Article Text

Prevention of neurological disease in later life
  1. Anthony G Rudda,
  2. Charles D A Wolfeb,
  3. Robin S Howarda
  1. aGuy’s and St Thomas’ Trust, London, UK, bUMDS, St Thomas’ Hospital, London, UK
  1. Dr Anthony G Rudd, Elderly Care Department, St Thomas’ Hospital, Lambeth Palace Road, London SE1 7EH, UK.

Statistics from

This article focuses on the prevention of neurological disease in the post-infant phase of life and considers both the public health and individual strategies that have been employed for primary prevention. Secondary prevention is considered for certain conditions—for example, stroke.

To be able to devise primary preventive strategies, sound epidemiological evidence is required. Factors to consider are case definition, study design, and the resultant confidence with which interpretation of the results can be made. For many of the neurological conditions discussed case definition is problematic. In stroke, for example, a clinical definition is used in most research studies. The problem of silent cerebral infarction is consequently omitted. Technological advances, particularly in imaging, also potentially lead to changes in diagnostic acumen. For many neurological conditions there are detailed clinical case series but when assessing risks, most evidence has been based on case-control studies with the resultant biases.

This review primarily considers stroke, as it is the major cause of adult neurological death and disability. With current knowledge of risk factors it is the condition in which there is most scope for primary prevention. A brief review of some of the other major neurological diseases is included to illustrate areas where work has been reported or is required. The perspective in this review is that of the developed world.

Burden of neurological disease

The number of deaths from neurological diseases in the United Kingdom declines from childhood to middle age from around 10% in 1-4 year olds to 5% in 25-34 year olds, with a small proportion thereafter. However, deaths from cerebrovascular disease increase from 35 years and represent the third most common cause of death in the total population.

The Office of Population Censuses and Surveys (OPCS) death data for 19931 detail deaths associated with neurological disorders (table 1):

Table 1

Deaths in England and Wales 1993

There is little data on morbidity from routine statistics but morbidity returns from general practice give a clue to the burden of symptoms (table 2).2 Overall, neurological and sense organ complaints represent the fourth most common cause for presentation, constituting 17% of consultations. This is highest in the under 5 age group with a steady increase with advancing age. Neurology symptoms alone constituted a consultation rate of 1732/10 000 patients in 1991-2. One per cent of the general practice consultations are for migraine.

Table 2

Consultation rates in general practice 1995

Table 3 3 summarises infectious diseases with neurological sequelae reported to OPCS communicable disease statistics (CDC) in 1993 in England and Wales.

Table 3

Infectious disease notifications

Vascular disease: stroke

Many of the risk factors identified for stroke have been those also associated with coronary heart disease, having been identified in observational studies mainly considering the outcome of coronary heart disease, such as the Multiple Risk Factor and Intervention Trial4 and Framingham studies.5 There are risk factors which are specific to stroke but generally the attributable risk fractions have not been accurately estimated. Most studies have only considered the risk factors for total stroke or for ischaemic strokes and relative risks are weighted towards describing the risk factors for thrombotic events, which do, however, constitute 76% of strokes. The fall in mortality from stroke, although encouraging, remains largely unexplained and recent incidence data do not suggest that the decline is continuing.


When considering risk factors for stroke it is important to distinguish between factors which are amenable to reduction and those which are not. In the first group the evidence on how to effectively reduce their prevalence will be presented. The broad categories under which risk can be considered are outlined:

Table 4

Risk factors for stroke

Inherited biological traits

These include age, sex, and genetic predisposition.

Social characteristics

These include ethnic background with cultural effects and socioeconomic status. These traits overlap with behavioural and biological factors.

Environmental factors

These are factors such as temperature and other climatic variables which may affect seasonal and geographical variations in stroke incidence and mortality.

Physiological characteristics

These include blood pressure, serum factors such as cholesterol and glucose, and obesity.

Behavioural characteristics (table 5)

These include smoking, alcohol intake, dietary factors, and the use of oral contraceptives and hormone replacement therapy.

Table 5

Behavioural traits

Inherited biological traits

Age—Both in the United States and the United Kingdom the risk of stroke doubles with each successive decade over the age of 55.6 7 In the United Kingdom population projections predict an increasingly aged population, at increased risk of stroke. Malmgren et al 8 estimate, however, that this will result in a net increase of only 4% of moderately or severely handicapped people. The incidence rates increase from 0.2/1000 population per year aged 45-54 to 10/1000 in people aged over 85.7

Sex—The Framingham study showed an overall 30% increased incidence in men, which was slightly higher in those aged under 65.5 In the Oxfordshire Community Stroke Projects in the United Kingdom, the odds of a male sustaining a first stroke were 26% greater than that of a female.9

Fetal and infant influences—Some groups have attributed the inequalities in health seen in Great Britain to fetal and infant influences.10 There is controversy over the early origins of cardiovascular disease.11 Stroke has been linked to earlier maternal and neonatal mortality.12 If this shift in paradigm for the origins of cardiovascular and cerebrovascular disease is correct, research and treatment of intrauterine diseases may be an alternative strategy to the activities of risk factor reduction in middle age. However, more detailed studies of these influences on cardiovascular and stroke risk factors in childhood and early adulthood are required.

Family history—A family history of stroke in any first degree relative was an independent predictor of stroke in women only of a middle class white cohort, with a relative risk of 2.3 after controlling for age, cholesterol, blood pressure, cigarette smoking, and diabetes.13 Welin et al 14 in a Swedish cohort study of middle aged men, showed that men with mothers who had died of a stroke had a threefold increased risk of stroke and the population attributable risk was estimated to be 0.19. Studies of the occurrence of stroke among twins indicate that deaths from stroke are up to 20 times more common than expected in both monozygotic and dizygotic twins, which implies that a shared family environment is an extremely strong risk factor for stroke.15

Bromberg et al 16 estimated that subarachnoid haemorrhage occurs almost seven times more often in first degree relatives than in second degree relatives and this is the first report to distinguish between the type of relative and risk. This means that the lifetime risk of subarachnoid haemorrhage is between 2% and 5% in first degree relatives. Therefore, screening for unruptured aneurysms should at least be considered in first degree relatives of patients with subarachnoid haemorrhage.

