ABSTRACT
The possible effect of the
apolipoprotein E polymorphism on the development of ischemic cerebrovascular
disease has not been sufficiently investigated, and controversial results were
obtained from the few existing studies. In this study, our goal was to
determine the possible role of the apolipoprotein E polymorphism in stroke
patients. Genotyping of apolipoprotein E carried out on 79 patients (26
thrombotic, 20 embolic, 26 lacunar, and 7 miscellaneous), and in 126 age and
sex matched controls who were free of cerebrovascular disease. In addition,
serum apolipoprotein E, A I, C III and B and lipoprotein (a) levels were
determined. The prevalence of well-known vascular risk factors was
significantly higher in the patients. The e2
allele was found to be significantly lower in patients (3.16 %) than in
controls (8.34%) (c2
= 4.37, p<0.05). Patients
with large-vessel stroke or lacunar stroke had higher triglyceride and lower
HDL levels. In all stroke subtypes, apolipoprotein A I levels were lower than
those in controls, and the ratio of apolipoprotein B to A I was higher. A
stepwise logistic regression showed that; the presence of vascular stroke was
related to e4
allele, diabetes, high systolic blood pressure, and high apolipoprotein E
serum levels, but inversely related to e2
allele and apolipoprotein A I levels. Epsilon 2 may protect individuals
against stroke even if the e2
carrying patients have higher apo E levels, and e4
may be a genetic risk factor together with other well-known vascular risk
factors. Large population-based studies are needed to clarify the exact
relationship between stroke and lipid metabolism.
INTRODUCTION
Stroke is the third major cause of death
and long term disability in industrialized countries. Intracerebral and
subarachnoid haemorrhages account for only 15 % of all strokes whereas the
other 85 % are caused by cerebral ischemia and can be distinguished according
to the cause, clinical syndrome or the arterial distribution. Stroke due to
large artery atherosclerosis causes infarcts, which are larger than 1.5 cm
along with the territory of the major intracerebral arteries. Cardiac embolism
also gives a similar clinical picture to large-artery atherosclerosis. In
small artery occlusion, which is caused by lipohyalinosis of small perforating
vessels, small brain stem or subcortical lesions of less than 1.5 cm are
detected. The other rare causes of strokes are vasculitis, hemotologic
disorders, migraine or oral contraceptives (van Gijn and van der
Worp,1995).Cerebrovascular diseases are not only induced by classic vascular
risk factors including hypertension, diabetes mellitus, cigarette smoking, but
also by genetic factors (Pullicino et al.,1997).
Previously, the role of apolipoprotein E (apo
E) polymorphism in atherosclerotic events has been shown (Utermann,1987). Apo
E is a protein which acts as a ligand for low density lipoprotein (LDL)
receptors and affects the hepatic binding, uptake and catabolism of different
lipoproteins. It also has a function in the repair response to tissue injury.
Increased levels of apo E concentration have been shown at sites of peripheral
nerve injury and regeneration (Mahley,1988). It has been found that the gene
for apo E is located on chromosome 19. Six major isoforms of the apo E gene
exist, each contains a pair among three major alleles e2,
e3,
e4
which encode the protein isoforms E2, E3, E4 (Siest et al.,1995; Mahley,1988).
Clinical and postmortem studies have shown that the e4
allele is associated with pathologies such as coronary artery diseases or
Alzheimer's disease (Kosunen et al.,1995; Strittmatter and Roses, 1995, Wang
et al.,1995; Corder et al.,1993;Saunders et al.,1993). In patients with
ischemic cerebrovascular diseases (ICVD), only a limited number of studies
clarifying the role of apolipoproteins and the role of genetic polymorphisms
in lipoprotein metabolism have been performed. Studies searching for the
effect of the apo E polymorphism are also rare and have produced contradictory
results (Ferruci et al.,1997; Kessler et al.,1997; Schmidt et al.,1997;
Kuusisto et al.,1995; Couderc et al.,1993; Pedro-Botet et al.,1992).
In this study, we aimed to find out the
relationship, if one exists,between ICVD, lipids, apolipoprotein levels and
apo E polymorphism. Moreover we searched for a possible relationship between
apo E polymorphism and subtypes of stroke.
