Ten year cerebrovascular morbidity and mortality in 68 year old men with asymptomatic carotid stenosisBMJ 1995; 310 doi: https://doi.org/10.1136/bmj.310.6990.1294 (Published 20 May 1995) Cite this as: BMJ 1995;310:1294
- Mats Ogren, senior lecturer in epidemiologya,
- Bo Hedblad, senior lecturer in epidemiologya,
- Sven-Olof Isacsson, professor of social and preventive medicinea,
- Lars Janzon, professor of epidemiologya,
- Gudrun Jungquist, senior lecturerb,
- Sven-Eric Lindell, professorb
- a Department of Community Health Sciences, Lund University, Malmo General Hospital, S-21401 Malmo, Sweden
- b Department of Clinical Physiology, Lund University, Malmo General Hospital, Malmo, Sweden
- Correspondence to: Dr Ogren.
- Accepted 7 March 1995
Objective: To study the natural course of carotid artery stenosis detected by ultrasonography.
Design: Prospective cohort study. Baseline examination in 1982-3 included ultrasound examination of carotid arteries, measurement of anklebrachial blood pressure index, and detection of atrial fibrillation by 24 hour ambulatory electrocardiography.
Setting: Malmo, a city in southern Sweden with 230000 inhabitants.
Subjects: 470 men aged 68 years randomly selected from the population.
Main outcome measures: Incidence of stroke and transient ischaemic attack and all cause mortality during 10 years of follow up in relation to carotid stenosis, leg artery disease (ankle-brachial blood pressure index below 0.9), and atrial fibrillation.
Results: Fifty men had a stroke; six of these were haemorrhagic. Another 11 had a transient ischaemic attack. Eighteen of the men with carotid stenosis (21.6 events/1000 person years) and 43 of the men with normal carotid arteries (14.8 events/1000 person years) had a stroke or transient ischaemic attack (P=0.188). Men with atrial fibrillation had an increased rate of cerebrovascular events (36.7/1000 person years (P=0.048). The highest rate was found in men with asymptomatic disease of the leg arteries (38.6/1000 person years) (P<0.001). The increased risk of stroke or transient ischaemic attack in this group remained after multivariate analysis (relative risk 2.0; 95% confidence interval 1.1 to 3.7).
Conclusions: In this cohort carotid stenosis was not associated with an increased risk of stroke. Part of this lack of association was explained by the high mortality from ischaemic heart disease in men with severe stenosis. Twenty seven of the 61 cerebrovascular events, however, occurred in men who had normal carotid arteries, normal ankle pressure, and no atrial fibrillation.
Asymptomatic atherosclerosis in carotid and leg arteries is a common occurrence among elderly people
Both conditions were found to be associated with an excess cardiovascular mortality
Compared with leg artery disease, asymptomatic carotid stenosis was not found to be associated with an increased risk of stroke
Assessment of peripheral atherosclerotic disease by measurement of arm and ankle blood pressure is a useful method for identifying patients with an increased risk of stroke
The clinical spectrum of carotid artery disease ranges from an incidentally discovered bruit to a stage with advanced arteriosclerotic lesions and subsequent stroke. Modern ultrasound techniques have extended the range of disease even further, and several studies have published results on the occurrence of asymptomatic carotid stenosis in patients with different arteriosclerotic diseases.1 2 3 4
We have previously reported on a 25% prevalence of carotid stenosis including a reduction by 30% or more in internal diameter in 68 year old men selected at random from the general population.5 This study, which was carried out in 1982-3, showed a relation between history of stroke or transient ischaemic attack and carotid stenosis. The 470 men, who belong to the prospective cohort study of men born in 1914 from Malmo in Sweden, have been followed up since this initial health examination, and it is now possible to describe the 10 year incidence of cerebrovascular disease in relation to prevalence of carotid stenosis at 68 years of age.