Social characteristics

Socioeconomic status (SEC)—in the United Kingdom Acheson and Sanderson17 estimated a relative risk of 1.6 in the lowest (SEC V) compared with the highest (SEC 1) group. There is also a tendency for the SMR (standardised mortality ratio) for stroke to be higher in northern regions of the United Kingdom and inner city districts where the proportion of SEC IV and V reside.18This may reflect numerous factors such as social deprivation, environmental and geographical differences, and dietary differences. The British Regional Heart Study found no significant difference in risk of stroke between socioeconomic groups after adjustment for other risk factors.19

Ethnic group —In the United States mortality data consistently show higher death rates from stroke among black communities, probably partly because of a higher prevalence of ischaemic heart disease and hypertension in these communities.20-24 In the United States Kittner et al 25 documented a relative risk of stroke incidence of 1.4 for black females and 1.1 for black males after adjustment for age, hypertension, and diabetes mellitus. The possibility that a specific type of stroke accounts for racial differences has been suggested.26 Even among black communities there is an association between darker skin colour and hypertension in lower socioeconomic groups.27

Balarajan28 showed significantly differing mortality rates between ethnic groups in England and Wales. In general, ethnic groups that experienced lower mortality rates from ischaemic heart disease in the 1970s showed the greatest improvement over the next decade. For ischaemic heart disease the highest rates were found in those of Indian origin. Those born in the Caribbean, United States, old Commonwealth, and western Europe had the lowest rates. For stroke the highest rates were in AfroCaribbeans followed by Africans, Indians, and Irish and showed a decline of 28% from 1970-2 to 1979-83. There were, however, classification problems in this analysis and no adjustment was made for social class or other factors. These data only relate to country of origin and not ethnic group and therefore do not take into account the second generation ethnic groups. The highest mortality rates for stroke are in Indian men, which are 53% above the national average, and 25% in Indian women. These data reinforce the need to establish the prevalence of the more conventional risk factors for stroke in large representative samples of these communities. A 76% excess of deaths in Caribbean men has been reported by Cruickshank et al.29 These differences in mortality have never been satisfactorily explained.29 30

A study by Chaturvedi et al 31 found median systolic blood pressure in AfroCaribbean men to be 6 mm Hg higher than in European men and 17 mm Hg higher in AfroCaribbean women than in European women. Haines et al 32 however, did not find a significant difference between blood pressure levels in white and black subjects attending a London inner city, general practice, but found that more black than white people were being treated for hypertension.

Hypertension alone cannot account for the high stroke rates in black populations. A further important consideration is the difference in the prevalence of diabetes mellitus between ethnic groups. Higher mortality and hospital admission rates have been found in AfroCaribbeans than for other ethnic groups.29 33 Miller et al 34 surveyed 209 men (81% of all registered in a general practice) and found that hyperglycaemia was present in 29% of AfroCaribbeans compared with 9% of Europeans. Gaines et al 24 reviewing the literature on black and white differences in risk factors for stroke conclude that apart from hypertension, smoking, diabetes, and obesity are more prevalent in black subjects.

In the United States, Howard et al 35estimated that socioeconomic status accounted for less than 25% of the excess male deaths from stroke among black people aged 45-64 and in black women it did not significantly reduce the estimated excess mortality from stroke.

Environmental factors

Seasonal variation in stroke mortality has been reported by many groups.36-39

The possible mechanisms may relate to the effects of temperature on blood pressure and clotting mechanisms and further research is required on the environment in which the stroke occurred in relation to blood pressure and haemostatic factors.

Physiological characteristics

Hypertension: observational and ecological data—Observational studies have all shown hypertension to be the single most important risk factor for stroke and one which has been shown to be eminently reducible by antihypertensive treatment in the randomised controlled trial setting. The attributable risk factor has, however, varied from 0.6 in the Framingham study40 to 0.75 in the Whitehall study.41-43

Secular trends in hypertension and its cardiovascular sequelae are of great public health importance. The decline in mortality from both stroke and coronary heart disease predate effective antihypertensive therapy and the reasons remain unclear as to the contribution of antihypertensive therapy to this decline.44

Bonita and Beaglehole45 suggested that at least three quarters of the decline in stroke mortality in the United States in the period 1970-80 was due to factors other than antihypertensive treatment.

Hypertension intervention data: meta-analysis—MacMahon and colleagues46 undertook a meta-analysis of nine major prospective observational studies in subjects aged 25-84, 98% male. After correction for regression dilution bias, prolonged differences in usual diastolic blood pressure of 5, 7.5, and 10 mm Hg were respectively associated with at least 34%, 46%, and 56% less stroke events and for the large majority of subjects, whether conventionally hypertensive or normotensive, a lower blood pressure should eventually confer a lower risk of vascular disease. This effect was seen throughout the blood pressure range and there was no evidence of any threshold below which levels of diastolic blood pressure were not associated with lower risk of stroke. Therefore for populations in which strokes are common a substantially lower diastolic blood pressure should eventually confer a substantially lower relative risk of vascular disease. The absolute benefits of a lower blood pressure would be greatest for large countries with high stroke rates, such as China. They concluded that there is still a need for better understanding of the dietary and other determinants of population blood pressure levels.

Collins et al 47 analysed 14 randomised trials of antihypertensive drugs. Long term differences seen in observational studies of 5-6 mm Hg in usual diastolic blood pressure were associated with about 35%-40% less stroke events. Stroke was reduced in trials by 42 (SD 6)% suggesting that virtually all the epidemiologically expected stroke reduction appears rapidly. For stroke the overview provides direct and highly significant evidence that both fatal and non-fatal strokes are prevented within just a few years of lowering blood pressure, even if the diastolic blood pressure is below 110 mm Hg at the start of treatment. There were proportionally similar sized reductions among those with mild, moderate, and severe hypertension.47 This suggests that as part of a primary care programme, patients with borderline blood pressures of 140-160 mm Hg systolic and 90-100 mm Hg diastolic should also be included for consideration of approaches to reduce blood pressure.