MATERIALS AND METHODS
Patient selection:
The study group consisted of patients
admitted to the Neurology Department of Akdeniz University Hospital and to the
other two major hospitals in the city, who had suffered from an ischemic event
or patients admitted to the Neurology Out-patient Clinics for follow-up, who
had previously been diagnosed as stroke patients within the last 15 months.
Each patient underwent a complete physical and neurological examination by a
neurologist. The control subjects were free of cerebrovascular diseases and
were matched with the patients for sex and age on the basis of age classes (
40-49, 50-59, 60-69, 70-80 ).
Nearly all the patients underwent
computerized tomography (CT) or magnetic resonance imaging (MRI) and
electrocardiography (ECG) analysis. On the basis of clinical symptoms and
findings of diagnostic tools, the patients were assigned into one of the
following categories: (1)- Large vessel disease-strokes due to pathologies of
extracranial arteries either thrombotic or embolic, (2)- Lacunar stroke due to
small deep infarctions in the territory of small perforating arteries of the
brain where localization of these infarcts was confirmed by CT or MRI and the
matched clinical symptoms, (3) Stroke due to miscellaneous causes which
includes cardiac emboli or unknown pathomechanisms. The consciousness level
(based on the Glasgow Coma Scale), functional status (based on the Barthel
Index) and the level of disability (based on the Rankin Scale) were also
determined in patients.
Potential confounders:
Vascular risk factors and associated
vascular diseases, based on the individual's personal history, results of a
physical examination and appropriate laboratory findings were recorded for
patients and controls. Those included; hypertension, diabetes mellitus (DM),
tobacco and alcohol consumption, use of oral contraceptives, body mass index
(BMI), history of migrane, ischemic heart diseases, arrhythmias, family
histories of hypertension, diabetes mellitus and stroke. Arterial hypertension
was considered to be present if an individual had a history of hypertension or
was using antihypertensive agents or if the systolic blood pressure (SBP)
exceeded 140 mmHg or the diastolic blood pressure (DBP) exceeded 90 mmHg. DM
was considered to be present if fasting glucose levels were exceeded 7.78 mmol/L
or if the individual was using antidiabetic agents.
We only included patients with neurologic
symptoms resulting from focal cerebral ischemia and excluded patients with
intracerebral or subarachnoid haemorrhage. Patients with stroke resulting from
vasculitis, migraine, oral contraceptive use or from trauma were also
excluded.
Apo E genotyping:
DNA was prepared from whole blood (Miller
et al.,1988). For genotyping of common apoE isoforms, amplification of apoE
sequences was carried out (Hixon and Vernier,1990). Electrophoresis of the
samples was performed on 10% polyacrylamid gel, after digesting the amplified
products with the HhaI restriction enzyme. The detection of restriction
fragments was performed by staining with ethidium bromide under UV light.
Determination of parameters of lipid
metabolism:
Serum apo E concentrations were determined
immunoturbidimetrically by using a kit supplied by Daicchii Pure Chemicals
Co., Ltd., Tokyo, Japan. Total cholesterol, HDL-cholesterol (after
precipitation procedure with phosphotungustic acid and Mg+2 ions)
and triglyceride concentrations were determined enzymatically by using CHOD/PAP
and GPO/PAP methods respectively and the calculation of LDL-cholesterol was
performed by using Friedewald formula. Apo AI, apo B and lipoprotein (a) [Lp(a)]
levels were determined in a Behring Nephelometer. Quantification of total apo
E and total apo C III particles and apo E present in the particles without apo
B (apo E LP non B) and apo C III present in particles without apo B (apo C III
LP non B) was carried out by an electroimmunodiffusion technique in an agarose
gel supplied by Sebia (Issy-les-Moulineaux, France).
