Clinical outcome in 287 consecutive young adults (15 to 45 years) with ischemic stroke

Subjects and methods.

Results.

Of 287 patients, 247 (86.1%) attended at least one neurologic visit, 179 (62.4%) attended two visits, 156 (54.4%) attended three visits, and 142 (49.5%) attended four visits or more. The final contact with the patients was a neurologic visit in 214 (74.6%) patients and a telephone call with the patient, a relative, or the general practitioner in the 73 (25.4%) remaining. Of 214 patients whose last contact was a neurologic visit, 139 (48.4% of the whole group) were seen >3 years after onset. The final neurologic visit was made by a neurologist from the Stroke Department in 201 patients (70.0% of the whole group), from another hospital in 11 patients (3.8%), and by a private neurologist in 2 patients (0.7%). The median duration of follow-up was 3 years. No patient was lost to follow-up. At the end of the follow-up period, 204 patients (71.1%) were treated with antiplatelet therapy (aspirin in 157, clopidogrel in 21, aspirin dipyridamole in 15, ticlopidine in 11) and 37 were treated with oral anticoagulant therapy (12.9%). Of the 46 remaining, 21 did not receive any antithrombotic drug anymore (7 dissections, 3 patients because of intolerance, 4 patients who decided to stop, and for an unknown reason in 7), the current therapy remained unknown in 3, and 22 patients were dead. An antihypertensive treatment was prescribed in 54 (91.5%) of the 59 survivors with arterial hypertension. Oral contraceptives were discontinued by 44 of the 45 survivors who were current users. Only 22 of 99 current smokers who survived stopped smoking.

Over a 3-year period of follow-up, 22 (7.7%) patients died: 13 during the first year (all from cerebral lesions) and 9 after the first year after stroke (cardiac in 4, cancer in 2, trauma in 1, and unknown in 2). The median delay between stroke and death was 1 year, ranging from 0 (death at the acute stage) to 3 years. The average annual mortality rate of the cohort was 4.5% (95% CI 2.1 to 6.9%) during the first year and 1.6% (95% CI 0.6 to 2.6%) during the subsequent 2 years. Ten patients (3.5%) had recurrent cerebral ischemia (one atheroma, one cardioembolism, two other definite, and six undetermined), none of them leading to death. Of 10 recurrent strokes, 4 (1.4%) occurred during the first year and 3 (1.0%) during each of the following 2 years. Two patients had a myocardial infarction (one atheroma, one cardioembolism) after 2 and 3 years. Nineteen patients developed epileptic seizures, which occurred in 15 during the first year and in 4 >1 year after onset.

The breakdown of recurrent stroke, myocardial infarction, epileptic seizures, and deaths in stroke subtypes at the end of the follow-up period is presented in table 3.We found no significant differences in mortality, recurrence rate, dependency, and epilepsy 1) between patients examined at a neurologic visit and those followed by telephone interview, 2) between patients directly admitted in the stroke unit and those referred from another hospital, or 3) according to the antithrombotic regimen at discharge and at the last visit or contact. The influence of those factors on the rate of myocardial infarction was not evaluated because only a few patients developed myocardial infarction. At the end of the follow-up period, 151 patients had an mRS score of 0, 72 a score of 1, 26 a score of 2, 10 a score of 3, 2 a score of 4, 4 a score of 5, and 22 a score of 6. Of 265 survivors, 249 (94.0% of survivors, 86.8% of all patients) were independent (mRS = 0 to 2). In two patients with an mRS score of 1 and in three with a mean mRS of 2, the disability was due to focal upper limb dystonia.

A bivariate comparison of demographic factors, risk factors for stroke, causes of stroke, stroke characteristics, and outcome events between patients with bad (mRS = 3 to 6) and good (mRS = 0 to 6) outcomes is reported in table 4.

