HEALTH SCIENCE JOURNAL

Volume 6, Issue 3 (July September 2012)

_ORIGINAL ARTICLE_

Risk factors associated with Epilepsy: A case-control study

Athanassios Vozikis1, John E. Goulionis2, Dimitrios Nikolakis3
1.MSc, PhD, Lecturer, University of Piraeus, Economics Dept.
2.PhD, Researcher, University of Piraeus, Department of Statistics and Insurance Science

3. MSc, Researcher, University of Piraeus, Department of Statistics and Insurance Science

ABSTRACT
Background: Epilepsy is a chronic neurological condition and the worlds common serious brain disorder.
The aim of this case-control study is to identify the significance of certain risk factors for epilepsy. The
risk factors examined are age, head trauma, central nervous system infections e.g. meningitis, abnormal
perinatal history, family history, socioeconomic status and parental consanguinity.
Methods: We designed a case-control study for 700 patients (512 males and 188 females) attending the
outpatient neurology clinic of a University General Hospital in Athens-Greece during a two years (20072009) period. In this study the patients were examined and their personal and medical data were taken.
For each patient, the phenomenon of epilepsy was studied in relation to different risk factors. The odd
ratios have been computed to see the significant factors associated with epilepsy.
Results: Positive family history for epilepsy, head trauma, sex, febrile and abnormal perinatal history and
low education, were shown to have a statistically significant association with epilepsy.
Key words: Statistics in Medicine; Epilepsy; Risk factors for Epilepsy; Odds ratio.
CORRESPONDING AUTHOR

Athanassios Vozikis
Lecturer, University of Piraeus, Economics Dept.
80, Karaoli & Dimitriou street
18534 Piraeus, Greece
Tel: 210 4142280
Email: avozik@unipi.gr

INTRODUCTION

pilepsy

chronic

worldwide have epilepsy, with almost

neurological disorder characterized

90% of these people being in developing

seizures.1

countries.2 Epilepsy is more likely to

disabling

occur in young children or people over

condition, which affects approximately

the age of 65 years; however it can occur

3% of the world population during their

at

lifetime.

controlled, but cannot be cured with

by

is

recurrent

Epilepsy

is

common

unprovoked
a

About

common

50

million

people

any

time.

Epilepsy

is

usually
Page | 509

E-ISSN: 1791-809X

Health Science Journal All rights reserved

www.hsj.gr

Quarterly scientific, online publication by Department of Nursing A,

Technological Educational Institute of Athens

medication, although surgery may be

spheres, but polls suggest it is generally

considered

decreasing with time, at least in the

in

difficult

cases.

Socioeconomic factors may influence the

developed world.13

risk of the disease in many ways.3 Some

Many

studies have found increased risks of

activities

cerebrovascular disease, brain tumors,

epilepsy. The most commonly prohibited

and neurodegenerative disease among

activities involve operation of vehicles or

people who have occupational exposure

machinery or other activities in which

to chemicals4, but few studies have

continuous

reported on the association between

However, there are usually exceptions

specific occupations and incidence of

for those who can prove that they have

epilepsy.5

Socioeconomic

jurisdictions
to

forbid

persons

certain

suffering

vigilance

is

from

required.

status

and

stabilized their condition. There is an

carry

outgoing debate in bioethics over who

risk

of

should bear the burden of ensuring that

hospitalization for epilepsy.6 Low income

an epilepsy patient does not drive a car

and low education are associated with an

or fly an airplane.2

increased risk among both men and

Several studies have found that fatalities

women.7 Risk is increased for men and

caused by seizures that occurred while

women in certain occupational groups.8

driving were relatively rare.13-15

occupation

sometimes

significantly

increased

Over 30% of people with epilepsy do not

seizure

control

available

even

with

the

medications.9-11

best

Materials and Methods

Epilepsy

Research Methodology

surgery is an option for patients whose

A total of 700 patients (512 males and

seizures remain resistant to treatment

188 females) were examined. The risk

with anticonvulsant medications who

factors

also

localization-

considered the sex (sex = 0, if patient is

related epilepsy; a focal abnormality that

female and sex = 1, if patient is male).

can be located and therefore removed.

