (Omnipaque) and meglu- mine metrizoate (Isopaque cerebral). Neuro- radiology 1988; 30:69–72. 6. Waugh JR, Sacharias N. Arteriographic com- plications in the DSA era. Radiology 1992; 182:243–246. 7. Grzyska U, Freitag J, Zeumer H. Selective ce- rebral intraarterial DSA: complication rate and control of risk factors. Neuroradiology 1990; 32:296–299. 8. Dion JE, Gates PC, Fox AJ, Barnett HJ, Blom RJ. Clinical events following neuroangiogra- phy: a prospective study. Stroke 1987; 18: 997–1004. 9. Earnest F, Forbes G, Sandok BA, et al. Com- plications of cerebral angiography: prospec- tive assessment of risk. AJR Am J Roentgenol 1984; 142:247–253. 10. Eisenberg RL, Bank WO, Hedgcock MW. Neurologic complications of angiography in patients with critical stenosis of the carotid artery. Neurology 1980; 30:892–895. 11. Faught E, Trader SD, Hanna GR. Cerebral complications of angiography for transient ischemia and stroke: prediction of risk. Neu- rology 1979; 29:4–15. 12. Heiserman JE, Dean BL, Hodak JA, et al. Neu- rologic complications of cerebral angiography. AJNR Am J Neuroradiol 1994; 15:1401–1407. 13. Komiyama M, Yamanaka K, Nishikawa M, Izumi T. Prospective analysis of complica- tions of catheter cerebral angiography in the digital subtraction angiography and mag- netic resonance era. Neurol Med Chir (To- kyo) 1998; 38:534–539. 14. Leffers AM, Wagner A. Neurologic complica- tions of cerebral angiography: a retrospective study of complication rate and patient risk factors. Acta Radiol 2000; 41:204–210. 15. Mani RL, Eisenberg RL, McDonald EJ Jr, Pol- lock JA, Mani JR. Complications of catheter cerebral arteriography: analysis of 5,000 pro- cedures. I. Criteria and incidence. AJR Am J Roentgenol 1978; 131:861–865. 16. McIvor J, Steiner TJ, Perkin GD, Greenhalgh RM, Rose FC. Neurological morbidity of arch and carotid arteriography in cerebrovascular disease: the influence of contrast medium and radiologist. Br J Radiol 1987; 60:117–122. 17. Olivecrona H. Complications of cerebral an- giography. Neuroradiology 1977; 14:175–181. 18. Takahashi M, Kawanami H. Complications of catheter cerebral angiography: an analysis of 500 examinations. Acta Radiol Diagn (Stockh) 1972; 13:248–258. 19. Aspelin P. Effect of ionic and non-ionic con- trast media on morphology of human eryth- rocytes. Acta Radiol Diagn (Stockh) 1978; 19: 675–687. 20. Aspelin P. Effect of ionic and non-ionic con- trast media on whole blood viscosity, plasma viscosity and hematocrit in vitro. Acta Radiol Diagn (Stockh) 1978; 19:977–989. 21. Gawel M, Burkett M, Rose FC. Platelet acti- vation following cerebral angiography. Acta Neurol Scand 1980; 61:240–243. 22. Laerum F. Injurious effects of contrast media on human vascular endothelium. Invest Ra- diol 1985; 20:S98–S99. 23. Popadic A, Witzmann A, Amann T, et al. The value of intraoperative angiography in sur- gery of intracranial aneurysms: a prospective study in 126 patients. Neuroradiology 2001; 43:466–471. 24. Vitaz TW, Gaskill-Shipley M, Tomsick T, Tew JM Jr. Utility, safety, and accuracy of intra- operative angiography in the surgical treat- ment of aneurysms and arteriovenous mal- formations. AJNR Am J Neuroradiol 1999; 20:1457–1461. 25. Derdeyn CP, Moran CJ, Cross DT, Grubb RL Jr, Dacey RG Jr. Intraoperative digital sub- traction angiography: a review of 112 con- secutive examinations. AJNR Am J Neurora- diol 1995; 16:307–318. 26. Leow K, Murie JA. Cerebral angiography for cerebrovascular disease: the risks. Br J Surg 1988; 75:428–430. 27. Theodotou BC, Whaley R, Mahaley MS. Complications following transfemoral cere- bral angiography for cerebral ischemia: re- port of 159 angiograms and correlation with surgical risk. Surg Neurol 1987; 28:90–92. 28. Cloft HJ, Joseph GJ, Dion JE. Risk of cerebral angiography in patients with subarachnoid hemorrhage, cerebral aneurysm, and arterio- venous malformation: a meta-analysis. Stroke 1999; 30:317–320. 29. Skalpe IO. Complications in cerebral angiog- raphy with iohexol (Omnipaque) and me- glumine metrizoate (Isopaque cerebral). Neuroradiology 1988; 30:69–72. 30. Wishart DL. Complications in vertebral an- giography as compared to non-vertebral ce- rebral angiography in 447 studies. Am J Roentgenol Radium Ther Nucl Med 1971; 113:527–537. 31. Antonovic R, Rosch J, Dotter CT. The value of systemic arterial heparinization in trans- femoral angiography: a prospective study. Am J Roentgenol 1976; 127:223–225. 32. Davies KN, Humphrey PR. Complications of cerebral angiography in patients with symp- tomatic carotid territory ischaemia screened by carotid ultrasound. J Neurol Neurosurg Psychiatry 1993; 56:967–972. 528 ⅐ Radiology ⅐ May 2003 Willinsky et al Radiology