Carotid artery occlusion: Natural history

Size: px

Start display at page:

Download "Carotid artery occlusion: Natural history"

Pierce Barnett
6 years ago
Views:

Carotid artery occlusion: Natural history Stephen C. NichoUs, M.D., Ted R. Kohler, M.D., Robert O. Bergelin, B.S., Jean F. Primozich, B.S., Ramona L. Lawrence, B.S., and D. E. Strandness, Jr., M.D., Seattle, Wash.

1 Carotid artery occlusion: Natural history Stephen C. NichoUs, M.D., Ted R. Kohler, M.D., Robert O. Bergelin, B.S., Jean F. Primozich, B.S., Ramona L. Lawrence, B.S., and D. E. Strandness, Jr., M.D., Seattle, Wash. During a 5-year period, 212 patients (170 men and 42 women, median age 65 and 64 years, respectively) were diagnosed as having internal carotid artery occlusion. Mean follow-up was 24.9 months. Five-year cumulative survival and stroke-free rates by lifetable analysis were 62% and 75%, respectively. Deaths were due to stroke in 7 of 40 patients (17%) and were of cardiac origin in 22 of 40 patients (55%). The strokes were ipsilateral in 20 of 31 patients (65%). No statistically significant difference between the sexes could be demonstrated for either death or stroke, nor was age correlated with stroke during follow-up. Diabetes and hypertension increased the risk of stroke, whereas gender and aspirin consumption had no discernible effect. Endarterectomy of the opposite carotid artery did not significantly affect the natural history but did reduce the stroke rate in the territory of the operated artery. Presenting symptoms were useful for estimating prognosis. Twenty-two of 111 patients referred for stroke (20%) suffered a further stroke and 21 of 111 patients (19%) died (three were stroke-related), whereas of those patients referred for transient ischemic attack (TIA), only 2 of 42 patients (5%) suffered a stroke and none died. TIAs occurred in 23 patients (11%) during follow-up, and these were premonitory for stroke in three cases (13%). The limited value of TIA in predicting stroke and the high mortality rate unrelated to stroke in this group are important considerations when therapy is considered for these patients. (J VASC SURG 1986; 4: ) The natural history of patients with internal carotid artery (ICA) occlusion is important when the role of contralateral endarterectomy is assessed. It has been reported that the prognosis after occlusion occurs is similar to that for lesser degrees of stenosis, although often an initial severe neurologic event has occurred. ~ The initial deficit that occurs with occlusion may be more severe when the contralateral ICA is normal than when it is diseased. 2 Much of the difficulty in documenting new neurologic events is due to difficulty in identifying and classifying such events against the background of often severe residual deficits. For this reason, some authors have included only those cases with mild initial neurologic deficits. However, since the long-term prognosis may depend on the severity of the initial neurologic insult, evaluation and categorization of all patients are needed. Of particular interest is the group of patients that is admitted to the hospital with "silent" occlusions. Most previous studies of this group have used postmortem or ICA ligation data. 3"4 However, ICA li- From the Division of Vascular Surgeu, Department of Surgcq,, Universiff of Washington School of Medicine. Presented at the First Annual Meeting of the Western Vascular Society,, Laguna Niguel, Calif., Jan , Reprint requests: S. C. Nicholls, M.D., Dept. of Surgery, Harborview Medical Ctr., 325 9th Ave., Seattle, WA gation is usually performed to treat trauma or intracerebral aneurysms rather than atherosclerosis and therefore may have a unique natural history. Although recent data suggest that the stroke rate after ICA occlusion occurs is no different from that of a matched population, s other studies report annual rates between 2% and 6%6,7--approximating the stroke rate occurring after transient ischemic attack (TIA).8 Although it might be expected that ipsilateral ischemic events would cease after ICA occlusion, 9 TIA and stroke distal to the occlusion are well documented. 7 Several mechanisms have been proposed for this--embolization through collateral channels from the ipsilateral common or external carotid arte~ or from the proximal "stump" of the occluded artery, t0 propagation ofthrombus to the circle of Willis or embolization from the occluded portion into pial arteries, H and, rarely, embolization from the contralateral artery through collateral branches. In fact, recent reports suggest that most neurologic events occurring after occlusion are on the same side6,7; this is the theoretic basis for extracranial-intracranial bypass grafting. Survival data must separate stroke-related death from that caused by other factors. Previous studies indicate that the principal cause of death in patients with TIAs is myocardial infarction and that carotid 479