This study has a number of strengths.

The goal for these procedures is total

For

have

symptomatic

control of epileptic seizures.

12

of

epilepsy

example,

our

investigated.

study

We

population

In most

included a well-defined open cohort of

cultures, persons with epilepsy have

the entire population of Greece. The

been

even

present study also has some limitations.

to

Although the national database includes

nowadays, in both the public and private

data on the entire population, it only

stigmatized,

imprisoned.

shunned

Stigma

or

continues

Risk factors associated with Epilepsy: A case-control study

Page | 510

HEALTH SCIENCE JOURNAL

Volume 6, Issue 3 (July September 2012)

incorporates information about hospital

admissions.

patient

products of the two pairs of diagonal

unavailable.

elements in the above 2 2 table, and for

Another limitation is that we were

this reason, is sometimes referred to

unable to test the validity of the epilepsy

as

diagnoses. However, it seems likely that

association between two factors can be

the diagnoses are valid. Moreover, to

tested via odds ratio. The null hypothesis

help

to be tested in this case is:

treatment

Data
are

counter

on
rather

any

out-

This estimate is the ratio of the

potential

validity

the

cross-product

problems, we only used main diagnoses

of epilepsy recorded in the hospital
registers, i.e., instances in which the
main

cause

of

hospitalization

was

epilepsy. This increased the possibility

that the diagnoses of epilepsy were valid.

or equivalently:

H 0 : ln 0 .
Note that testing above hypothesis is
same as testing:
H 0 : 1 0

in simple logistic model:

p
ln
0 1 X .
1 p

leads to a success in terms of the odds of

that event. When two sets of binary data
are to be compared, a relative measure of

The null hypothesis:

H 0 : 1

the odds of success in one set relative to

that in the other is referred to as the

or:

H 0 : ln 0

odds ratio.
Suppose a 2 2 table is given as
Number

Number

of

of

successes

failures

Dataset 1

Dataset 2

Then the odds ratio is given as:

p 1 p1 ad
1

.
p 2 1 p 2 bc

The

H 0 : 1

It is sometimes helpful to describe the

chance that a binary response variable

ratio.

may be tested using the test statistic:

ln

S .E.ln

which has an approximate standard

normal

distribution.

100 1 %

An

confidence

approximate
interval

for

ln is constructed as:
ln Z /2 S.E.ln .

(2.1)

For example, a 95% confidence interval

for ln is given by:
Page | 511

E-ISSN: 1791-809X

Health Science Journal All rights reserved

www.hsj.gr

Quarterly scientific, online publication by Department of Nursing A,

Technological Educational Institute of Athens

ln 1,96 S.E.ln .
The

confidence

interval

significant at 1% level of significance. A

given

by

equation (2.1) on inversion will give us

Z /2 S . E .ln

odds ratio is (1.624, 3.180) and hence

the 95% confidence interval for the true

the confidence interval for as:

Z /2 S . E .ln

95% confidence interval for the true log

odds

ratio

is

(5.072,

24.047).

The

confidence interval shows a significant

If the interval contains unity, it indicates

relationship between injury and epilepsy

independence; otherwise an association

(Table 2).

is indicated.

In the case of the relationship between

epilepsy and meningitis, the estimated

Results

odds of having epilepsy are 4.057 times

Odds Ratio Analyses

more for those who have meningitis as

In the first case (relationship between

compared to those who have not. The

sex and epilepsy) the estimated odds of

estimated log odds ratio is 1.400 and its

having epilepsy are 1.084 times more for

asymptotic

males as compared to females. Hence the

which is highly significant at 1% level of

males are more likely to have epilepsy

significance. A 95% confidence interval

than females. The estimated log odds

for the true log odds ratio is (0.800,

ratio

asymptotic

2.000) and hence the 95% confidence

standard error is 0.187, which is highly

interval for the true odds ratio is (2.227,

significant at 1% level of significance. A

7.391). The confidence interval shows a

95% confidence interval for the true log

significant relationship between epilepsy

odds ratio is (-0.286, 0.448) and so the

and meningitis (Table 3).