Occlusive leg artery disease has been found to be associated with an increased risk of cerebrovascular disease.6 Others have shown a relation between carotid stenosis and non-invasively detected leg artery disease and also that asymptomatic leg artery disease in men with carotid stenosis is associated with an increased risk of future myocardial morbidity.7 We compared cerebrovascular morbidity in men with and without an ankle-brachial pressure index <0.9, which can be regarded as a sign indicating the presence of substantial stenosis of the arteries.8 9
Nineteen (4%) of the 456 men in the original cohort study had atrial fibrillation detected during 24 hour ambulatory electrocardiography in 1982-3.10 We assessed atrial fibrillation as a risk factor for stroke or transient ischaemic attack11 in the presence of carotid stenosis and leg artery disease.
Subjects and methods
The cohort study of men born in 1914, which focused on the epidemiology of cardiovascular and pulmonary disease, has been running since 1969.12 In 1982-3, all men born in even months in 1914 and residing in Malmo were invited to a health examination close to their 68th birthday. Of 621 invited, 500 took part in this examination. Four hundred and seventy (75.7%) of the 621 underwent carotid ultrasound examination and the measurement of arm and ankle blood pressure, and 456 (73.4%) underwent 24 hour ambulatory electrocardiography.
CAROTID ARTERY EXAMINATION
A reduction of the internal diameter of the carotid artery of 30% of more in one or both arteries was used for the definition of carotid stenosis.5 A quarter of the men fulfilled this criterion. Forty seven (10%) had a reduction in diameter exceeding 45%. In men with bilateral stenosis the most severely affected side was used for classification of the degree of lumen reduction.
The ultrasound examination of the carotid arteries was made with a continuous wave Doppler with a 5 MHz transducer (Dopscan 1050, Carolina Medical Electronics, King, North Carolina). With Dopscan 1050 it is possible to detect stenoses which reduce the internal carotid artery diameter by 30% or more with a sensitivity of 97% and a specificity of almost 100%. A reduction in diameter of 30% corresponds to a maximum frequency shift of 3.1 kHz, and a stenosis of >45% corresponds to a cut off point of 4.3 kHz.13 The examination was made with the subject in the supine position after 10 minutes of rest. Measurements were made in the common carotid artery 2 cm beneath the bifurcation; in the proximal internal carotid artery from the point of the vessel where the most abnormal Doppler signals were found; and 2 cm distally of the obstruction. If no Doppler signals could be obtained in the proximal and distal part of the internal carotid artery the vessel was diagnosed as totally occluded.
History of stroke at the baseline examination was based on the results of a questionnaire, clinical examination, and hospital records. The final judgment was made by a neurologist who was unaware of the result from the ultrasound examination.5
RECORDING OF ANKLE-BRACHIAL BLOOD PRESSURE INDEX
The recording system consisted of pulse sensors (mercury in Silastic strain gauges) placed on the big toes and thumbs; two Wheatstone bridges with amplifiers to record changes in the resistance of the strain gauges; blood pressure cuffs (18x60 cm to measure systolic ankle pressure and 12x35 cm to measure the systolic upper arm pressure); a pressure transducer (Siemens-Elema EMT 746 with amplifier EMT 311) to record cuff pressures; and a six channel ink jet recorder (Siemens-Elema; Mingograph).14 15 Duplicate recordings were made with the subject in the supine position and the arithmetic average used. For each leg, an ankle-brachial blood pressure index was calculated by dividing the systolic pressure in the ankle by the highest systolic pressure in the upper arm.
Leg artery disease was defined as an ankle-brachial pressure index <0.90 in one or both legs.*RF 8, 9 Subjects with a pressure index >/=0.90 in both legs were considered free from leg artery disease.
