Etiology of ischemic stroke among young adults of Serbia

Background/Aim. Etiology of ischemic stroke (IS) among young adults varies among countries. The aim of the study was to investigate the causes and risk factors of IS in the young adults of Serbia. Methods. A total of 865 patients with IS, aged 15 to 45 years, were treated throughout the period 1989–2005. Etiologic diagnostic tests were performed on the patient by the patient basis and according to their availability at the time of investigation. The most likely cause of stroke was categorized according to the TOAST (Trial of ORG 10172 in Acute Stroke Treatment) criteria. Results. There were 486 men and 379 women, with 19% of the patients ≤ 30 years old. Large artery arteriosclerosis and small artery disease were confirmed in 14% of the patients, and embolism and other determined causes in 20%. Undetermined causes made up 32% of the patients, mostly those (26%) with incomplete investigations. Smoking (37%), hypertension (35%) and hyperlipidemia (35%) were the most common risk factors. Rheumatic heart diseases and prosthetic valves were the most common causes of IS. Arterial dissections and coagulation inhibitors deficiency were detected in a small number of patients. Conclusion. Etiology of IS among Serbian young adults shares characteristics of those in both western and less developed countries.


Introduction
Ischemic stroke (IS) among young adults is a rare event that makes 5-10% of all stroke patients 1-3 .Also, the causes of IS in the young differ among various countries and from those in older population.Cardioembolism and other determined causes, mostly arterial dissections, were reported as the main causes of IS among young adults 1, [4][5][6][7][8][9][10][11][12] .Large artery atherosclerosis and small artery disease had rarely outnumbered other defined causes of stroke 2, [13][14][15][16] .However, the atherosclerosis is the main etiologic factor among the young of developing countries, as it used to be in developed countries more than two decades ago.
The aim of this study was to find out if risk factors and causes of IS among the young adults of Serbia are similar to those in other countries.

Methods
This prospective study included a total of 865 consecutive patients, 15-45 year old, with first ever transitory ischemic attack (TIA) or completed IS treated in the Department for Emergency Neurology from January 1989 to December 2005.In 1989, initially as a part of the Master thesis, a research project investigating etiology of stroke in young adults was started and continued until today.Clinical diagno-sis of IS was confirmed by brain computed X-ray tomography (CT) in 834 of the patients, brain magnetic resonance imaging (MRI) in 21 of the patients, brain nuclear scintigraphy in 7 and in 3 of the patients by autopsy.
Hypertension, diabetes mellitus, hyperlipidemia, cigarette smoking, previous TIA or stroke, heavy alcohol consumption, drug abuse, previous history of migraine, current oral contraceptive use and family history of stroke among first-and second-degree relatives were considered as risk factors for IS.Hypertension was regarded present when a patient had previously been advised to take antihypertensive drugs, or when blood pressure was > 140/90 mmHg in two different occasions at least 7 days after the stroke onset.Diagnosis of diabetes mellitus was established if a patient had already been treated or according to the criteria of the World Health Organization (WHO) 17 .Hyperlipidemia was considered risk factor when fasting blood total cholesterol was > 6.2, LDL -cholesterol (LDL-C) was > 4.2 mmol/l or HDL -cholesterol (HDL-C) < 1.1 or triglycerides