Statistical analysis:
Standard statistical procedures from the
BMDP statistical software were used. For major risk factors and potential
confounders, differences between case and control groups were tested by using
Student's t test, ANOVA, Dunnett's test and c2
analysis. As the distribution of
triglyceride and Lp(a) levels were skewed, the log transformed values were
used. To estimate the effect of the apo E polymorphism on the risk of stroke,
whilst simultaneously adjusting for possible confounders, unconditional
multiple logistic regression was used with maximum likelihood estimation of
the regression coefficients and their standard errors. All the potential
confounding factors were systematically tested in the regression models: sex,
hypertension, diabetes, cigarette and alcohol consumption, family history of
stroke and hypertension (as categorical variables), and age, SBP and DBP, BMI,
cholesterol, triglyceride, apolipoprotein and lipoprotein levels (as
contunious variables). The fitted model included DM, SBP and serum apo E and
apo AI levels as confounders. Adjusted odds ratios for stroke were calculated
for e2
allele [(e2/
e2)
+ (e3/)
+ (e2/
e4)]
and for e4
allele [(e3/
e4)
+ (e4/
e4)]
by taking the e3
[(e3/
e3)]
allele as a reference. Significance levels were set at 0.05 in all cases.
RESULTS
We studied 79 patients ( mean-/+SD
age, 62.9-/+8.9
years, range 40-80 years, 50.6% male) and 126 controls (mean-/+SD
age, 58.6+/-8.81 years, range 40-77 years, 47.6% male). The prevalence of
genotypes which are shown in Table 1a. and the allele frequencies are shown in
Table 1b.
Table
1a: Apo E genotypes in controls and patients
|
|
Controls
|
Patients
|
|
|
n
|
%
|
n
|
%
|
| e2/e2
|
1
|
0.80
|
0
|
0.00
|
| e2/e3
|
17
|
13.50
|
4
|
5.06
|
| e3/e3
|
87
|
69.05
|
60
|
75.95
|
| e3/e4
|
19
|
15.07
|
13
|
16.45
|
| e4/e4
|
0
|
0.00
|
1
|
1.27
|
| e2/e4
|
2
|
1.58
|
1
|
1.27
|
Table
1b: Apo E allelic frequency (Miscellaneous group contains patients with stroke
due to cardiac embolism and with unknown cause)
|
|
Controls
|
Patients
|
Patients
|
|
|
|
(Total)
|
Large vessel
stroke
|
Lacunar
|
Miscellaneous
|
|
|
|
|
Thrombotic
|
Embolic
|
|
|
|
e2
(%)
|
8.3
|
3.2*
|
1.9
|
2.5
|
5.8
|
0.0
|
|
e3
(%)
|
83.3
|
86.7
|
94.2
|
90.0
|
75.0
|
92.9
|
|
e4
(%)
|
8.4
|
10.1
|
3.9
|
7.5
|
19.2*
|
7.1
|
*p<0,05
Difference between controls and patients (Dunnettis test)
No statistically significant
difference was found between patients and controls for e3
and e4
allele frequencies. But e2
was found to be much more prominent in controls than in patients (c2
=4.37, p<0.05). The general
characteristics of the patients and controls are seen in Table 2.
Table
2: General characteristics of controls and patients.
|
|
Controls
(n=126)
|
Patients
(n=79)
|
| Age
(years)
|
58.6-/+8.8<
/p>
|
62.9-/+8.9*
*
|
|
|
(40-77)
|
(40-80)
|
| Male
ratio (%)
|
47.6
|
50.6
|
| SBP
(mmHg)
|
123.2-/+17.8
|
144.6-/+17.1
***
|
| DBP
(mmHg)
|
81.1-/+10.1
|
91.2-/+11.9
***
|
| Hypertension
(%)
|
27.8
|
79.0***
|
| DM
(%)
|
0.8
|
28.4***
|
| Cigarette
Smoking (%)
|
|
|
| Past
|
15.1
|
25.9***
|
|
Current
|
29.4
|
50.6***
|
| Alcohol
Consumption (%)
|
|
|
| Past
|
2.4
|
1.2
|
|
Current
|
7.9
|
23.5
|
| BMI
(kg/m2)
|
30.62-/+3.53
|
30.52-/+4.35
|
| Family
History (%)
|
|
|
| Hypertension
|
32.5
|
30.9
|
|
DM
|
15.1
|
22.2
|
|
Stroke
|
23.8
|
49.4
|
**p<0,05,
***p<0,001: Difference between controls and patients (Student t
test or c2
test)
No statistically significant sex
differences were found between patients and controls but the mean age of
patients was higher than those of controls (p<0.05). SBP and DBP were found
to be higher in patients than in controls (p<0.001) especially in older
individuals. Of the patients, 79% were found to be hypertensive and 28.4% to
be diabetic. No significant difference was found between patients and controls
for the family history of hypertension and DM, but on the other hand 49.5% of
the patients had a positive family history for stroke, compared to 23.8% of
the controls (c2=13.40,
p<0.05). Current cigarette and alcohol consumption was found to be much
more prominent in patients than in controls (c2
=20.72, p<0.001, c2
=10.02, p<0.05 respectively).