RISK FACTORS FOR ARTERIOSCLEROTIC DISEASE
Systolic and diastolic (phase V) blood pressures were measured in the right arm with the subject in the sitting position after 15 minutes of rest. Blood pressure was recorded to the nearest 5 mm Hg. Smoking habits were assessed by means of a structured questionnaire. The men were divided into never smokers (less than 1 g of tobacco a day for less than one year); former smokers (those who had stopped smoking at least a month before investigation); and current smokers. Plasma cholesterol concentrations were analysed by standard methods and expressed as mmol/l. Diagnosis of atrial fibrillation was based on the findings of 24 hour ambulatory electrocardiography.10
FOLLOW UP DATA
Mortality—The men were followed up from the baseline study in 1982-3 until their death or 31 March 1993. Mean (range) follow up time was 102.9 (3.2-128.4) months. Data on mortality were obtained from the mortality register of the Swedish National Bureau of Statistics. In 65% of the deaths necropsy was performed. Mortality was expressed as deaths per 1000 person years of observation. Cases coded 430.0-438.9 according to the International Classification of Diseases, eighth revision (ICD-8) were counted as deaths from cerebrovascular disease and cases coded 410.0-412.9 as deaths from ischaemic heart disease.
Rate of cerebrovascular events—The follow up period was from the cross sectional study in 1982-3 until the first cerebrovascular event, death, or 31 March 1993. Status was updated on each subject until this date. Mean (range) follow up was 97.0 (2.8-128.4) months. Cerebrovascular events were traced through the local patient administrative system, to which both of the two hospitals in the city are linked. End points were validated by study of the hospital records, results from computed tomography, and necropsy reports. A cerebrovascular event was defined as stroke by using the WHO definition16 (ICD-8 codes 430.0-434.9 and 436.0-438.9) or as transitory ischaemic attack17 (codes 435.0-435.9). Rates of cerebrovascular events were expressed as events per 1000 person years of observation.
Mortality and rate of cerebrovascular events in men with signs of carotid stenosis and in men with signs of leg artery disease were compared with rates in men without corresponding signs of disease. Computations were made by using the Kaplan-Meier method with the generalised Wilcoxon test.18 Risk factors at baseline were compared in men with and without carotid stenosis who did and who did not suffer a cerebrovascular event during follow up by using analysis of variance comparison of means and X2 tests for comparison of proportions. Cox's proportional hazards model was used to study mortality and rate of cerebrovascular events in relation to carotid stenosis and arm-ankle pressure index, systolic blood pressure, smoking, and presence of atrial fibrillation, when previous stroke was controlled for.19
ALL CAUSE MORTALITY AND RATE OF CEREBROVASCULAR EVENTS
Of the 470 men who were examined at 68 years of age, 26 (5.5%) had a history of stroke. Sixty one (13.0%) had at least one cerebrovascular event during follow up (table I). Of these, 54 (88.5%) had no previous episode of cerebrovascular disease. The cerebrovascular event was classified as stroke in 50 (82.0%) of the 61 cases. Ten of these had a transient ischaemic attack before their stroke. Eleven (18%) of the 61 men had a transient ischaemic attack only. Six men had haemorrhagic stroke. Seven men died within two weeks after the onset of stroke. The diagnosis of stroke was confirmed by computed tomography or by findings at necropsy in 38 (76.0%) of the 50 cases.
During follow up 158 men (33.6%) died. Of these, 16 died from focal cerebrovascular disease and 51 from ischaemic heart disease. Of the 121 men who were invited to participate in the health examination at 68 years of age but who declined to participate, 73 (60.3%) died during the follow up period. Three of these died from stroke and 17 from ischaemic heart disease.
INCIDENCE OF CEREBROVASCULAR DISEASE AND ALL CAUSE MORTALITY IN RELATION TO CAROTID STENOSIS
Fifteen of the 113 men with carotid stenosis involving a reduction of >30% in internal diameter had a stroke and three had a transient ischaemic attack, corresponding to a cerebrovascular event rate of 21.6/1000 person years. This rate was not significantly different from the rate in men with normal carotid arteries (14.8/1000 person years) (P=0.188) (table II; fig 1). Six out of 43 events (14%) were of the vertebrobasilar type in men who had normal carotid arteries. In men with carotid stenosis no such events occurred. The rate of cerebrovascular events in men with a stenosis >45%—that is, the top 10% of the distribution—was 20.7/1000 person years (7/48).