Cardiac disease—After age, sex, and hypertension, cardiac disease is said to be the most important risk factor for stroke. In the hypertensive population of Framingham, the ratio of myocardial infarction to stroke was two in men and one in women, whereas it was six in men and three in women who were not hypertensive.5The study also showed in men that in the presence of coronary heart disease the relative risk of stroke was 2.5 in the presence of congestive heart failure, and 4.4 in the presence of left ventricular hypertrophy. The risks were more pronounced in females (3.2, 4.3, and 7.9 respectively). The Evans County Study26 found similar but weaker evidence for this. Conversely, Welin and colleagues14 in multivariate analysis, did not find coronary heart disease or left ventricular hypertrophy to be significant factors.

Left ventricular hypertrophy—Ten per cent of persons aged 30-62 can expect to have ECG changes of left ventricular hypertrophy within 10 years48 between 15%-57% and echocardiography up to 98%.49 Left ventricular hypertrophy on ECG indicated a high risk of stroke in the Framingham study (relative risk 7.5 in men and 8.6 in women). To some extent this is a reflection of the severity and duration of the associated hypertension, but the residual effect of left ventricular hypertrophy shown by ECG is substantial (fourfold to eightfold) after adjustment for a coexistent increase in blood pressure.48

Atrial fibrillation—Atrial fibrillation affects 2% of the population (0.5% at 50-59 to 8.8% at 80-89) and the Framingham Study has estimated that atrial fibrillation accounts for between 7% and 31% of all strokes in patients under the age of 60 and increases the risk of stroke by a factor of five (1.5% in the fifth decade to 23.5% in the eighth decade).50 51 Up to 20% of those with atrial fibrillation who have a first stroke have a further episode within one year.52 In addition, atrial fibrillation is associated with a greater early mortality in patients admitted to hospital with acute stroke53 and a greater risk of recurrent stroke54 although these associations were not found in the Framingham Study.55 Kopecky et al 56 suggested that lone atrial fibrillation in patients under 60 at diagnosis was not associated with increased risk of stroke but the cohort study did not include subjects with cardiovascular disease, hypertension, or diabetes at the onset. People with atrial fibrillation without underlying rheumatic heart disease are five to six times more likely to have a first stroke or transient ischaemic attack compared with a control population of matched age and blood pressure50; in those with atrial fibrillation and rheumatic heart disease the risk is 17 times greater than in controls. Gustafsson et al,57 in an ecological analysis in Sweden, suggested that treatment with anticoagulants or aspirin is cost effective provided that the risk of serious complications of haemorrhage is kept low (savings of £2 million per million inhabitants/year).

Trial evidence—Non-rheumatic atrial fibrillation increases the risk of stroke and five randomised controlled trials of the use of warfarin in the primary prevention of stroke have shown consistent reductions in the region of two thirds with moderate anticoagulation long term.58-64 Warfarin reduces the incidence of ischaemic stroke or embolus from 4.5% to 1.4% per year (31 events per 1000 treated). Only two of these trials also reported a beneficial effect of aspirin in patients with atrial fibrillation.58 65 The Stroke Prevention in Atrial Fibrillation Trial reported that primary events or death were reduced by 58% (P=0.01) by warfarin and 42% by aspirin (P=0.02). The risk of significant bleeding was 1.5%, 1.4%, and 1.6% per year in patients assigned to warfarin, aspirin, and placebo respectively. In this trial, very stringent exclusion criteria were applied for eligibility for the warfarin group and monitoring was extremely thorough, which is reflected in the low risk of bleeding. In the non-trial setting, where criteria are not always met, the benefits will be reduced.

In the SPAF II study,66 although warfarin seemed more effective than aspirin for the prevention of ischaemic strokes, this agent was associated with more haemorrhagic strokes, which are generally more disabling than ischaemic strokes. In patients less than 75 years of age, without a history of hypertension, heart failure, or previous thromboembolism, the absolute risk of ischaemic stroke or embolism with aspirin was 0.5% per year. Previously warfarin has been recommended unless there were contraindications for anticoagulation. In the light of the SPAF II results treatment with aspirin can now be recommended for many patients because of its efficacy, safety, cost, and ease of administration. Warfarin is indicated for high risk patients such as those who have had a previous transient ischaemic attack or stroke.

Although those over 75 have the highest risk of a stroke from atrial fibrillation, they are at higher risk of bleeding with anticoagulants. These trials need to be replicated in the primary care setting if management is to be based in primary care. Further trials of the benefits and risks of various combinations of doses of aspirin and warfarin are also required.

The European Atrial Fibrillation Trial Study Group67concluded that anticoagulation is effective in reducing the risk of recurrent vascular events in patients with non-rheumatic atrial fibrillation with a recent transient ischaemic attack or minor ischaemic stroke. In absolute terms 90 vascular events (mainly stroke) are prevented if 1000 patients are treated with anticoagulants for one year. Aspirin prevents 40 such deaths. The same group have analysed the optimal intensity of anticoagulation.68

The Copenhagen Atrial Fibrillation Aspirin Study69 showed a 12% reduction in serious vascular events (stroke, myocardial infarction, or vascular deaths) and the Antiplatelet Trialists Collaborations overview of 25 trials70 reported a 25% reduction across all types of vascular death with aspirin therapy. Most of the trials to date evaluating the benefit of aspirin in these patients have produced results which fail to reach significance and larger trials are required to determine the benefit. The SPAF II study has continued to randomise patients between aspirin and warfarin and the results will inform this debate.66

Community stroke registers have estimated the prevalence in first time stroke patients of cardiac emboli to be between 8%-20% and ischaemic heart disease between 13%-38%.71 The scope for secondary prevention is therefore considerable.