2 480 Nicholls et al. Journal of VASCULAR SURGERY -._> lo~~k----~~. Female 0.8 = 0.6 > All c- O 0 c~ 13._ 0.4 i1 1'2 2'4 3'6 4'8 6'0 Male Fig. 1. Life-table analysis: cumulative survival rate by sex at 5 years. Men (n = 170), 57%; women (n = 42), 86%. No statistical difference demonstrated (p = 0.66). endarterectomy may reduce neurologic complications but cannot improve survival.~2 Survival data for patients with carotid occlusions are similar. 6a3'~4 We undertook a retrospective study of patients with carotid occlusion identified in our vascular laboratory to ascertain their risk for subsequent stroke or death. MATERIAL AND METHODS Of 2932 patients referred for noninvasive carotid evaluation at the University of Washington during a 5-year period between January 1979 and January 1984, 219 (7.4%) had an ICA occlusion at the time of study. Three patients whose occlusions resulted from neck irradiation in the treatment of malignant disease and four others lost to follow-up were excluded. The remaining 212 patients arc the subject of this article. At the time of initial evaluation, 111 patients had a history of stroke (98 ipsilateral, 9 contralateral, and 4 nonspecific), 42 patients had TIAs, 35 patients were asymptomatic, and 24 patients had nonhemispheric symptoms. Of those patients admitted with stroke, 42 had suffered the event more than 6 months before the occlusion had been documented. Other causes for referral were the presence of a cervical bruit, postendarterectomy follow-up, evaluation of nonhemispheric symptoms, or assessment before cardiac surgery. The asymptomatic patients with bruits and those patients who had had carotid endarterectomy were placed in our ongoing research protocols for these subgroups and were followed up at 3- and 6-month intervals. Those patients who had not been followed up regularly were contacted by telephone to ascertain their current status. When required, further data were obtained from the attending physician. The diagnosis was established by results of ultrasonic duplex scanning with spectral analysis. The accuracy of this method in diagnosing occlusion when compared with arteriography has varied between 93% and 100%. TM The Fisher-Irwin test for statistical analysis of tabular data was used.~7 Life-table analysis was performed for the end points of stroke and death. Comparison of median values was performed with the Mann-Whitney procedure, is Survival analysis and comparison of survival distributions among subgroups were done with the life-table program included in the Statistical Package for the Social Sciences. ~9 Risk ratios were estimated by means of the Cox proportional hazards model. 2 RESULTS The study population consisted of 170 men with a median age of 65 years (range, 42 to 90 years) and 42 women with a median age of 64 years (range, 41 to 88 years). The average follow-up period was 24.9 months. Arteriograms were available for comparison in 89 of 212 patients (40%) and the agreement with duplex scanning data was 94%. The cumulative survival rate at 5 years was 62% (40 deaths) (Fig. 1). Thirty-five deaths occurred in the group of men (seven were stroke related). Five women died but none of these deaths was related to stroke. The cumulative 5-year survival rate for women was 86% and for men, 57% (p = 0.66). Other causes of death