The mortality in men with carotid stenosis >30% was 52.2/1000 person years. Men with normal carotid arteries had a mortality of 36.1/1000 person years (P=0.030) (table II; fig 2). Of the 47 men with carotid stenosis who died during follow up, five died from cerebrovascular disease and 22 from ischaemic heart disease. Of the 48 men who had carotid stenosis exceeding 45%, 21 died during follow up, 14 of them from ischaemic heart disease.
INCIDENCE OF IPSILATERAL STROKE AND TRANSIENT ISCHAEMIC ATTACK
In the 65 men who had right carotid artery stenosis, 11 cerebrovascular events occurred. Of these, four were localised to the ipsilateral cerebral hemisphere, five affected the contralateral hemisphere, and two affected both hemispheres (table I). Among the 65 men with left sided carotid stenosis eight events occurred: five ipsilateral, two contralateral, and one bilateral (table I).
INCIDENCE OF CEREBROVASCULAR DISEASE AND ALL CAUSE MORTALITY IN MEN WITH LOW ANKLE-BRACHIAL PRESSURE INDEX
Thirty six per cent (24/66) of the men with leg artery disease had carotid stenosis >30%. The rate of cerebrovascular events in men who had an ankle-brachial pressure index below 0.9 in either leg (38.6/1000 person years) was almost three times as high as it was in men with normal arm-ankle indices (13.5/1000 person years) (P<0.001) (table II). Of the men with leg artery disease, the rate of cerebrovascular events in men without carotid stenosis was 40.7/1000 person years, and in men with carotid stenosis it was 34.8/1000 person years (fig 1).
Men with asymptomatic leg artery disease had an all cause mortality of 82.2/1000 person years, which was 2.5 times higher than that in men with normal pressure indices (34.2/1000 person years) (P<0.001) (fig 2).
ATRIAL FIBRILLATION IN RELATION TO INCIDENCE OF CEREBROVASCULAR DISEASE AND ALL CAUSE MORTALITY
Of the 456 men participating in ambulatory electrocardiography, 19 (4.2%) had atrial fibrillation. Of these, 17 had chronic and two had intermittent atrial fibrillation. Men with atrial fibrillation had a higher rate of cerebrovascular events (36.7/1000 person years) (P=0.048) and higher total mortality (79.4/1000 person years) (P=0.005) than men without fibrillation (corresponding rates 14.9/1000 and 35.3/1000, respectively; table II).
RATE OF CEREBROVASCULAR EVENTS IN RELATION TO ARTERIOSCLEROTIC RISK FACTORS
In the risk factor analysis men who did and who did not develop cerebrovascular disease during follow up were compared after stratification according to presence and absence of carotid stenosis at 68 years of age (table III). In men who had no signs of carotid stenosis at 68, mean systolic blood pressure at that age was significantly higher in men who during follow up suffered a stroke or a transient ischaemic attack (161 mm Hg; 95% confidence interval 153 to 169 mm Hg) than it was in the men who remained free from cerebrovascular disease (152 mm Hg; 150 to 154 mm Hg).
Multivariate analysis of mortality and cerebrovascular morbidity in relation to carotid stenosis, armankle pressure index, previous stroke, and risk factors showed that the presence of a pressure index <0.9 in one or both legs at 68 years of age was independently associated both with an increased mortality (relative risk 2.0; 95% confidence interval 1.3 to 3.0) and with an increased rate of cerebrovascular events 2.0; 1.1 to 3.7) during 10 years of follow up (table IV). The association between carotid stenosis and mortality that was found in the univariate analysis did not remain in the Cox's proportional hazards model (1.2; 0.8 to 1.8). No significant association was found between carotid stenosis and rate of cerebrovascular events (1.2; 0.7 to 2.2). Systolic blood pressure was found to be independently associated with an increased rate of cerebrovascular events, with an 8.2% rise of relative risk per 5 mm Hg increment of systolic blood pressure.