95% confidence interval for the true

In the case of family history and epilepsy

odds

The

the estimated odds of having myocardial

confidence interval shows a significant

infarction are 3.266 times more those

relationship between sex and epilepsy

who have a family history of epilepsy as

(Table 1).

compared to those who do not have

In the next case (relationship between

family history of epilepsy. The estimated

injury and epilepsy) the estimated odds

log odds ratio is 1.184 and its asymptotic

of having epilepsy are 11.044 times more

standard error is 0.231, which is highly

for persons with injury than those who

significant at 1% level of significance. A

do not have injury. The estimated log

95% confidence interval for the true log

odds ratio is 2.402 and its asymptotic

odds ratio is (0.731, 1.637) and hence

is

0.081

ratio

is

and

its

(0.751,

1.564).

standard

error

is

0.306,

standard error is 0.397, which is highly

Risk factors associated with Epilepsy: A case-control study

Page | 512

HEALTH SCIENCE JOURNAL

Volume 6, Issue 3 (July September 2012)

the 95% confidence interval for the true

low

odds

The

deprivation) are also risk factors for

confidence interval shows a significant

epilepsy.16-18A potential limitation is that

relationship between family history and

there have been large changes in the

epilepsy (Table 4).

labor market in Greece during the study

In the last case (relationship between age

period. Lack of information on duration

and epilepsy) the estimated odds of

of employment was partly remedied by

having epilepsy are 0.747 times more for

the

persons having age less than 60 years

maintained the same occupation through

than those having age equal to or greater

two consecutive censuses. The results

than 60 years. Hence the persons having

showed

age equal to or greater than 60 years are

concordant occupational titles was 72%,

more likely to have epilepsy than those

suggesting a reasonable quality for the

who have age less than 60 years. The

census data. Moreover, it is important to

estimated log odds ratio is 0.292 and its

compare consistency within this study

asymptotic

0.188,

and between studies, as well as biological

which is highly significant at 1% level of

plausibility, before inferring causality. In

significance. A 95% confidence interval

addition, early onset may influence a

for the true log odds ratio is (-0.660,

persons choice of occupation, which

0.076)

confidence

may in turn have influenced the results.

interval for the true odds ratio is (0.517,

Low socioeconomic status is a risk factor

1.080). The confidence interval shows a

for the development of epilepsy. Low

significant relationship between age and

socioeconomic status is associated with

epilepsy (Table 5).

social

ratio

and

is

(2.077,

standard

so

the

5.136).

error

is

95%

education

analysis

that

and

and

of

socioeconomic

individuals

the

proportion

economic

who

of

deprivation,

unemployment, and low income, which

Discussion and Conclusions

in turn are associated with risk factors

In our research study, we had no data on

like incidence of birth defects, trauma,

individual

infection, and poor nutrition.19 Low

risk

factors

for

epilepsy,
for

socioeconomic status may influence the

is

risk of epilepsy through risk factors that

associated with several individual risk

are the same for epilepsy as for injury,

factors for cardiovascular disease

cardiovascular disease, and Alzheimers

however,

we

socioeconomic

did
status,

adjust
which

and

Alzheimers disease, many of which (like

disease.