were > 1.7 mmol/l 18 .Current cigarette smoking was verified when a patient had smoked > 5 cigarettes per day at least for one year, and as an ex-smoker a patient who had stopped smoking more than one year ago.Alcohol consumption was regarded chronic heavy drinking when a patient regularly took > 3 heavy drinks per day (> 36 g of alcohol/day) and as an acute alcohol intoxication when a patient had taken > 48 g of alcohol during preceding 24 hours 19 .Migraine was defined according to the criteria of International Classification of Headache Disorders 20 .Oral contraceptives were considered risk factor if they were used at any time in a 3-month period before the stroke 21 .Heart diseases were categorized as high-risk or low-risk cardio embolic diseases 22,23 .Low-risk cardioembolic diseases were considered potential cause of stroke only in the absence of other more possible etiological factors.
The diagnostic protocol for young IS patients included medical history, cardiac and neurological examinations, assessment of risk factors and appropriate laboratory tests.Routine laboratory tests, ophthalmoscopic exam, electrocardiogram (ECG) and chest X-ray radiography were done in nearly all the patients.Other diagnostic tests for identifying etiology of IS were performed according to the patient by the patient selection.For example, detailed coagulation or immunological laboratory tests or angiography were not performed in the presence of high risk cardioembolic diseases or an echocardiography was not mandatory for the patients with normal cardiac physical findings and high grade carotid stenosis concurrent with infarct localization.During this 17-year long period not all of the contemporary diagnostic tests were accessible and some of the routine diagnostic procedures were unavailable during the period of economic sanctions.Transthoracic echocardiography (TTE) was done in 451 of the patients (52%) and transesophageal (TEE) in only 38 of the patients.Vascular examination was done in 458 patients (53%) of which 289 ones were receiving cerebral or MR angiography (MRA) and 268 ones Doppler ultrasound exams.Blood lipids were analyzed in 412 of the patients (48%).Immunological testing (antinuclear antibodies, anti-DNA, antineutrophil cytoplasm antibodies, immune complexes, C3, C4, VDRL) was done in 375 of the patients (43%), antithrombin III (AT III) in 81 of the patients, protein C (PC) in 67, protein S (PS) in 32, lupus anticoagulant (LAC) in 107, anticardiolipin antibodies (ACA) in 98 and homcistein in 26 of the patients.Routine and immunological cerebrospinal fluid testing were done in 417 (48%) and 285 (32%) of the patients, respectively.
Initially, most likely cause of stroke was categorized as atherosclerotic disease, cardioembolism, nonatherosclerotic arteropathy, haematological disorder or undetermined cause according to previously described definitions 22 .Later, we revaluated the data and used the TOAST criteria to reassign the most likely cause of IS into one of 5 groups: 1) large artery atherosclerosis (LAA), 2) small artery disease (SAD), 3) embolism (EMB), 4) other determined causes (ODC), and 5) undetermined cause (UDC) with distinguishing unknown, uninvestigated and a multiple causes subgroup 23 .
For the statistical analysis, data were stored on SPSS version 13.0 software.Statistical evaluation was made by means of t-test and one-way ANOVA for numerical variables and χ 2 and ANOVA tests for proportions with confidence intervals.Multinominal logistic regression was performed for the evaluation of a possible effect of age, sex and risk factors on stroke subtypes.P < 0.05 was considered statistically significant.