At the end of the follow-up period, of 265 survivors, 8 (3.0%) did not want to work for personal reasons, 142 (49.5%) returned to the same work (10 with adjustment in their occupation), 30 (10.5%) found a new job, and 43 (15.0%) were considered by the social insurance as unable to work for medical reasons: The mRS scores were 3 to 5 in 15 patients, 2 in 15 patients, 1 in 12 patients, and 0 in 1 patient with depression. Of the 42 remaining who wanted to work but were unemployed, 30 were already unemployed before stroke and 12 lost their work after stroke despite an apparently good recovery (mRS = 0 in eight patients and 1 in four); these 12 patients had attended school for 8 to 11 years. Twenty patients (7.0%), including 14 men and 6 women, spontaneously reported changes in their family structure with divorce.

Discussion.

Our study showed that 3 years after an acute ischemic stroke occurring between 15 and 45 years of age, 1) the annual mortality rate was 4.5% during the first year and then 1.6%; 2) the annual rate of stroke recurrence was 1.4% during the first year and then 1.0%; 3) there was a 0.2% annual rate of myocardial infarct; 4) 87.5% of patients were independent (Rankin score = 0 to 2); 5) factors associated with a good outcome (mRS = 0 to 2) at the bivariate analysis were decreasing age, absence of alcoholism or diabetes mellitus, and presence of hypercholesterolemia, or treatment with statin or fibrate, or oral contraceptive therapy; 6) the cause of stroke, stroke recurrence, and occurrence of myocardial infarction or epileptic seizure did not influence the outcome; 7) 4.2% of patients lost their work after stroke despite an apparently good recovery; 8) 7.0% of patients reported divorce; and 9) if most patients were still treated according to guidelines in most domains, only 22% of smokers gave up cigarette smoking.

Our study included one of the largest cohort of young adults with ischemic stroke followed up in a single center. No patient was lost to follow-up, suggesting that our results are valid in terms of case fatality. The 74.6% rate of neurologic visits at the end of the follow-up, with telephone contacts in the remainders, is close to the rates seen in other studies.9 This study was observational, meaning that no test was done specifically without being part of the current practice in our center, except in patients included in another study. The major advantage of such an observational study is to evaluate patients in conditions as close as possible to true life and to include all consecutive patients admitted during the study period. However, with patients having had a diagnostic test only when this test was thought to be useful, a small degree of heterogeneity in the diagnostic workup is possible. Therefore, the frequency of factors that can be identified only with optional procedures, such as interatrioseptal abnormalities in TEE, may be underestimated. Accordingly, because no reference to ethnicity is allowed in patients’ files according to the French law, this variable cannot be taken into account in any observational study. However, this is not a major issue, because the proportion of non-Caucasian patients in our area is very low, and we can consider that almost all (and maybe all) patients were Caucasian. We also cannot exclude an underestimation of toxic angiopathies in our study population because a toxicology screen was not performed in all patients, but only when no obvious cause was found with the very first examinations, provided the patient gave consent, as required by the French law.

Studies conducted in university hospitals may suffer recruitment bias9: Tertiary-level hospitals may enroll more complicated cases,20 but they may also recruit patients with etiologies requiring specific investigations that are not available in other centers. Our center is the only one providing the expertise of a senior in-house neurologist and neuroradiologist, with neuroimaging facilities 24 h/day, for a population of almost 4 million inhabitants. Therefore, a recruitment bias probably explains the high rate of cervical artery dissections. Changes in the number of patients recruited in 1995 and after are probably the consequence of the presence 24 h/day of a senior neurologist and the availability of neuroimaging facilities after this date. This organization might have led to admission of patients who would probably have been admitted in other hospitals before 1995 and probably skewed the recruitment toward younger patients and causes requiring specialized neurologic and neuroradiologic expertise, such as dissections.

We included only patients with arterial ischemic strokes. Many other studies also included patients with hemorrhagic strokes,6-8,25⇓⇓⇓ subarachnoid hemorrhages,6-8,25⇓⇓⇓ and TIA,4,9,14,18,19⇓⇓⇓⇓ who have different mortality and recurrence rates. Accordingly, young ischemic stroke patients accounted for only 4.8% of patients admitted in our Stroke Department during the study period. This proportion is in the lower range of the proportion of young patients in most stroke units, ranging from 4 to 12%,11 and is lower than the percentage of young patients admitted in our Stroke Department.29 The most likely explanation is that we excluded patients with TIA, venous strokes, or cerebrovascular disorders without stroke (e.g., dissections and cerebral venous thrombosis without stroke). Moreover, our Stroke Department admits 15% of patients whose final diagnosis is not a cerebrovascular disorder,28,29⇓ and many of these patients are young. Causes of ischemic stroke encountered in our study are similar to those of young ischemic stroke patients reported in other Western studies,4,7-9,14,18-20⇓⇓⇓⇓⇓⇓⇓ with a slightly higher proportion of dissections. Conversely, several classic causes of ischemic strokes in the young, such as Sneddon syndrome, cardiac fibroelastoma, and cardiac myxoma, were not found in our patients.