3 Volume 4 Number 5 November 1986 Carotid artery occlusion 481 a Ipsilateral 0 8 All " Ipsilateral.~ > All ~ (/) o_ ~c o 0.4- o_ ~ Stroke o - z~,,,---~ti A a_ '2 2~4 3'6 4'8 6'0 Fig. 2. Life-table analysis: cumulative stroke-free and transient ischemic attack (T/A)-free rates. Five-year cumulative rates were as follows: all strokes, 75%; ipsilateral strokes, 84%; all TIAs, 81%; ipsilaterial TIAs, 91%. Closed circles, stroke; open triangles, TIA. Table I. Outcome according to presenting symptoms Presenting symptoms Stroke during follow-up Death during follow-up Death from stroke Stroke (n = 111) 22 (20%) 21 (19%) 3 (14%) TIA(n = 42) 2 (5%) 4 (9%) 0 Nonhemispheric 4 (17%) 6 (25%) 3 (50%) (n = 24) Silent occlusion (n = 35) 3 (9%) 9 (26%) 1 (11%) TIA = transient ischemic attack. were cardiac disease, 22 of 40 patients (55%); malignancy, 3 of 40 patients (7.5%); infected gangrene of diabetic limbs (two patients); and ruptured aortic aneurysm (one patient). The cumulative stroke-free survival rate at 5 years was 75% (Fig. 2). Of the 31 strokes that occurred, 18 (64%) were ipsilateral, five were contralateral, and five were nonspecific or of the brain stem type. The median age of those patients suffering stroke during follow-up was 66 years (n = 31), not statistically different from the nonstroke group (65 years, n = 191). The cumulative proportion of patients who were free of stroke at 5 years was 77% for women and 72% for men (p = 0.64). The overall cumulative survival rate for those patients who were free of TIAs at 5 years was 82%, for ipsilateral TIAs, the survival rate was 92% (Fig. 2). Eleven of the 23 events were in the territory of the occluded artery. Three TIAs were premonitory for stroke. Eight patients had bilateral occlusion. During follow-up three patients died (all from causes other than stroke) and none suffered stroke. The difference in the mortality and stroke rates was not significantly different in this subgroup compared with those patients who had unilateral occlusion (p = 0.28 and p = 0.19, respectively). Of those patients who had a history of stroke, 22 of 111 (20%) suffered a further stroke and 21 of 111 patients (19%) died. Three of these deaths (14%) were stroke related. Of those patients referred for the treatment of TIA, only 2 of 42 (5%) suffered a stroke and none subsequently died. Stroke and mortality rates for the other referral groups are listed in Table I. The association between the degree of contralateral disease and death from any cause during followup reached statistical significance (p = 0.04, Fig. 3). The degree of contrallateral disease did not show a correlation with the risk for stroke during follow-up (p = 0.83) or even for stroke in the territory of the patent carotid artery (p = 0.73, Fig. 4). Contralateral endarterectomy. Fifty-five patients underwent endarterectomy of the ICA opposite the occlusion. The 5-year survival rate was 82%

4 482 Nicholls et M. Journal of VASCULAR SURGERY 1.0 = ~ e -'A o~ 0.81._~ co 0.6- ".~ S 0.4- o. o o_ e =.L E 12 2'4 3'6 4~8 6~3 Fig. 3. Life-table analysis: cumulative survival by severity of contralateral disease (percentage of vessel diameter reduction). A, normal (n = 2); B, 1% to 15% (n = 50); C, 16% to 49% (n = 53); D, 50% to 99% (n = 98); and E, occluded (n = 8). Statistical significance demonstrated (p = 0.04). 1.0 " - =A,E o, 0 8 t ~ g D "!1 I 0.6 C 0.4 a_ ~2 2 ~, 3; 4'8 6'0 Fig. 4. Life,table analysis: cumulative stroke-free rate for hemisphere ipsilateral to the patent carotid artery grouped by severity of disease in that artery. For explanation of categories, see Fig. 3. No statistical significance demonstrated (p = 0.73). in this group and 54% in those patients who did not have surgery; this difference was not statistically significant (Fig. 5). The proportion of patients who were free of stroke at 5 years in the operated (77%) and nonoperated (75%) groups was similar (p = 0.64). Endarterectomy did not alter the stroke rate for the hemisphere ipsilateral to the occlusion (5-year stroke-free rates were 89% for the oper- ated group and 78% for the nonoperated group, p = 0.83). For the hemisphere ipsilateral to the endarterectomy, the 5-year stroke-free rate was 92% (nonoperated) and 100% (operated) (p = 0.15). The survival rate for patients without stroke at 5 years was 80% for those patients not taking aspirin and 74% for those patients with a history of aspirin ingestion, an insignificant difference (p--0.73).