Results
There were 379 women and 486 men with IS.Table 1 shows main demographic characteristics and risk factors in different subtypes of IS.Mean age of women (36.7±7.3) was significantly lower than that of men (37.6±7.0)(p = 0.045, 0.98, 95%CI: 0.02-1.94).In a group of patients with age up to 30 years women predominated (53%) and after that age men were more affected (58%).Carotid area was involved in 80% of patients, and 9% of patients had TIA.
Among hypertensive patients only 50% of them had been treated before the stroke and among those who did not take antihypertensives nearly 66% of the patients knew that they had hypertension for more than 5 years.Twenty-five out of 74 patients had insulin-dependent type of diabetes mellitus.Fifty-six patients had increased triglycerides, 18 high total cholesterol, 55 patients had both disorders and 16 patients had alone HDL-C decrease.Of the current smokers 93% smoked more than 10 cigarettes per day.
The angiography, carried out on the average 10 days after the stroke onset, showed pathological findings in 77% of all the investigated patients.In 75 patients of the LAA group who received angiography significant intracranial ste- nosis was present in 65% of them.Duplex ultrasound examination of cervical arteries discovered hemodynamically significant stenotic changes in only 15% of the examined patients.
High-risk embolic heart diseases were found in 59% of the patients with cerebral embolism (Table 2).Overall, there were 37 patients with atrial fibrillation and 12 patients with intracardial thrombus.Four out of 26 patients with rheumatic heart disease (RHD) were not aware of its presence until stroke onset.Among 14 patients with RHD and atrial fibrillation, eight had not received antithrombotic drugs before the stroke onset, 2 received antiplatelet agents and only 4 pa- tients had taken anticoagulants but with initial prothrombin time beyond therapeutic level in 3 of them.Also, the initial prothrombin time was in the therapeutic range in only 7 of 20 patients with mechanical prosthetic valves.One patient with mechanical valve had been treated with aspirin and another had not received any antithrombotic therapy.Mitral valve prolapse (MVP) was present in overall 27 of the patients and atrial septal aneurysm in 16.TTE showed pathological findings in 51% of all the examined patients.In 38 patients with previously normal TTE, subsequently performed TEE discovered abnormal findings in 27 of them, mostly atrial septal aneurysm or MVP.
In the ODC group the nonarteriosclerotic arteropathies were diagnosed in 56% of the patients (Table 3).In 8 pa-tients, cerebral angiographic findings with clinical signs, blood and CSF results suggested the diagnosis of possible isolated CNS angiitis.Among the viral cerebral vasculitides CSF immunological tests confirmed HIV or Epstein-Barr virus infection in two patients each and varicella zoster and herpes simplex type 1 in one patient.In one patient common viral antibodies were not detected.
Primary antiphospholipid syndrome (APL) was confirmed in 7 patients as a cause of IS and secondary one in three.Among 107 tested patients, LAC was present in six with two of them having systemic lupus erythematosus (SLE).Low-or medium-high positive ACA were detected in 13 of 98 tested patients.Six of these patients had primary APL syndrome, two had SLE and one Sjögren disease.In two of the patients with positive ACA the cause of IS was posttraumatic arterial dissection in one and the presence of intracardial thrombus in another one.One patient with positive ACA who had MVP with thickened valves was classified to embolic type of stroke and in one woman postpartum benign angiopathy was responsible for the stroke onset.Decreased AT III was present in 5 of 81 and decreased PC in 4 of 67 tested patients.Protein S was normal in all 32 tested patients.Hereditary thrombophillia was confirmed in one patient with AT III deficiency and in another one with both AT III and PC deficit.In one puerperal woman with low values of activated PC resistance who developed fatal MCA stroke and acute myocardial infarction genetic investigation was not done.
Overall, there were 43 women who had stroke during pregnancy or immediately postpartum.In 14 of them no other possible cause of stroke was detected.Only two patients out of 66 with migraine had possible migraine-induced stroke.Current oral contraceptive use was recorded in only 10 women, and in one woman who had taken them for 8 years the autopsy revealed signs of arterial intimal hyperplasia and complete carotid occlusion.In 3 of the patients the onset of stroke was related to marijuana or heroin abuse.
In the UDC group the cause of stroke was not discovered in 52 of the patients (6%) despite complete investigations.In only six patients more than one possible mechanism of stroke was detected.There were 223 patients (26%) without complete investigations.In 102 of these patients neuroimaging showed large-artery brain infarction and lacunes in 31.There were 72 patients who had 2 or more major risk factors for stroke.