Previous transient ischaemic attack— Whisnant40 estimated the attributable risk of a previous transient ischaemic attack to be 10% and estimates of a 13-fold increased risk of stroke for subjects who have had a transient ischaemic attack have been found in the first year after a stroke and a sevenfold risk for the seven years thereafter. Eighty per cent of transient ischaemic attacks originate from emboli in the region of the carotid artery. The risk of atrial fibrillation increases with age but has an overall prevalence of 0.5%-2%. About 20% of patients with first time stroke have had one or more preceding transient ischaemic attacks. The risk of stroke after a transient ischaemic attack is estimated at 5% a year, which corresponds to a 13-fold increased risk of stroke.72

Carotid artery stenosis—Several large trials have reported interim or final results assessing the benefits of carotid endarterectomy and to determine the level of perioperative risk at which the procedure becomes acceptable: the European Carotid Surgery Trial (ECST),73 the North American Symptomatic Carotid Endarterectomy Trial (NASCET),74 and the Veterans Affair Cooperative Studies Program,75 all published in 1991 with similar findings. The first two trials recruited patients with a history of a minor ischaemic event and severe stenosis of the origin of the ipsilateral internal carotid artery (70%-99%). The risk of stroke or death within 30 days of surgery was 7.5% in ECST and 5.8% in NASCET. If the patient survives stroke free, the risk of stroke in the ipsilateral territory over the next few years, assessed by life table analysis, was 2.8% for surgical patients and 16.8% for controls at three years in the ECST study and 9% for surgical patients and 20% for controls in the NASCET study. The beneficial effect of surgery is such that between five and 10 patients must be operated on to prevent one from having a stroke. Only half the strokes prevented are disabling. If all those patients with severe carotid artery stenosis who had had a transient ischaemic attack underwent surgery the overall reduction in the total incidence of stroke would be between 0.2% to 0.5% but there would be a resultant fivefold increase in carotid surgery.76 77 Endarterectomy is not indicated for most, possibly all, patients with moderate symptomatic carotid stenosis.78

The lack of conclusive evidence for endarterectomy in asymptomatic disease led to the establishment of two multicentre trials, the Asymptomatic Carotid Arterosclerotic Study (ACAS) and in its European counterpart the Asymptomatic Carotid Surgery Trial (ACST). Recent ACAS data suggest that if stenosis is greater than 60% endarterectomy is beneficial with a relative risk reduction of 55% but absolute risk reduction of 5.9%.79 The problem with the study is that 9% of the study group did not receive the allocated treatment and median follow up time was short.80

Secondary prevention after transient ischaemic attack or ischaemic stroke—The Antiplatelet Trialists Collaboration performed a meta-analysis of trials assessing the benefit of aspirin therapy in secondary prevention of stroke and estimated that long term anticoagulant therapy reduces the risk of stroke and other major vascular events by about 25%.81 This benefit has been shown clearly for doses of aspirin of 300 mg to 1300 mg/day but it remains unclear whether low dose aspirin would be effective for certain high risk groups.

There have been two recent clinical trials suggesting the benefit of lower doses of aspirin. The Dutch transient ischaemic attack trial82 compared 30 mg aspirin with 300 mg, and in the first group, stroke myocardial infarction or vascular death occurred in 14.7% versus 15.2% in the 300 mg group. However, there were too few events for the difference to reach significance. The second trial, the Swedish Aspirin Low dose Trial,83 was the first to show a clear beneficial effect of aspirin with daily doses of less than 300 mg. This group compared the effect of 75 mg aspirin with placebo and showed a significant 17% risk reduction in the treatment group compared with placebo.

The United Kingdom Transient Ischaemic Attack Study Group84 results also suggested a similar beneficial effect for different aspirin doses above 300 mg, but that side effects are probably dose dependent. The beneficial effect of aspirin seems to be similar in people with hypertension and diabetes, in all age groups, and in both sexes.

Ticlopidine is another antiplatelet agent which has been shown to be effective as a secondary preventative agent with similar if not greater efficacy than aspirin, but which is associated with significantly more side effects. It is recommended for those who are unable to tolerate aspirin. Interestingly, its mechanism of action is different from aspirin and there is some evidence that combined aspirin and ticlopidine therapy has a greater beneficial effect than each individually, but with considerably increased bleeding.85-87 A related thienopyridine drug (clopidogrel) has been shown to have fewer side effects but offers few advantages over aspirin.88

With regard to secondary prevention after a transient ischaemic attack, it is important to note that only 12% of strokes are preceded by transient ischaemic attack and only about half of these are reported to a doctor. Therefore if secondary prevention is introduced at this stage, this could prevent no more than 5% of all strokes.

The Antiplatelet Trialists Collaboration70 undertook an overview of randomised trials of prolonged antiplatelet therapy. They concluded that there is as yet, no clear evidence on the balance of risks and benefits of antiplatelet therapy in primary prevention among low risk subjects. Among a much wider range of patients at high risk of occlusive vascular disease than is currently treated routinely (for example, unstable angina, suspected acute myocardial infarction, stroke, transient ischaemic attack, and other high risk groups such as those with stable angina or peripheral vascular disease), some years of antiplatelet therapy, with 75-325 mg aspirin a day or some other antiplatelet regimes, offers worthwhile protection against myocardial infarction, stroke, and death.

The same group89 indicates that antiplatelet therapy, chiefly aspirin alone or aspirin plus dipyridamole, greatly reduces the risk of vascular occlusion in a wide range of patients at high risk of this complication, but when treatment should be started and for how long requires further study.

Compliance and safety with warfarin are less likely to be assured when warfarin is prescribed in general practice than in the selected trials.90

Obesity—Many studies fail to distinguish obesity from its associated feature, hypertension, and it is not clear how much of an independent effect obesity has on the incidence of stroke. The Framingham study provides some evidence for obesity as an independent factor5; however, much of the effect of obesity is mediated through other risk factors. Larsson et al 91 showed clearly that an increased waist:hip ratio was associated with stroke but was not an independent long term predictor.

Diabetes mellitus—The Framingham study documented a relative risk of stroke of 2.6 in men and 3.1 in women with diabetes and an attributable risk of about 10%.55 More recently Stegmayr and Asplund,92 using population based registry data, estimated the risk at 4.1 in men and 5.8 in women. They estimated that the attributable risk was 18% in men and 22% in women and 50 strokes are annually directly attributable to diabetes per 100 000 population, although how the authors controlled for other factors such as hypertension and atrial fibrillation is unclear.