Table
3: Plasma lipid profile in controls and patients according to stroke subtypes.
(Miscellaneous group contains patients with stroke due to cardiac embolism and
with unknown cause)
|
|
Controls
|
Patients
|
Patients
|
|
|
|
(Total)
|
Large
vessel stroke
|
Lacunar
|
Miscellaneous
|
|
|
|
|
Thrombotic<
/b>
|
Embolic
|
|
|
| Apo
E
(mg/L)
|
43.4-/+11.5
|
46.5-/+12.1
|
48.2-/+10.9
|
45.3-/+12.7
|
47.0-/+13.6
|
41.3-/+8.3<
/p>
|
| Cholesterol
(mmol/L)
|
5.49-/+1.14
|
5.13-/+1.14
|
5.13-/+1.14
|
5.51-/+1.13
|
4.94-/+1.10
|
5.33-/+1.17
|
| Triglyceride1
(mmol/L)
|
1.78-/+0.83
|
1.97-/+
1.08
|
1.93-/+0.88
|
2.15-/+1.14
|
2.00-/+1.31
|
1.52-/+0.51
|
| HDL
(mmol/L)
|
0.80-/+0.28
|
0.64-/+0.22
***
|
0.65-/+0.24
**
|
0.68-/+0.22
|
0.57-/+0.17
**
|
0.76-/+0.26
|
| LDL
(mmol/L)
|
3.87-/+1.02
|
3.63-/+0.96
|
3.52-/+0.95
|
3.84-/+0.93
|
3.50-/+0.93
|
3.87-/+1.13
|
| Apo
AI
(g/L)
|
1.45-/+0.25
|
1.21-/+0.29
***
|
1.19-/+0.38
**
|
1.26-/+0.21
*
|
1.17-/+0.28
**
|
1.30-/+0.24
|
| Apo
B
(g/L)
|
1.09-/+0.25
|
1.10-/+0.25
|
1.12-/+0.23
|
1.15-/+0.26
|
1.06-/+0.28
|
1.04-/+0.30
|
| Apo
B/Apo AI
|
0.76-/+0.19
|
0.95-/+0.29
***
|
0.97-/+0.30
**
|
0.93-/+0.27
*
|
0.93-/+0.22
**
|
0.83-/+0.32
|
| Lp
(a)1
(mg/L)
|
0.18-/+0.14
|
0.36-/+0.22
*
|
0.43-/+0.26
*
|
0.12-/+0.10
|
0.22-/+0.19
|
0.21-/+0.19
|
| Lp
C III
(mg/L)
|
27.30-/+10.1
1
|
26.80-/+9.29
|
26.66-/+9.77
|
27.87-/+9.11
|
26.63-/+9.54
|
24.85-/+6.34
|
| Lp
C III NB (mg/L)
|
20.42-/+8.13
|
20.15-/+6.78
|
20.04-/+6.79
|
20.30-/+7.13
|
19.70-/+6.08
|
20.51-/+5.22
|
| Lp
C III B (mg/L)
|
6.87-/+3.90
|
6.65-/+4.40
|
6.61-/+4.12
|
7.57-/+5.14
|
6.93-/+4.70
|
4.34-/+3.78
|
| Lp
E
(mg/L)
|
45.98-/+16.6
0
|
55.16-/+25.0
4*
|
57.90-/+26.9
6*
|
47.58-/+16.8
3
|
55.52-/+29.4
5*
|
55.67-/+28.8
4
|
| Lp
E NB (mg/L)
|
38.13-/+16.6
5
|
45.52-/+23.9
8*
|
47.46-/+26.2
5
|
40.64-/+16.2
4
|
47.51-/+27.6
3*
|
45.04-/+25.9
2*
|
| Lp
E B (mg/L)