484 Nicholls et al. Journal of VASCULAR SURGERY 0.8 0.6" TEA + TEA - p=o.15 0? t 1'2 i4 6b Fig. 5. Life-table analysis: cumulative survival for patients with and without endarterectomy (TEA).

5 484 Nicholls et al. Journal of VASCULAR SURGERY " TEA + TEA - p=o.15 0? t 1'2 i4 6b Fig. 5. Life-table analysis: cumulative survival for patients with and without endarterectomy (TEA). Operated group (TEA+): n = 155, n = 108, n = 84, n = 46, and n = 24 at 1, 2, 3, 4, and 5 years, respectively. Nonoperated group (TEA -): n = 55, n = 46, n = 33, n = 20, and n = 15 at 1, 2, 3, 4, and 5 years, respectively. clusion has been examined in several studies and reveals larger variation than the mortality data. In four studies that gave incomplete or no information about the vascular territory relative to the side of the occlusion, the stroke rate ranged from 7% to 25%. ~'2'22'2s Reported rates for ipsilateral stroke also vary widely, ranging from zero to 28% (Table III). 7,ls'2s'26 Only two other studies report life-table data for stroke. A cumulative stroke-free survival rate of 83% at 5 years was reported by Sacquegna et al. ~4 and of 85% by Furlan, Whisnant, and Baker, 6 who also noted that two thirds of the strokes were ipsilateral. These rates are somewhat lower than the present study, in which the 5-year cumulative stroke-free rate was 75%, and the rate for the ipsilateral hemisphere was 84% (an annual adjusted ipsilateral stroke rate of 3%). Therefore, the pooled data suggest that the annual stroke rate ipsilateral to an occluded artery is in the range of 2% to 3%. There is greater variation in the reported rates of TIA. Furlan, Whisnant, and Baker 6 reported 13 of 138 patients (9%) with this symptom, although the side was not specified. Saquegna et al) 4 reported a 4% rate of ipsilateral TIA; Bogousslavsky et al.zs reported 34% and Cote, Barnett, and Taylor, 7 50%. Persson et al.s reported that 15 of 96 patients had TIAs, nine of which were ipsilateral. In this study, the incidence was 11% (23 of 212 patients), with only three TIAs being premonitory for stroke. The incidence of "silent" occlusions in this study was 17% 35 of 212 patients. Other reported rates vary from 7% to 28%. s,6,~3 These data underline the importance of the collateral circulation at the time of ocdusion. Fields and Lemak 2 observed that the associated neurologic deficit is less severe in patients with severe disease in the carotid artery contralateral to the occlusion, presumably as a result of greater collateral development. Some severe associated events may be caused by embolism rather than decreased flow. H,27 The role of endarterectomy for the ICA opposite the occlusion did not affect the natural history significantly, although there was a trend towards improved survival. This may reflect the selection of patients at better risk for surgery. There was an improvement in the stroke rate on the side of endarterectomy, although this did not reach statistical significance; no strokes in the territory of the operated arteries occurred during follow-up. CONCLUSIONS The annual stroke rate in the territory of an occluded ICA is 3% and constitutes two thirds of all strokes in such patients. The 5-year survival rate (62%) is similar to that of other groups with known carotid disease and death caused by cardiac disease exceeds that from all other causes combined. Patients who have a history of stroke when the occlusion is diagnosed have a significantly higher rate of subsequent stroke and death than those who are admitted to the hospital with TIA. Diabetes and hypertension arc associated with increased risk of stroke, whereas gender and aspirin consumption have no significant impact. The occurrence of TIA is not helpful in the prediction of patients at increased risk for stroke occurring after carotid occlusion. Endarterectomy of