Diabetic patients with proteinuria or ischaemic ECG abnormalities are at highest risk of new cerebrovascular disease.93

Infection—Syrjanen et al,94 in a case-control study, estimated that febrile infection in the previous month to a stroke in patients aged under 50 inferred a relative risk of nine (95% CI 2.2-80). The most common infection was respiratory (80%), with a relative risk of 14.5 (95% CI 1.9-112.3) estimated in conditional logistic regression analysis for matched pairs. The study used community controls. A case-control study, again using community controls, by Grau et al 95 indicated that recent infection, primarily of bacterial origin, may be a risk factor for cerebrovascular ischaemia in older as well as younger patients.

Homocysteinuria—Premature arteriosclerosis and thromboembolic events are well known complications of homozygous homocysteinuria due to cystathionine syntheses deficiency which damages vascular endothelial cells. Heterozygosity for homocysteinuria can be assessed after methionine loading and is said to occur in one in 70 of the population and predisposes to premature occlusive arterial disease, including cerebral disease.96 Perry at al 97 reported in the British Regional Heart Study that total homocysteine concentrations are a strong independent risk factor for stroke with a graded increase in the relative risk of stroke in the second, third, and fourth quartiles of the total homocysteine distribution (odds ratio 1.3-2.8). Nygard et al 98 in a survey of a community in western Norway, showed that raised plasma total homocysteine concentrations were associated with major components of the cardiovascular risk profile—that is, male sex, old age, smoking, high blood pressure, raised cholesterol, and lack of exercise.

Sickle cell disease—Serjeant99 estimated that a patient with sickle cell disease has a roughly 10-fold increased risk of stroke compared with the general population. The risk is greatest for homozygous HbSS subjects, but heterozygote carriers also carry significant risk.

Fibrinogen—Increase in the fibrinogen concentration increases blood viscosity, which may further enhance the risk of thrombus formation. Wilhelmsen et al 100showed a significant relation with stroke when the systolic blood pressure was above 200 mm Hg and fibrinogen 5-6 g/l, when a 12% incidence of stroke was seen compared with 0 if the systolic was between 100-120 mm Hg and fibrinogen 1-3 g/l. The Framingham study5 showed a significant independent relation of fibrinogen with stroke but the coefficient was not significant. Welinet al 14 estimated the population attributable risk to be 0.18 and in multivariate analysis fibrinogen was shown to be an independent risk factor. Qizilbash et al,101 in a case control study adjusting for other variables, estimated the odds ratio for ischaemic stroke to be 1.78 (0.91-3.18) for fibrinogen greater than 3.6 g/l.

Clotting factors—In the Northwick Park Study, Meade and colleagues102 showed raised factor VII concentrations to be the strongest predictor of myocardial infarction five years before the event, followed by raised cholesterol and fibrinogen concentrations.

There has also been recent attention to antiphospholipid antibodies such as anti-cardiolipin antibody and lupus anticoagulant. These may be found in association with various systemic diseases, but most commonly in systemic lupus erythematosus. In some instances they are an isolated finding with no clinical correlates of connective tissue disease.103 These antibodies are associated with increased risk of thrombotic episodes both in venous and arterial vessels. In a study of 100 consecutive patients admitted to hospital with first ever strokes, Chakravarty et al 104 found that 21% were positive for anticardiolipin antibody compared with none in 100 age matched controls. In the patient group which was positive for anticardiolipin antibody, the three month case fatality was 62% compared with 21.5% in the antibody negative group. In addition, the proportion of patients with significant disability after stroke (Barthel score 0-9) was significantly higher in the antibody group. The authors suggested that the presence of anticardiolipin antibodies may represent an independent prognostic marker for mortality and outcome from stroke. Studies have been on selected groups and long term follow up of patients is required to assess whether the presence of antibodies is an epiphenomenon.

Packed cell volume—Increased blood pressure and cigarette smoking increase the packed cell volume and account for most of the association with stroke. The Framingham study showed a significant independent correlation in the under 65s only.5Welin et al 14 was not able to show an association. Yiyohara et al 105 found that a low packed cell volume was important as an independent risk factor in women.

Behavioural characteristics

Physical activity—In the British Regional Heart Study physical activity was inversely associated with risk of stroke independent of coronary risk factors, heavy drinking, and pre-existing ischaemic heart disease or stroke (relative risk 1 for inactivity, 0.6 moderate activity, and 0.3 vigorous activity).106 Also in the Framingham study at the 14 year follow up point there was a similar inverse relation between physical activity and risk of stroke107 but this was not significant after adjustment for age and the classification of physical activity was broad.

Lipids—The story is far less clear or consistent than for coronary heart disease. In the Framingham study there was a negative association of low density lipoprotein cholesterol with stroke, especially in women over 65 after other risk factors had been accounted for. There was no clear cut relation with triglycerides and no protective effect of high density lipoprotein.5 55

In a meta-analysis on 10 prospective studies Qizilbash et al 101 108 described a pooled relative risk of 2.9 (95% CI 1.43-5.87) for people with a total cholesterol of greater than 5.7 mmol/l compared with lower concentrations. In white people, the population attributable risk of stroke due to a cholesterol concentration of greater than 5.7 mmol/l has been calculated at 0.2. In a case-control study the same group found fibrinogen and lipids to be important risk factors for ischaemic stroke. Adjusted for other variables the odds ratios of ischaemic events were 1.73 (0.9-3.29) for total cholesterol >6 mmol/l; 1.34 (0.69-2.61) for low density lipoprotein >3.5 mmol/l; and 0.32 (0.15-0.69) for high density lipoprotein >1.2 mmol/l.109

At the six year follow up point in the Multiple Risk Factor Intervention Trial Research Group (MRFIT) study, Iso et al 110 showed that a reduced serum cholesterol concentration was associated with a threefold increased risk of intracranial haemorrhage after controlling for age, smoking, diastolic blood pressure, and race or ethnic group. The group stressed that the excess risk was almost exclusively among men with high blood pressure and that intracranial haemorrhage accounted for only 2.9% of all deaths from cerebrovascular disease. There was an overwhelming positive association between high serum cholesterol and death from non-haemorrhagic stroke and total cardiovascular disease.

