Corresponding author, Prof. Pierre-Marie Preux is the Plos One editorial board member. Drs Robert Sebbag and Daniel Gérard are employees of Sanofi France and the authors have received financial support from the department of access to medicines, Sanofi, Gentilly, France, there are no patents, products in development or marketed products to declare. This does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials.
Conceived and designed the experiments: DB P-MP MD MD-C SO SC. Performed the experiments: DB KC PH MV CH MD P-MP. Analyzed the data: DB. Contributed reagents/materials/analysis tools: DB P-MP MD-C KC. Wrote the paper: DB. Overall supervision: P-MP MD-C MD SO SC DG RS.
Identify epilepsy-associated factors and calculate measures of impact, stigma, quality of life (QOL), knowledge-attitude-practice (KAP) and treatment gap in Prey Veng, Cambodia.
This first Cambodian population-based case-control study had 96 epileptologist-confirmed epilepsy cases and 192 randomly selected matched healthy controls. Standard questionnaires, which have been used in similar settings, were used for collecting data on various parameters. Univariate and multivariate regression was done to determine odds ratios. Jacoby stigma, 31-item QOL, KAP etc were determined and so were the factors associated with them using STATA software. Treatment gap was measured using direct method.
Multivariate analyses yielded family history of epilepsy, difficult or long delivery, other problems beside seizures (mainly mental retardation, hyperthermia), and eventful pregnancy of the subject's mother as factors associated with epilepsy. There was high frequency of seizure precipitants esp. those related to sleep. Population attributable risk (%) was: family history (15.0), eventful pregnancy of subject's mother (14.5), long/difficult birth (6.5), and other problem beside seizures (20.0). Mean stigma (1.9±1.1, on a scale of 3) was mainly related to treatment efficacy. Mean QOL (5.0±1.4 on a scale of 10) was mainly related to treatment regularity. Cause or risk factor could be determined in 56% of cases. Treatment gap was 65.8%.
Factors in pre- and perinatal period were found to be most crucial for epilepsy risk in Cambodia which inturn provides major prevention opportunities. A global action plan for treatment, stigma reduction and improvement of QOL should be set-up in this country.
Medical and social domains of epilepsy in Cambodia are poorly investigated. Based on a first-ever large population-based door-to-door representative study (N = 16510), the lifetime prevalence of epilepsy in Prey Veng province is 5.8/1000
This population-based study was conducted in Prey Veng province and was an extension of the earlier population-based epilepsy prevalence study that was conducted in the same population
To determine case histories and KAP, we used detailed questionnaires that are already used in several population-based surveys in similar REP settings. Information on stigma was derived using the Jacoby stigma questionnaire
There were 96 cases and 192 matched controls. There were slightly more (53.1%) male epilepsy cases and the median age of the cases was 24.0 years, SD±13.6, range 3–70. Those <20 years represented 65.6% of the cases and the most frequent age group was 12–20 years (30.2%). On the other hand, 53.2% controls were males and the median age of controls was 28.0 years (SD±14.1, range 3–77). Five (2.6%) controls had some health problem, namely two subjects with chronic headache and one subject each with sleep disturbance, gastric ulcer, and weakness. All controls had a normal neurological clinical exam.
Generalised epilepsy consisted of tonic-clonic seizures (76.0%), absence (1.0%) and myoclonic seizures (3.1%) whereas partial epilepsy consisted of simple and complex partial seizures (12.5%), and olfactory partial seizures (1.0%). Mixed seizures consisted of 6.2% cases and 0.2% cases remained unclassified.
Factors associated with epilepsy and their respective odd's ratios are provided in
Parameter | OR crude | 95% CI | p-value |
Genetic factors | |||
Family history of epilepsy (any degree relative) | 2.9 | 1.4–5.6 | 0.0018 |
Perinatal factors | |||
Uneventful pregnancy of the subject's mother | 0.08 | 0.0–0.2 | <0.0001 |
Long or difficult pregnancy | 12.6 | 3.5–45.4 | <0.0001 |
Premature birth | 2.7 | 1.4–5.0 | 0.0013 |
Post-birth crying | 0.3 | 0.1–0.9 | 0.03 |
Use of drugs during pregnancy | 5.4 | 2.0–14.7 | 0.0009 |
Trauma/hospitalization related factors | |||
Currently hospitalised for any cause | 4.3 | 1.7–10.5 | 0.0014 |
Head injury with loss of consciousness before seizure onset | 7.4 | 2.3–23.5 | 0.0006 |
Comorbid factors | |||
Neurological sequel of any disorder | 28.5 | 3.6–222.8 | 0.001 |
Beside seizure, any other disorders | 8.2 | 2.9–23.1 | <0.0001 |
Others | |||
Mental retardation | 122.1 | 16.3–910.8 | <0.0001 |
Footnote: CI: Confidence interval; OR: Odds ratio.