|
7.85-/+4.65
|
8.73-/+7.35
|
9.05-/+7.73
|
7.23-/+6.10
|
8.01-/+5.54
|
10,62-/+7,94
*
|
1test
on log-transformed variables
*p<0.05, **p<0.01,
***p>0.001 : difference between controls and patients (Dunnettis test).
Table 3 shows plasma lipid profile
according to stroke subtypes. Although apo E serum concentration seems to be
higher in patients, no statistically significant difference was found between
patients and controls, but e2
allele carrying patients had higher apo E levels than the e2
carrying controls (p<0.001) while lower apo E serum concentration was found
in patients carrying e4
allele than in controls (p<0.05). Total cholesterol, triglyceride and LDL-cholesterol
levels were found to be similar in patients and controls, but they were shown
to be higher in e4
carrying controls than in patients (p<0.05). HDL-cholesterol levels were
found to be higher in controls than in patients especially in female
individuals (p<0.001). In e3
and e4
carrying controls, apo A I levels were higher than in patients (p<0.001),
but we were unable to find any significant differences between patients and
controls for apo B and Lp C III particles. On the other hand, total Lp E
particles and Lp E NB particles were significantly higher in patients,
especially in e2
and e3
carrying individuals (p<0.05).
We divided individuals according to sex
,age groups and either e2
allele carriers (e2+)
or e2
allele non-carriers (e2-)
in order to find out the relationship between age, sex and the occurance of
stroke. We could not find any relationship between age, sex and stroke
occurance in the e2+
group, but we found that, in the e2-
group we had many more male patients in the 70-80 years age group than
controls (c2
=6.78, p<0.05). Although the
number of female patients in this group was high, there was no statistical
significance of this.
No statistically significant difference in
cholesterol and triglyceride levels was found between stroke subtypes and
controls. In thrombotic and lacunar groups, HDL-cholesterol levels were found
to be lower than those in controls (p<0.05). Apo A I levels were also low
in those groups (p<0.05) and the ratio of Apo B to apo A I was higher
(p<0.05). Levels of Lp E and Lp E NB particles were only found to be higher
in thrombotic and lacunar infarct patients (p<0.05).
The parameters showing the outcome of
stroke, Glasgow Coma Scale, Barthel Index or Rankin Scale, didn't show any
significant difference between male or female patients and there was also no
difference according to the allele distribution and stroke subtype. But
patients carrying the epsilon 2 allele seemed to show a tendency to have a
better outcome than others.
Multiple logistic regression analysis
indicated that apo E e4
genotype (e3
versus others) (p<0.05), DM (p<0.05), apo E level (p<0.05) and SBP
(p<0.05) were significant and independent predictors of stroke, but on the
other hand apo E e2
genotype and apo A I levels were found to be inversely associated with the
presence of stroke (Table 4).
Table
4:Results of the multiple logistic regression analysis.
| Variable
|
b
|
SE
|
d.f.
|
p
|
Odds
Ratio
|
95%
C.I.
|
|
|
|
|
|
|
|
|
| Apo
E Genotype*
|
|
|
2
|
0,0015
|
|
|
|
e2
|
-2.02
|
0.52
|
1
|
|
0.13
|
0.04-0.37
|
|
e4
|
0.28
|
0.38
|
1
|
|
1.32
|
0.62-2.82
|
| DM**
|
1.78
|
0.61
|
1
|
0.0112
|
5.93
|
1.75-20.08
|
| Apo
E (mg/L)
|
0.05
|
0.01
|
1
|
0.0060
|
1.05
|
1.03-1.07
|
| Apo
AI (g/L)
|
-2.89
|
0.91
|
1
|
0.0078
|
0.05
|
0.01-0.34
|
| SBP
(mmHg)
|
0.088
|
0.015
|
1
|
0.0001
|
1.09
|
1.06-1.12
|
*e3
is the reference
**Absence of DM is
reference
DISCUSSION
This study was designed as a case-control
study. Due to the probable differential effect of apo E polymorphism by age,
to be able to give false positive results, we had to choose the controls
carefully according to the age groups designed by the cases. Therefore even
though the mean ages of controls and patients differ significantly, this is
attributed to the above reason and we believe that it is the best
epidemiologic strategy in order to avoid confusion due to interactions between
the effects of this polymorphism and age.