Cigarette smoking—A meta-analysis of the relation between cigarette smoking and stroke provided strong evidence of an excess risk of stroke among smokers, with an overall relative risk of 1.5 (1.4-1.6).111

Alcohol—Alcohol is a risk factor for hypertension and studies have shown a positive association between alcohol consumption and stroke which may be linked with hypertension rather than alcohol acting independently.112 There is evidence that chronic and heavy drinking (>60 g ethanol/day) are associated with increased risk of all types of stroke.112 The role of low or moderate alcohol intake remains unclear.

Evidence from over 19 prospective cohort studies provides no strong evidence of any association for total stroke except for an apparently modest increase in risk in the highest level of alcohol intake. When the data are segregated according to the two broad categories of stroke type, consistent patterns emerge. For haemorrhagic stroke, there is strong evidence for an association, with typically about a threefold excess risk for higher levels of alcohol consumption. By contrast, there is equally strong evidence against any material adverse association with ischaemic stroke.112-116

Potassium—In the INTERSALT study117 potassium was negatively associated with blood pressure. Khaw et al 118 estimated the relative risk of stroke in a Californian cohort in those whose potassium intake was in the lowest tertile compared with the top two tertiles to be 2.6 in men and 4.6 in women. In multivariate analysis a l0 mmol increase (one serving of fresh fruit and vegetable) in daily potassium intake was associated with a 40% reduction in risk of stroke associated mortality which was independent of other dietary variables and known cardiovascular risk factors.

The oral contraceptive—There are reports of a fourfold increase in haemorrhagic stroke in people over 35, particularly those who smoke.119 The relative risk of ischaemic stroke is about three and the absolute risk is estimated at one in 10 000 oral contraceptive users. The WHO Collaborative Study on Cardiovascular disease and Steroid Hormone Contraception estimates that between 13%-18% of all strokes in women aged 20-44 in Europe and the developing countries, respectively, are attributed to the use of oral contraceptives.120

Hormone replacement therapy—Meade and Berra121reviewed 10 cohort studies and estimated a 15% reduction in stroke deaths in women on hormone replacement therapy, although there was considerable variation between centres. The Nurse’s Study in the United States reported no significant reduction in ischaemic stroke with combined oestrogen and progesterone hormone replacement therapy, with a suggestion of an increased risk of stroke in the subgroup of women taking the highest doses of unconjugated oestrogen.122

Vitamin C—Acheson and Williams123 in an ecological analysis showed a possible protective effect of vitamin C in fresh fruit and vegetables on occurrence of stroke.


Strategies to reduce incidence or recurrence of stroke can be either directed to those at high risk for a particular factor or the whole population. Both approaches are probably required to reduce the incidence of stroke but the optimum strategy is still to be determined. The high risk approach may involve population screening or opportunistic case finding and the selective treatment of the few people at highest risk. An alternative, complementary, strategy is to try and produce a small reduction in the risk factor in every person in the community by some form of intervention.

There have been several major intervention trials at a community level, largely set up to determine the effectiveness of risk factor reduction strategies on cardiovascular disease but with cerebrovascular mortality also being recorded. Most of the evidence is also based on males.124-126

The task force on arteriosclerosis convened by the National Heart and Lung Institution in the United States recommended that a Multiple Risk Factor Intervention Trial (MRFIT) in those at high risk of cardiovascular disease (aged 35-57) due to combinations of raised lipids, hypertension, and cigarette smoking be set up to determine whether an intervention programme would result in a significant reduction in mortality from cardiovascular disease, non-fatal myocardial infarct, and cerebrovascular mortality.

After 10.5 years of follow up a 13% reduction in fatal stroke in the special intervention group was found.4 When considering follow up, black subjects and smokers were more likely to discontinue their participation between screening examinationss.126The effect that healthier lifestyle practices had on control subjects is not known but a possible effect may have been detected in the intervention arm.

In north Karelia, Finland127 a whole community package to represent “real life” tested the feasibility of such intervention programmes by using existing services and other resources. In the 10 years 1972-82 in 30-59 year old subjects there were small but significant reductions in smoking, cholesterol, and blood pressure. Despite several studies showing significant reductions in risk factors this has not been formulated into similar reductions in death rates for coronary heart disease or stroke in the study populations. The World Health Organisation factors study of multiple intervention for risk factors for coronary heart disease showed no significant reductions in deaths from coronary heart disease after six years of intervention.128

However, Vartiainen et al 129 estimated the extent to which the changes in the main cardiovascular disease risk factors (blood pressure, smoking, and serum cholesterol concentration) could explain the changes in mortality from stroke in north Karelia during the past 20 years. They found that two thirds of the fall in mortality from stroke in men, and half in women, could be explained by changes in the three above cardiovascular disease risk factors.

General health checks by nurses in England are ineffective in helping smokers to stop smoking, but they help patients to modify their diet and total cholesterol concentration.130 The public health importance of this dietary change depends on whether it is sustained. The British family heart study is a randomised controlled trial in general practices in 13 towns in Britain to measure the impact of a programme of cardiovascular screening and lifestyle intervention led by nurses.131 As most general practices in the United Kingdom are not using such an intensive programme the changes in coronary risk factors achieved by the voluntary health promotion package for primary care are likely to be even smaller. It seems that this style of approach to the population through primary care alone is not going to produce large reductions in the risk of cardiovascular disease. Instead more effective legislation needs to be in place to control the use of tobacco and promote the consumption of healthy food.


A knowledge of the role and prevalence of risk factors for stroke is necessary before any attempts can be made to reduce them. However, to develop a health promotion strategy for primary and secondary prevention of stroke, important considerations are the patients’ own attitudes; their perception of personal risk perceived as controllable or hereditary and therefore fixed, and their desire to modify their controllable risk behaviour.

Very little work has been done in this field with regard to stroke, patients perceptions or the social image, and meaning of stroke in society. A study of patient perception of cardiovascular risk by Silagyet al 132 showed that the health risks of smoking and lack of exercise were recognised by most people. Attitudes towards diet were more complex. Being overweight was the main predictor of a person’s perception of diet as harmful to health. Interestingly there were major sex differences in attitude towards the role of diet in that obese women were more likely to think that their diet was harmful than did obese men. Motivation to reduce risk factors was high for smoking but low for exercise.