Discussion
This is one of the largest published series of IS in young adults reported from a single medical center.This fact remains even if we exclude patients with TIA and all patients without complete investigation.Also, this study was one of the longest prospective studies that investigated causes of IS among young adults using predesigned protocol.However, for the most of the study period (1991-2000) our country was exposed to the civil war and economic sanctions that made up diagnostic facilities becoming poorer than in the beginning of the study period.For these reasons, a small proportion of our patients received vascular examination, TEE and update coagulation studies compared with other similar studies 2, 6,8,10,12,14,24 .Although our study was prospective, lack of all necessary diagnostic procedures during this period resulted in 26% of the incompletely investigated patients.In the Canadian study such patients made up 22% and in a French one only 9.5% of all the patients 10,24 .
The demographic characteristics and presence of hyperlipidemia and smoking were similar to those of other western countries and the frequencies of hypertension and diabetes mellitus were closer to those in Asian studies (Table 4) 1, 2, 9, 13, 14, 24, 25 .In addition, LAA and SAD groups made up one-quarter of all patients, which is higher than in most western studies and less than in Middle or Far East countries (Table 4) 1, 2, 5,6,10,12,14,15,24,26 .We believe that our proportion of LAA was underestimated because this diagnosis requires a positive finding of significant artery stenosis.Among our patients without full investigation there were even 73 of them who had at least two major risk factors for stroke, as well as CT findings of large artery infarction.It is highly possible that a substantial number of these patients would have had LAA if more vascular examinations had been done.Totally 14% of SAD patients is higher than the most of western studies reported 1,5,6,9,10,12,24 .This may well be the consequence of the inadequate preventive measures and poor control of risk factors during the past years.The fact that even half of our patients with hypertension had not received antihypertensive treatment before their IS and that of these as many as two thirds knew that they had hypertension for at least 5 years was the best illustration of an improper health education and poor medical care.
The proportion of cerebral embolism in our study was not different from the most recent studies except that RHD or prosthetic valves made up as much as a quarter of all these patients (Table 3) 2, 6, 9, 10, 14, 15 .During previous years, poor facilities for detection and follow-up of patients at risk resulted in an inadequate control of RHD and anticoagulation level.Only 4 out of 14 patients with RHD and AF received anticoagulants before the stroke onset.Lacking of preventive programs was also reported in India and Mexico where RHD and prosthetic valves made up as high as two thirds of all cardioembolic strokes in young adults 7,27 .In most other studies MVP, patient foramen ovale (PFO) and, atrial septal aneurysm (ASA) were common sources of cardiac embolism 1, 6, 9, 10, 12-14 .In our series these heart conditions were undiagnosed because of the low rate of the applied TTE and TEE.Out of only 38 patients who underwent TEE, in 27 were detected related pathological findings.This confirmed the importance of TEE in discovering potential causes of IS in the young as was documented in many other studies 1,2,10,12,28,29 .However, some of these abnormalities are common in the general population and detecting them does not mean that they had a causal role in the stroke.It is advisable to perform a full vascular examination and to exclude more probable causes of stroke in these patients.
The ODC was reported as the most common subtype of IS in the young (Table 4) 2, 5, 6, 9, 10, 14, 15, 24 .In our series this group of patients it was as common as embolism, which was also found by some other authors 1, 25 .In nearly all recent studies an arterial dissection was the most often detected separate cause of IS in young people, with a frequency of 14 to 20% of performed cerebral angiograms 2, 10, 12, 14, 24 .The proportion of arterial dissections among angiograms (3.5%) in our study was more in accordance with some earlier studies when they were not so readily recognized and made up 0.4 -10.1% of angiographic findings 4,6,8,13 .Nowadays, cross-sectional MRI in combination with MRA is the method of choice for initial diagnosis of craniocervical artery dissections 30 .A low proportion of arterial dissections in our study may imply inexperience of our clinicians and neuroradiolo-gists in detecting this entity, but even more it reflects the unavailability of MR neuroimaging during the past years.Moyamoya disease was detected in 5% of the performed angiograms and this was closer to the findings of Far East studies than to those of western countries 2, 10, 13, 14, 24 .It is not quite clear to us what were the reasons for this higher proportion of Moyamoya disease among our patients.However, Adams et al. 6 also reported higher appearance of this cause in the young with IS.
We confirmed a haematological disorder as a cause of IS in 4.3% of all our patients.Others reported this proportion being 4.1 -5.7% 6,10,24 .ACA were positive in 13% of those tested and in 7 patients we regarded APL syndrome as a cause of IS.In the selected groups of young patients with IS, positive ACA were detected in up to 23% of patients 2, 7, 14, 31, 32 .However, some studies with a low frequency of ACA even questioned the relevance of their role in unselected young stroke patients 1,6,33 .This notion prevails even more for a hereditary deficiency of coagulation inhibitors 32 .A systemic review of the literature showed that the prevalence of inherited deficiencies of PC, PS and AT III was low in unselected IS patients.Even in young stroke patients deficiency of coagulation inhibitors was rare and mostly of the acquired type 1, 2, 6, 34 .Only in a few studies of young stroke patients inherited deficiencies of coagulation inhibitors were reported with frequencies of 4 -7.5% 12,14,35 .Out of a small number of the tested patients, we documented hereditary deficit of AT III in two patients, of whom one also had PC deficit, and very low APCR ratio in one puerperal woman whose death precluded genetic analysis for Factor V Leiden mutation.

Conclusion
Our study on young ischemic stroke patients is one of the largest reported, showing that our patients did not differ in their demographic characteristics from western stroke patients.Similar to the patients of Middle or Far East, our young people had higher proportion of hypertension and lacunar strokes.Embolism and other defined causes were the most common etiologic subtypes of IS, but still with very high proportion of rheumatic valvular diseases.Arterial dissections and coagulation disorders might have been underestimated.Migraine, oral contraceptive use and drug abuse were rarely associated with IS in our study.It is our assumption that the poor diagnostic facilities, low health education and inadequate preventive measures during past years are responsible for some of the differences.During the last few years, there have been some indications that the etiology of stroke in the young of Serbia is more similar to the reports of developed countries.

Table 4 Demographic characteristics, risk factors and causes of ischemic stroke (IS) in the young less than 45 years old in different studies
Jovanović DR, et al.Vojnosanit Pregl 2008; 65(11): 803-809.