There are a limited number of studies
performed on members of the Turkish population about apo E polymorphism. In a
wide range population based study, Mahley et al. reported the allele
frequencies of the Turkish population as; 6.1% for e2,
86% for e3
and 7.9% for e4
(Mahley et al.,1995). In the present study, we found 6.3% for e2,
84.6% for e3
and 9.1% for e4.Thes
e results are comperable with the study of Mahley et al.
Since stroke is one of the major causes of
death and long term disability, current approaches are based on preventing the
occurrence of a first or recurrent stroke by reducing the risk factors,
especially high blood pressure, hyperlipidemia, susceptibility to thrombus
formation or platelet aggregation (Iso et al.,1989). Such measures may be
partly responsible for the prevention of stroke. Besides these factors,
environmental or genetic factors are also important (Graffagnino et al.,1994;
Kiely et al.,1993). Recent data showed that individuals with genetic
predisposition are thought to be much more prone to diseases such as coronary
artery diseases or Alzheimer's disease. But data about the genetic
susceptibility of stroke are insufficient and this field is of interest for
researchers.
To date, compared to other pathologies,a
limited number of studies have been performed on ICVD patients in different
populations. Pedro-Botet et al. reported that the e4
allele was more prevalent in ICVD patients and could be a predisposing genetic
marker (Pedro-Botet et al.,1992). In our study, we were unable to find any
significant difference between patients and controls for e4
allele distribution, but e4
prevalence was higher in patients with lacunar infarcts. From this point of
view, our study would be compatible with that of Pedro-Botet et al., but there
was no age discrimination being possibly related to the importance of e4
allele in that of Pedro-Botet et al.. Previously Kuusisto et al. reported that
carrying the e4
allele cannot be regarded as an important risk factor for coronary heart
disease or stroke in elderly individuals (Kuusisto et al.,1995). From this
point of view; there is some conflict between our study and these other two.
We found that the e4
allele is more common in ICVD patients with lacunar etiology and these
findings support the previously reported data (Kessler et al.,1997). It was
reported by Couderc et al. that, the E3/E2 phenotype was much more frequent in
patients and concluded that the epsilon 2 gene may be a risk factor for
cerebrovascular morbidity together with the associated risk factors of
diabetes, hypertension and obesity (Couderc et al.,1993). In fact, in our
study population the prevalence of the E3/E2 genotype was much more frequent
in controls than in patients and also epsilon 2 allele frequency was found to
be more common in controls. It is definitely the case that carrying an epsilon
2 allele is associated with a lower risk of ischemic stroke (Ferruci et
al.,1997). These findings are consistent in young and middle-aged individuals
but not confirmed in older populations (Kuusisto et al.,1995). Ferruci et al.
reported that the epsilon 2 allele was associated with a lower risk of
ischemic stroke but that this limited protective effect was confined to the
age range 70 to 79 years (Ferruci et al.,1997). Our results are consistent
with this study. We also found that the epsilon 2 allele was much more
prominent in controls than in patients but we could not detect any age range
in which epsilon 2 was much more protective. We observed that individuals
within the age range of 70 to 80 years whoever carrying the epsilon 4
allele,were much more prone to stroke than others. This conflict may be due to
the mean age of the selected subjects. In the study by Ferruci et al., the
subjects were more than 70 years old and the authors concluded that results
should be verified with a wider age range. In our study population, the age
range was from 40 to 80 years and it can be confirmed that epsilon 2 allele is
associated with a lower incidence of stroke except in the age group between 70
and 80 years. Schmidt et al. determined the apo E polymorphism in patients
with microangiopathy -related cerebral damage which manifests itself with
white matter abnormalities and lacunar lesions in MRI in elderly individuals (Schimidt
et al.,1997). It was concluded that the E3/E2 genotype was significantly and
independently associated with such abnormalities. In fact in our study of
patients with lacunar infarcts, epsilon 4 was the frequently associated allele
when compared with controls.It is also important to note that epsilon 2 allele
carrying patients had higher apo E and Lp E non B levels.
There are some existing results which have
been published showing that the concentration of apo E is an important factor
for the affinity of lipoproteins to specific apo E receptors and showing that
higher apo E levels could accelerate cholesterol entrance into the cell a