There are also important differences in attitudes and lifestyle between ethnic groups reflecting cultural differences in lifestyle and social habits. Cigarette smoking and alcohol consumption seem to be higher in white than black men133 and may form an integral part of social activities in this population, which may therefore be resistant to change. The control of weight is influenced by attitudes to body size held by different cultural groups and by the practical problems associated with changes in diet in a family context. Differences in attitudes towards western medical practice between different cultural groups is an important factor determining compliance with recommended drug therapies for primary and secondary stroke prevention.

When developing a programme of stroke prevention, not only does the potential benefit in stroke reduction need to be considered by individual risk factor modification, but also the patients’ own perception of personal and social costs of behavioural change.

Diabetes mellitus

Diabetes is a group of disorders characterised by chronic hyperglycaemia, which is a major cause of morbidity and early mortality. Neuropathy is a common complication and can contribute to ulceration of the feet and painless arthropathy. Neuropathy, defined as the presence of bilateral impaired vibration perception, was present in one study in 17% of diabetic patients aged 20-59, 42% in patients aged 60-69, and 10% of patients aged 70 or over.134Autonomic neuropathy can lead to postural hypotension, nocturnal diarrhoea, impotence, an atonic bladder with overflow incontinence, and vasomotor phenomena. Myocardial infarction and strokes resulting from macrovascular disease are the commonest causes of death in people with diabetes over the age of 50.135

The Diabetes Control and Complications Trial Research Group136 indicated that intensive therapy with insulin with careful monitoring reduced the incidence of neuropathy by 60 (95%CI 38-74)%. The sample size was too small to detect differences in macrovascular disease in this young group of patients but a non-significant 41% reduction in events was seen.

Non-vascular events


In 1990 in the United Kingdom there were at least 4191 deaths from fracture of the skull and intracranial injury.137 Tiretet al 138 estimated in France that there were three peaks of admissions—in those under 5, those aged 25-24, and those older than 75. Sixty per cent were due to road traffic accidents and 30% due to falls.

National and local government policies are important in the prevention of accidents, as is the role of health promotion. Health services are relevant to reducing case fatality rather than preventing the disease occurring in the first place and will not be discussed further.

Accident prevention in young people may be possible by teaching them the skills to cope with traffic and the road systems. Teaching young children about crossing the road and giving bicycle lessons are important, as is improving their visibility. Thompson et al 139 found that bicycle safety helmets were highly effective in preventing head injury. Other effective measures include crash helmets for motorcyclists and riding hats for horse riders.

A major contributing factor in male deaths between the ages of 15 and 24 years is alcohol.140 Campaigns against alcohol may be effective. In Oklahoma legislation that increased the minimum age for the purchase of alcohol, made the limit of blood alcohol while driving 22 mmol/l, and facilitated licence withdrawal from suspected “drunk drivers”, may have been responsible for reducing motor vehicle fatalities and fatal crashes by a third.141 Other legislation such as the greater use of speed limits and the enforcement of regulations on the use of front and rear seat belts could help to reduce road traffic accidents or minimise the injuries that result from them.142

Head injuries caused by falls are a feature of children and elderly people. They may be prevented by identifying possible accident risks in the home and removing them.143 Particularly in elderly people, acute illness may present with falls. Early diagnosis and treatment, combined with skilled evaluation of the living environment to remove hazards, may prevent further falls.

The Royal College of Physicians Working Party on Prevention144 indicated that preventive measures would best be achieved through law enforcement, education and the use of helmets and seat belts. With regard to alcohol they suggested prices should rise, advertising should be curtailed, bottles should clearly state the alcohol content of the beverage, low alcohol beverages should be promoted, and there should be better coordination between all the agencies involved.

The Faculty of Public Health Medicine set targets for accidents reducing the deaths from 1790 in 1979 to 55 in 2000.145The Health of the Nation strategy has set targets for accidents to reduce the death rate among children aged under 15 by at least 33%, among 15-24 year olds by 25%, and among those aged over 65 by at least 33% by 2005.146


A recent conference in Edinburgh The challenge of the dementias identified that research in genetics and in imaging is beginning to throw light on susceptibility factors and to offer the potential for earlier diagnosis.147

With vascular causes of dementia now back on the agenda and with ideas about roles for oxidative stress, we may be seeing the first hints of preventive possibilities that march in parallel to improved physical and mental health.147 Yoshitake et al 148 performed one of the largest population based studies of incidence of dementia and risk factors in Japan with very high postmortem rates. In males the commonest cause of dementia was vascular in origin, contrary to reports from other parts of the world that Alzheimer’s disease is the predominant pathology.


The only risk factors for Alzheimer’s disease that have been unequivocally reported are age, sex, and family history, none of which are yet amenable to influence. There is some evidence to support the negative association between smoking and Alzheimer’s disease.149

The Rotterdam Study150 showed an inverse dose-response relation between education and dementia, in particular Alzheimer’s disease. This was not due to the confounding of cardiovascular disease. This confirms the Canadian Study of Health and Ageing,151which showed that those with less education were at higher risk of Alzheimer’s disease with an odds ratio of 4 (95% CI 2.49-6.43). The opportunity for primary prevention could therefore be directed at improving education levels in the population.


The associations of dementia with hypothyroidism, vitamin B12 deficiency, syphilis, and HIV are well established. Primary prevention of these conditions may reduce the burden of dementia in society. Prenatal screening for Down’s syndrome and Huntington’s chorea provides an opportunity for prevention of some cases of dementia.


Compared with stroke relatively little research into potential risk factors for multi-infarct dementia has been undertaken. It seems likely that there are close links between these two conditions. Currently two studies are under way looking at risk factors but using different methodologies and risk factors can only really be described as putative.152 153

Age, sex, race, education, and risks for cerebral infarction have all been cited as risk factors. Ladurner et al 154found that a history of hypertension was more common among cognitively impaired patients with cerebral infarction than in those with cerebral infarction without cognitive impairment. Gorelick et al 155 showed that a history of recent cigarette smoking and history of myocardial infarction, both of which are associated with atherosclerosis, independently predict cognitive impairment associated with multiple cerebral infarcts among predominantly African American multi-infarct patients. Furthermore, higher systolic blood pressure level exerted a protective effect. Tatemichi et al 156 on the other hand found that hypertension and cardiovascular disease were not associated with dementia after stroke but that diabetes mellitus was. The role of other risk factors needs to be clarified.