Risk factors for stroke were frequent in our patients, especially in those who had no clear causes for ischemic stroke: Of 124 patients with no clear cause of stroke, only 14 were free of any risk factor. Risk factors did not differ from those encountered in similar studies in Western populations.4,8,9,20⇓⇓⇓ Because of the young age of the patients, the rates of associated coronary disease, peripheral artery disease, cardiac failure, and previous stroke or TIA were low.

Because of the low rate of new events in this age category, our study, like all other studies, was underpowered, as confirmed by the wide CI. Moreover, although 82.6% of patients were admitted during the first 24 hours and 95.1% within 3 days of stroke onset, we cannot exclude a small under-representation of patients with very early death. Antithrombotic treatments at the end of the follow-up period were not optimal in 14 patients. However, we did not interview general practitioners on this question, and for those who were not seen at a neurologic visit but only had a telephone call, it is possible that the reason to stop antithrombotic treatments was appropriate. However, 8.5% of hypertensive patients were not under antihypertensive therapy, and we did not explore, in those who were treated, whether the treatment actually lowered blood pressure toward proper levels. If advice to stop oral contraceptive therapy was followed in 44 of 45 patients, only 22 of 99 smokers gave up smoking. This is a major problem in terms of prevention, because the effectiveness of cigarette withdrawal reduces the risk of new vascular event in stroke patients.41

We found annual mortality rates of 4.5% (95% CI 2.1 to 6.9%) after the first year and 1.6% (95% CI 0.6 to 2.6%) during the subsequent 2 years, which are close to the 6.3 and 2.8% found during the same periods in the L’Aquila cohort.9 Other studies found lower18,19⇓ or higher6-8⇓⇓ mortality rates because they did not evaluate separately ischemic stroke, TIA, and intracerebral hemorrhages. Moreover, because of different inclusion criteria, possible recruitment bias, small number of events, and width of the CI, comparisons between studies remain difficult. Although the absolute survival rate of our cohort was high, the occurrence of an ischemic stroke in young adults cannot be regarded as a benign event, when a 8.7% mortality rate is found after a mean follow-up of 3 years. Ischemic stroke in young adults is serious event, associated with an increased risk of death in the subsequent few years.9 The mortality rate was higher during the first year than during the following 2 years, but the follow-up was too short in our study to determine whether, as described in the L’Aquila cohort,9 the average annual mortality rate remains stable over time after the first year.

The annual risk of stroke recurrence was 1.4% during the first year and then 1.05% during the following 2 years, which is similar to that found in most studies.4,9,10,14,18-20,22,24,25⇓⇓⇓⇓⇓⇓⇓⇓⇓ Only two cases of myocardial infarction occurred, and both were delayed after the first year: This low rate is similar to that found in other series.6,9,14,20⇓⇓⇓ Epileptic seizures occurred in 6.6% of patients, that is, more frequently than stroke recurrence. This risk was not clearly evaluated in other series.

Stroke subtypes probably influence the occurrence of outcome events, but our study, like previous ones, was not powered to evaluate outcome events separately in each category. Patients with stroke of atherothrombotic and cardioembolic origins were at higher vascular risk, as shown by a tendency toward higher rates of recurrent strokes, myocardial infarction, and death. Conversely, patients with cerebral ischemia included in the “other definite causes” diagnostic group were at low risk of recurrence, myocardial infarction, and death. This finding may be the consequence of the important weight in this group of patients with cervical artery dissections, who have low risks of death and stroke recurrence after the acute stage.36,42⇓