6 Volumc 4 Number 5 November 1986 Carotid artery occlusion 483 Table II. Estimated risk ratios from Cox regression models fit to occlusion data Coronary p Value for Model Diabetes mellitus Hypertension artery disease Aspirin Gender addition to model ~ (0.06) (0.08) (0.07) (0.05) 3.00 (0.08) (0.08) 0.44 (0.07) (0.04) 0.42 (0.06) (0.04) 3.00 (0.08) 0.41 (0.06) (0.04) 2.94 (0.08) 0.41 (0.06) 1.35 (0.53) (0.04) 3.16 (0.07) 0.44 (0.08) 1.34 (0.55) 1.49 (0.40) 0.40 NOTE: A smoking covariate is included in each model but not shown. Values of p are in parentheses. *Values of p are for addition of a single factor to a model listed previously in the table. Table III. Outcome of carotid occlusion Sample size Mean Deaths Stroke TIA (No. of follow-up patients) (too) No. Cardiac Stroke All Ipsilateral All Ipsilateral Fields, Lemak (34%) 80 (40%) 89 Current study (55%) 7 (18%) (11%) 11 Bums et al.2~ (20%) 10 (66%) 5 Furlan et al (57%) 2 (7%) (9%) Saquegna et al.~ (53%) 4 (27%) (4%) 4 Persson et al. s (33%) 1 (8%) (6%) 9 Norrving, Nilsson (33%) (10%) 2 Cote et al (50%) 1 (25%) Dyken et al? Grillo, Patterson (70%) 1 (10%) Barnett et al~ Bogousslavsky et al. 2s (34%) 8 When analyzed further by sex, these differences remained insignificant (men, p = 0.37; women, p = 0.42). Risk factor analysis with a regression model. Nine separate regression models to estimate stroke risk were fitted to the study data (Table II). Values are expressed as risk ratios with p values shown in parentheses. Models 1 through 7 consider diabetes mellitus, hypertension, and coronary artery disease as risk factors singly and in combination. Smoking is included as a covariate in each model. Risk ratios for each risk factor are little changed by addition or deletion of the other factors and all are at or near the 5 % level of statistical significance. Model 7 best summarizes the effect of each factor on stroke risk while controlling for the other factors. Increased risk is noted for diabetes mellitus (2.46 ) and hypertension (3.16 ). Cardiac disease was associated with a reduced risk.of stroke (0.44 ). The reason for this is unclear, although the increased risk of death may have altered the exposure for stroke in this group. Coronary artery disease was associated with an in- creased incidence of non-stroke-related death. Aspirin ingestion (1.34 ) and gender (1.49 x ) were not significant risk factors for subsequent stroke. DISCUSSION The prognosis after carotid occlusion depends on presenting symptoms. A history of stroke at the time of diagnosis is associated with a high incidence of further stroke or death, whereas presentation with TIA has a more benign outcome. The cumulative 5- year survival rate of 62% in our study group is similar to other studies (62.3%, 14 77%, 6 and 74%11). The major cause of late death in patients with carotid artery occlusion is cardiac disease, not stroke. Results from our study, showing 22 of 40'(55%) cardiac deaths and 7 of 40 (17.5%) stroke-related deaths, are similar to those reported in other series with longterm follow-up of carotid occlusion (Table III). These rates are comparable to findings reported in patient groups with TIA or after carotid endarterectomy) 2 The prognosis for stroke occurring after ICA oc-