In Washington (United States)157 8% of males and 14% of females missed all or part of work or school due to a headache in a four week period before interview.

It is estimated that the annual cost of lost productivity due to migraine in the United States is between $1.2-17.2 billion dollars. Direct cost to health services is estimated at more than one million doctor visits per year.158

The available evidence supports the view that familial and genetic factors play a part in determining risk of migraine. The precise genetic pattern of transmission remains to be determined.159

With current knowledge there seems to be no role for primary prevention.


Table 6 indicates the putative risk factors for multiple sclerosis, most of which are not amenable to primary prevention. The evidence for trauma being a major cause of the disease is not strong.161

Table 6

Multiple sclerosis risk factors

Several viral infections have been considered, including measles and HTLVI as causes of multiple sclerosis. The evidence, however, that measles vaccination protects against multiple sclerosis is conflicting.162 163

No primary preventive strategy is possible with the current understanding of multiple sclerosis aetiology.


There have only been 10 or so incidence studies, and there are problems with case definition. The prevalence of parkinsonism increases greatly with age (14.9% at 65-74 years, 29.5% at 75-84 years, and 52.4% at 85+ years).164

Martyn and Osmond,165 in a case control study, found no evidence of any link between birthweight and weight at 1 year of age or domestic environment on subsequent risk. Cases were more likely than controls to have never smoked (odds ratio 2 (95% CI 1.1-3.6)). In a 26 year follow up of over 8000 males in the Honolulu heart programme, men who smoked had a reduced risk of developing the disease (relative risk 0.39).166

There has been a suggestion167 that rural living, well water consumption, and pesticide exposure increase the risk. Rural living may also explain the association with two ubiquitous rural associated pathogens coronavirus and Nocardia asteroides, these having recently been proposed to cause Parkinson’s disease.168 169


Different case-control studies have identified various risk factors including previous mechanical, clinical, and electrical trauma, heavy metal exposure and occupations dealing with leather, physical labour, and physical activity.170 No associated antecedent medical condition was found in one case-control study.171

A case control study in Scotland172 identified an association between excess fracture rate, manual occupation, and occupational exposure to lead and solvents and increased risk of motor neuron disease. No link was found with socioeconomic deprivation in childhood, childhood infections, or social class. If these findings are confirmed there is potential for primary prevention in this area.


There has been a wide variation of definitions of epilepsy used in epidemiological studies. In total population studies in developed countries defining epilepsy as recurrent unprovoked seizures, age adjusted incidence rates vary from 24 to 53 per 100 000 person-years. Studies of first unreported seizures provide somewhat higher estimates ranging from 26-70 per 100 000 person-years.173

In a study in Rochester, USA, epilepsy attributable to cerebrovascular disease (8% of cases) was the most important single identified aetiology among active prevalent cases. This was followed by epilepsy associated with neurological deficits from birth (5%). Head injury and infection of the CNS are the most important groups in middle adulthood and cerebrovascular disease and degenerative brain disease in elderly people.173 It is also associated with brain tumours, alcohol, and multiple sclerosis (see other sections).

Infectious diseases

One of the greatest developments in medical care has been the ability to protect children and adults from some of the infectious diseases that have several neurological sequelea, through immunisation and vaccination.

Active immunisation has been decisive in the control of polio and measles.


Encephalitis is rare, occurring in less than 1/1000 cases. Complications of encephalitis such as mental retardation, residual seizures, and hemiparesis occur in 25% of cases.

Vaccination started in the United Kingdom in 1968. By 1988 a combined measles, mumps, nd rubella (MMR) vaccine was introduced and coverage at two years is in the order of 92-93%. By 1991 notifications of the disease had reached an all time low although there were cases in older children prompting a campaign in the over 10s. A CDR review174 indicated a sharp rise in the number of cases of measles in 1994. Even though coverage rates are high the children who remain unvaccinated and those in whom vaccine fails to evoke immunity will accumulate, allowing measles transmission to continue with a resurgence every seven to eight years, and two dose strategies could be employed to overcome this.


The major organisms to consider for bacterial meningitis areHaemophilus influenzae (HI) type b, Neisseria meningitidis (NM), and Streptococcus pneumoniae(SP).175

Haemophilus influenza meningitis most often occurs in infants and young children and only exceptionally in the newborn or adults. The epidemiology of Haemophilus influenza meningitis is now changing because of increasing use of effective vaccines developed in the 1970s and composed of Haemophilus influenzatype b capsular polysaccharide. New vaccines with improved immunogenicity were developed in the 1980s by conjugating polysaccharide with protein antigens. Since 1990 they have been effectively used in children at least 2 months old.

In the United Kingdom nearly all Neisseria meningitidisdisease is caused by serogroups B and C. Group B causes 70% of infections and C 25% mainly in infants.

The development of vaccines that protect infants is a significant challenge. Phase two trials are underway for Group B. Group C vaccines do not protect the under 2s. Almost certainly a serogroup C conjugate vaccine is likely to be available for the United Kingdom in the next few years but the prospects for a group B vaccine are uncertain.

The control of an outbreak is crucial to the control of the disease with rifampicin, ciprofloxacin, and ceftraxone all being recommended for use in preventing secondary cases of meningococcal disease.


Poliomyelitis is rare in developed countries but continues to be a problem in developing countries. Continuing vaccination programmes and control of water supplies remain crucial in controlling the disease.


Stroke is responsible for over 80% of deaths from neurological causes in the United Kingdom. We have outlined the multiple risk factors associated with stroke, many of which have been shown to be effectively reduced in trials.

Whether reducing risk factors in individual subjects or at a population level results in a reduction in stroke mortality is less convincing. Identification of the key risk factors for stroke subtypes and effective strategies for reducing them is now required.

The lack of knowledge of the risk factors for many other neurological diseases limits the opportunity for effective prevention. This must be the major challenge for neurologists for future years.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.