The functional outcome was reported as favorable in most studies conducted in young patients who survived an ischemic stroke.4,6,9,11,14-19,21-24⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓ Of 265 survivors, 249 (94.0% of survivors) were independent (mRS = 0 to 2), including 151 patients (57.0% of survivors) free of any residual symptom or sign. Other studies found less favorable functional outcomes.9,20⇓ Possible explanations for such discrepancies may be that in our study, 1) the follow-up was shorter; 2) patients were recruited years after those studies with less favorable functional outcomes, when imaging techniques were able to diagnose minor strokes and prevention made substantial improvements; and 3) the breakdown of causes was slightly different. The low rate of dependency in this age category leads to the opinion that invasive and potentially dangerous acute stroke therapies should be used only in the small group of patients who are unlikely to survive without dependency. Although the risk of cerebral bleeding increases with age and thrombolysis is useful in selected patients, patients who are the most likely to benefit from thrombolysis are probably not in this age category.

We found no significant differences in mortality, recurrence rate, dependency, and epilepsy 1) between patients examined at a neurologic visit and those followed by telephone interview, 2) between patients directly admitted in the stroke unit and those referred from another hospital, and 3) according to the antithrombotic regimen at discharge and at the last visit or contact. This lack of difference may be due to the lack of power of the statistical analysis, the number of subjects in each group being too small to allow the detection of a small difference. One could, for instance, have expected a worse outcome in patients referred from another hospital, that is, more likely to be the most severe at onset, but this was not the case. We were, however, less surprised to find no difference of outcome according to the antithrombotic regimen for the following two reasons: 1) the small number of patients, which did not allow sufficiently powered statistics; and 2) the heterogeneity of patients in each antithrombotic regimen in terms of stroke etiology and therefore stroke outcome. The influence of those factors on the rate of myocardial infarction was not evaluated because of the small number of patients who developed myocardial infarction.

Factors associated with a good outcome (mRS = 0 to 2) at the bivariate analysis were decreasing age, absence of alcoholism or diabetes mellitus, and presence of hypercholesterolemia or treatment with statin or fibrate or oral contraceptive therapy. However, the cause of stroke, stroke recurrence, and occurrence of myocardial infarction or epileptic seizure did not influence the outcome. We should, however, be careful in the interpretation of these results: First, the study was not designed to identify predictors of outcome; second, a potential important predictor of outcome such as the intensity of the clinical deficit at onset was not recorded because of the heterogeneity in the delay of admission (0 to 7 days); finally, we did not perform a multivariate analysis because of the impossibility to take into account an important factor such as the severity of the deficit and also because of the small number of subjects in each category. The lack of association between the cause of stroke, stroke recurrence, and occurrence of myocardial infarction or epileptic seizures with the outcome is probably due to the lack of power of the statistics. The association between hypercholesterolemia with a good outcome was also reported in older subjects,43 leading to the hypothesis that cholesterol might have a neuroprotective effect.43 Another hypothesis may be that patients with hypercholesterolemia are more likely to receive statins or fibrates, which both have a neuroprotective effect in experimental stroke.44 The association of statin or fibrate therapy with a good outcome in our study supports this hypothesis, but without a multivariate analysis, this result should be considered with caution: We cannot exclude that patients with hypercholesterolemia and statin or fibrate therapy have different stroke subtypes with a better spontaneous outcome.45 Accordingly, the association of oral contraceptive therapy with a good outcome may just reflect that oral contraceptive therapies are less often given in women with medical conditions predisposing to stroke, comprising those women most likely to have a bad outcome.

We found that 15.0% of patients who were regarded by health insurance as able to work were still unemployed. Of course, two thirds of them were already unemployed before stroke, and the study was conducted during the worst period of unemployment in France. Factors such as subtle cognitive or behavioral changes, depression, and fatigue may have influenced social recovery,21 but we did not make a systematic neuropsychological and behavioral assessment because of the observational design of the study. Besides, our study was not designed to answer the question of whether social and familial impairments were more frequent after an ischemic stroke but to evaluate the frequency of various endpoints, including social and familial impairments. Without controls, we cannot know whether the frequency of social and familial impairments is higher than that observed in the general French population in this age category.