For

unskilled

workers,

the

Page | 513
E-ISSN: 1791-809X

Health Science Journal All rights reserved

www.hsj.gr

Quarterly scientific, online publication by Department of Nursing A,

Technological Educational Institute of Athens

increased risk of epilepsy might be

different sites, which may dilute the risk

traceable, at least in part, to short

estimation.

duration of employment and associated

exposure to organic solvents and other

lifestyle factors. Alcohol intake or abuse

chemicals (such as that experienced by

is a risk factor for epilepsy, so higher

launderers and dry cleaners) and high

alcohol

alcohol

consumption

in

certain

Pathways

consumption.

may

include

However,

our

occupational groups may help explain

findings do not allow inferences about

raised risk of epilepsy in those groups.

causal relationships, so the findings

There is a striking similarity in the list of

remain tentative.24-26

occupational

groups

with

high

prevalence of alcohol consumption and

occupational groups with raised risk of

REFERENCES
1.Adelw C, Ahlbom A, Feychting M,

epilepsy.20 It is also possible that drivers

Johnsson

higher risk of accident injuries may

Tomson T. Epilepsy as a risk factor for

partly lie in the causal pathway: seizures

cancer. J Neurol Neurosurg Psychiatry

are a common complication of head

2006;77(6):784-786.

injuries. It is noteworthy that increased

F,

2.Ben-Manachem

Schwartzbaum

E.

Toward

and

more

risks were found for male launderers and

pragmatic view of driving and epilepsy.

dry cleaners. Chemical exposures occur

Epilepsy curr 2004;4(4):133-134.

frequently in these occupations; for

example,

exposure

chemical

cleaning

to

solvents

agents.

and

Earlier

3.World

Health

Epilepsy:

Organization

Etiology,

(WHO).

epidemiology

and

prognosis, 2001.

epidemiological studies have reported

4.Frigerio R, Elbaz A, Sanft KR, Peterson BJ,

that solvent exposure increases the risk

Bower JH, Ahlskog JE, et al. Education

of epilepsy. We have no information

and occupations preceding Parkinson

about exposure to specific chemicals, so

disease: a population-based case-control

it is hard to identify which kind of

study. Neurology. 2005;65:157583.

chemical agent may have been causally

5.Denays R, Kumba C, Lison D, De Bels D.

associated with raised risk of epilepsy.

First

Increased risks for these workers are not

occupational nickel poisoning. Epilepsia

consistent

2005;46:9612.

in

different

sexes

and

epileptic

seizure

induced

by

cohorts.21-23 These inconsistencies may

6.Dockerty JD, Draper G, Vincent T, Rowan

partly be explained by men and women

SD and Bunch KJ. Case-control study of

working at different types of tasks and

parental age, parity and socioeconomic

Risk factors associated with Epilepsy: A case-control study

Page | 514

HEALTH SCIENCE JOURNAL

Volume 6, Issue 3 (July September 2012)

level in relation to childhood cancers. Int

traditional

J Epidemiol 2001;30(6):1428-1437.

1999;40(3):382-386.

7.Gaitatzis A, Carroll K, Majeed A and

Sander J. The epidemiology

of the

comorbidity of epilepsy in the general

population.

Epilepsia

2004;45:1613-

1622.

cultures.

14.Johnston

A,

Epilepsia

Smith

P.

Sudden

unexpected death in epilepsy. Expert Rev

Neurother 2007;7:1751-1761.
15.Kanner AM,

Balabanov A. Depression

and epilepsy: How closely related are

8.Hauser WA, Annegers JF,

Kurland LT.

Prevalence of epilepsy in Rochester,

Minnesota:

1940-1980.

Epilepsia

1991;32:429-445.

they? Neurology 2002;58:S27-S39.

16.Ludvigsson P, Hesdorffer D, Olafsson E,
Kjartansson O, Hauser WA. Migraine
with aura is a risk factor for unprovoked

9.Fisher R, Van Emde Boas W, Blume W and

Elger C.. Epileptic Seizures and Epilepsy:

seizures

in

children.

Ann

Neurol

2006;59(1):210-213.

Definitions proposed by the (ILAE) and

17.Murphy JM, Horton NJ, Monson RR,

the international Bureau for Epilepsy

Laird NN, Sobol AM, Leighton AH.

(IBE). Epilepsia 2005;46(4):470-472.

Cigarette

smoking

in

relation

to

10.Gruman J, von Korff M, Reynolds J,

depression: Historical trends from the

Wagner EH. Organizing Health care for

Stirling County Study. Am J Psychiatry

people with seizures and epilepsy. J

2003;160:1663-1669.