This chort study was designed to describe the natural course of carotid stenosis. In agreement with the ethics committee at Lund University we decided that the probands and their doctors should not be informed about the results from the carotid artery examination. The lack of an association between carotid stenosis and cerebrovascular disease should therefore not have been confounded by selective treatment of men with carotid stenosis to prevent stroke. Only four of the men with carotid stenosis were subject to carotid endarterectomy during follow up. Three of these had their cerebrovascular end point before surgery.
The relatively narrow 95% confidence interval around the point estimate of the relative risk does not indicate that the lack of association is explained by low power. A more plausible explanation is a concomitant higher mortality from other arteriosclerotic diseases in men with carotid stenosis, which agrees with conclusions of other studies.20 21 22 The mortality from ischaemic heart disease increased with the severity of carotid stenosis. Of the 34 men who had a carotid stenosis exceeding 50%, 17 died during follow up, 13 of them from ischaemic heart disease. Three of these had had a stroke. None of the 17 men with severe stenosis who survived during follow up, however, had a stroke.
The incidence of cerebrovascular events in relation to carotid artery disease must take into account the validity of the method used for detection of carotid stenosis. Dopscan 1050 has in comparison with angiograms been shown to have a sensitivity of 97% and a specificity of almost 100% for the detection of a reduction in the diameter of the carotid artery lumen 30% or more.13 Hence we could not explain the high proportion of cerebrovascular events in men with normal results of carotid ultrasound examination on the basis of an inadequate sensitivity.
Carotid stenosis is associated with an increased prevalence of leg arteriosclerosis.7 Non-invasively detected leg artery disease has previously been reported to be associated with stroke.6 In this study, a low ankle-brachial pressure index was found to be the strongest marker for cerebrovascular disease. These men had no clinical signs or symptoms of leg artery disease. The high incidence of myocardial infarction7 and stroke associated with a low arm-ankle index illustrates that leg artery disease is a sign indicating generalised arteriosclerosis. As the increased mortality associated with carotid stenosis disappeared in the multivariate analysis we cannot consider carotid stenosis to be a similar marker. The risk of stroke or transient ischaemic attack associated with leg artery disease was independent of carotid stenosis. Prediction of risk of future stroke or transient ischaemic attack can therefore be improved by measurement of the arm-ankle pressure gradient.
We have confirmed the well known risk of stroke associated with high systolic blood pressure and presence of atrial fibrillation.11 23 The association between stroke and high blood pressure, however, was not observed in men with carotid stenosis. This is probably because of selection bias—that is, hypertensive men with carotid stenosis had died from ischaemic heart disease.
An important finding in our study is the much increased risk of cerebrovascular disease in subjects with a reduced arm-ankle pressure index. The risk associated with a certain marker should not be assessed in relative numbers only. Seventy four per cent (45/61) of the cerebrovascular events occurred in men with normal ankle-brachial index, 70% (43/61) in men without carotid artery stenosis, and 44% (27/61) in men who had normal carotid arteries, normal arm-ankle index, and no atrial fibrillation. This suggests that in a non-selected group of men the event leading to stroke in several cases is triggered within the intracerebral vascular area. This observation should be considered in the clinical evaluation of patients with asymptomatic carotid artery stenosis.
This study was supported by grants from the Bank of Sweden Tercentenary Foundation, the Wallenberg Foundation, the Swedish Heart Lung Foundation, Emhold Lundstrom's Foundation, the Medical Faculty at the Lund University, and from research foundations administered by Malmo General Hospital.