7 Volume 4 Number 5 November 1986 Carotid artery occlusion 485 the carotid artery opposite the occlusion did not affect the overall survival or stroke rate but appeared to lower the incidence of stroke in the hemisphere ipsilateral to the endarterectomy. REFERENCES 1. Dyken ML, Klatte E, Oldrich JK, Spurgeon C. Complete occlusion of common or internal carotid arteries. Arch Neurol 1974;30: Fields WS, Lemak NA. Joint study of extracranial arterial occlusion. X. Internal carotid artery occlusion. JAMA 1976;235: Castaigne P, Lhermitte F, Gautier JC, Escovrolle R, Derousne C. Internal carotid artery occlusion. A study of 61 instances in 50 patients with post mortem data. Brain 1970;93: Handott AM, Millikan CH. Pathogenesis of cerebral interaction secondary to mechanical carotid artery occlusion. Stroke 1970;1: Persson AV, Griffey EE. Jaxheimer EC, Jewell ER. The natural history of total occlusion of the internal carotid artery. Vasc Diagn 1984;5(3): Furlan AJ, Whisnant JP, Baker HL. Long-term prognosis after carotid artery occlusion. Neurology 1980;30: Cote R, Barnett HJM, Taylor DW. Intemal carotid occlusion: A prospective study. Stroke 1983;14: Genton E, Barnett JHM, Fields WS, Gent M, Hoak JC. Cerebral ischemia: The role of thrombosis and antithrombotic therapy. Stroke 1977;8:150-75, 9. Murphey F. Treatment of the completed stroke with longterm anticoagulants: Six and one-half years experience. In: McDowell FH, Whisnant JP, eds. Cerebral vascular diseases. Fourth Princeton Conference, Barnett HJM, Peerless SJ, Kaufmann JCE. "Stump" of internal carotid artery--a source for further cerebral embolic ischemia. Stroke 1978;9: Einsiedel-Lechtape H. Neuroradiology of vascular and atrophic lesions in the geriatric age. Neuroradiology 1978; 16: Adams HP, Kassell NF, Mazuz H. The patient with transient ischemic attacks--is this the time for a new therapeutic approach? Stroke 1984;15: Grillo P, Patterson RH. Occlusion of the carotid artery: Prognosis (natural history) and the possibilities of surgical revascularization. Stroke 1975;6: Sacquegna T, DeCarolis P, Pazzaglia P, et al. The clinical course and prognosis of carotid artery occlusion. J Neurol Neurosurg Psychiatry 1982;45: Langlois YE, Rocderer GO, Chan A, et al. Evaluating carotid artery disease. Ultrasound Med Biol 1983;9: Langlois YE, Roederer GO, Chan A, et al. Postendartcrectomy cartoid ultrasonic duplex scanning. Ultrasound Med Biol 1983;9: Fleiss JL. Statistical methods for rates and proportions, ed 2. New York: John Wiley & Sons, Inc, Hollander M, Wolfe DA. Nonparametric statistical methods. New York: John Wiley & Sons, Inc, Hull CH, Nie NH. Statistical package for the social sciences. Release 9. Chicago: SPSS Inc, Breslow NE, Day NE. Statistical methods in cancer research. Vol 1. The analysis of case-control studies. IARC scientific publication no. 32. Lyon: International Aency for Research on Cancer, Burns J, Satiani B, Vasko JS. Long-term survival following total carotid artery occlusion. Cardiovasc Rev Rep 1984; 5: McDowell FH, Potes J, Groch S. The natural history of internal carotid and vertebral-basilar artery occlusion. Neurology 1961;1 l(suppl): Hardy WG, Lindncr DW, Thomas LM, Gurdjian ES. Aaaticipatcd clinical course in carotid artery occlusion, Arch Ncurol 1962;6: Norrving B, Nilsson B. Carotid artcrv occlusion: Acutc symptoms and long-term prognosis. Ncurol Rcs 1981;3: Bogousslavsky J, Rcgli F, Hungcrbuhler JP, Chrzanowski R. Transicnt ischcmic attacks and cxtcrnal carotid artery. A rctrospectivc studv of 23 paticnts with an occlusion of thc intcmal carotid artery. Strokc 1981; 12: Barnett IHM. Delayed ccrcbral ischcmic cpisodcs distal to occlusion of major ccrcbral artcries. Neurology 1978;28: Lhcrmitte F, Gautier JC, Dcroucsne C. Nature of occlusions of the middlc cercbral artcry. Ncurolog T 1970;20:82-5.

Internal Carotid Artery Occlusion: Clinical and Therapeutic Implications

94 Internal Carotid Artery Occlusion: Clinical and Therapeutic Implications VIVIAN U. FRITZ, M.D., CHRIS L. VOLL, M.D., AND LEWIS J. LEVIEN, M.D., PH.D. Downloaded from http://ahajournals.org by on November