Ambul Care Manage 1998;21:1-17.

11.Henry

TR,

seizures

Drury

in

candidates

with

I.

18.Park JO, Shin SD, Kim J, Song KJ, Peck

Non-epileptic

MD. Association between socioeconomic

lobectomy

status and burn injury severity. Burns

refractory

2009;35(4):482-490.

temporal
medically

seizures. Neurology 1997;48:1374-1382.

19.Matuja

WB,

Kilonzo

G,

Mbena

P,

12.Hesdorffer DC, Ludvigsson P, Olafsson

Mwango'mbola RL, Wong P, Goodfellow

E, Gudmundsson G, Kjartansson O,

P, Jilek-Aall L. Risk factors for epilepsy

Hauser WA. ADHD as a risk factor for

in

incident

community-based

unprovoked

seizures

and

epilepsy in children. Arch Gen Psychiatry

2004;61(7):731-736.

rural

area

in

Tanzania:

case-control

study.

Neuroepidemiology 2001;20(4):242-247.
20.Pramuka M, Hendrickson R, Zinski A,

13.Jilek-Aall L. Morbus Sacer in Africa:

Some religion aspects of epilepsy in

Van

Cott

AC.

psychosocial

self-

management program for epilepsy: A

Page | 515
E-ISSN: 1791-809X

Health Science Journal All rights reserved

www.hsj.gr

Quarterly scientific, online publication by Department of Nursing A,

Technological Educational Institute of Athens

randomized

pilot

study

in

adults.

Epilepsy Behav 2007;11:533-545.

21.Sidenvall R, Heijbel J, Blomquist HK,
Nystrm L, Forsgren L. An incident
case-control study of first unprovoked
afebrile

seizures

population-based
perinatal

in

children:

study

risk

of

pre-

factors.

A
and

Epilepsia

2001;42(10):1261-1265.
22.Thompson PJ, Upton D. The impact of
epilepsy

on

the

family.

Seizure,

1992;1:43-48.
23.Tromp SC, Weber JW, Aldenkamp AP,
Arends J, vander Linden I, Diepman L.
Relative influence of epileptic seizures
and epilepsy syndrome on cognitive
function. J Child Neurol 2003;18:407412.
24.Warnakulasuriya
cancer

risk

S.

Significant

associated

with

oral
low

socioeconomic status. Evid Based Dent

2009;10(1):4-5.
25.Wilson

SJ,

Bladin

PF,

Saling

MM.

Paradoxical results in the cure of chronic

illness: The "burden of normality" as
exemplified following seizure surgery.
Epilepsy Behav 2004;5:13-21.
26.Cascino G. D. Epilepsy: Contemporary
perspectives
treatment.

on

evaluation

Mayo

Clinic

and
Proc

1994;69:1199-1211.

Risk factors associated with Epilepsy: A case-control study

Page | 516

HEALTH SCIENCE JOURNAL

Volume 6, Issue 3 (July September 2012)

ANNEX
Table 1: Epilepsy versus sex.
Epilepsy
Sex
0
1
Total

0
56
144
200

1
132
368
500

Total
188
512
700

Table 2: Epilepsy versus head injury.

Epilepsy
0
193
7
200

Diabetes
0
1
Total

1
357
143
500

Total
550
150
700

Table 3: Epilepsy versus meningitis.

Epilepsy
Meningitis
0
1
Total

0
187
13
200

1
390
110
500

Total
577
123
700

Table 4: Epilepsy versus family history.

Epilepsy
Family History
0
1
Total

0
174
26
200

1
336
164
500

Total
510
190
700

1
160
340
500

Total
212
488
700

Table 5: Epilepsy versus age or < 60 years .

Epilepsy
Age
0
1
Total

0
52
148
200

Page | 517
E-ISSN: 1791-809X

Health Science Journal All rights reserved

www.hsj.gr