Parameter | OR | 95% CI | p-value |
Family history of epilepsy | 3 | 1.16–8.14 | 0.02 |
Eventful (hemorrhage, hypertension) pregnancy of the subject's mother | 9.9 | 2.9–33.5 | 0.0002 |
Long or difficult birth | 8.52 | 2.08–34.89 | 0.002 |
Other problems beside seizure |
13.66 | 2.68–69.48 | 0.001 |
After excluding other problems beside seizure (mental retardation, hyperthermia) from the model | |||
Family history of epilepsy | 2.6 | 1.0–6.6 | 0.04 |
Eventful (hemorrhage, hypertension) pregnancy of the subject's mother | 8.6 | 2.6–28.2 | 0.0004 |
Long or difficult birth | 8.5 | 2.1–33.1 | 0.002 |
Other problems beside seizure: mental retardation, hyperthermia.
We enquired about several seizure precipitants among our cases and their presence was positively reported in the following manner: having seizures within one hour of awakening from sleep (18.7%), lack of sleep (18.6%), during sleep (16.4%), stoppage of treatment (13.1%), emotional disturbance (12.0%), alcoholism (5.4%), menstruation (4.3%), and hyperventilation (3.1%). Rest did not report any seizure precipitant.
The difference in proportions of various risk factors between case and control population was calculated. Excess proportion of risk factors was observed among cases than controls in the following manner: abnormal pregnancy of the subject's mother (+22.3%, p<0.0001) followed by other problem beside seizures (+15.1%, p<0.0001), long or difficult birth (+14.1%, p<0.0001) and family history of epilepsy (+14.0%, p = 0.0006).
Case fraction (CF) is the proportion of exposed among total study population. These fractions were calculated for various risk factors in our population and were: 14.5% (95% CI 3–50) from positive epilepsy family history, 10.0% (95% CI 4–21) from eventful pregnancy of the subject's mother, 6.2% (95% CI 4.7–7.5) from long or difficult birth, and 7.6% (95% CI 5.9–8.2) from other problems beside seizures (mainly mental retardation and hyperthermia).
The Attributable risk indicates the amount of epilepsy that can be attributed to a factor among the group exposed to a risk factor. The fractions (in %) for various risk factors in our population were: 66.6% (95% CI 50–86) from positive epilepsy family history, 89.8% (95% CI 85–92) from eventful pregnancy of the subject's mother, 88.2% (95% CI 83–93) from long or difficult birth, and 92.6% (95% CI 88–95) from other problems beside seizures (mainly mental retardation and hyperthermia). These figures indicate that 66.6–92.6% of those exposed to one of these factors could avoid having epilepsy as outcome.
The population attributable risk percentage (PARP) expresses the increase of risk in the general population (over that for a population which had no exposure) which is due to the exposure. It is the measure of weight of risk factor on the occurrence of a disease and in turn indicates the proportion of cases that can be eliminated from the total population if the factor is eliminated. We calculated the attributable risk for each factor that we obtained by using standard formula for case-control studies. We obtained population attributable risk of 15.0% (95% CI 4–51) from positive epilepsy family history, 14.5% (95% CI 7–27) from eventful pregnancy of the subject's mother, 6.5% (95% CI 1–21.0) from long or difficult birth, and 20.0% (95% CI 9–41.0) from other problems beside seizures (mainly mental retardation and hyperthermia). This risk also implies that the relation of these factors with epilepsy is causal at least statistically and overall in Cambodia a cause/risk factor can therefore be evidently determined in at least 56.0% of cases. Causal Pie Model for population of Cambodia is given in
We enquired about stigma by using three-question jacoby scale (
Parameter | Not at all (%) | May be (%) | Certain (%) | No response (%) |
Others being uncomfortable | 26 | 45.8 | 27 | 1.2 |
Considered as inferior | 31.2 | 48.8 | 18.7 | 1.3 |
Being avoided | 43.7 | 41.6 | 13.5 | 1.2 |
Details are presented in
Parameter | Response (%) |
Epilepsy is contagious | Yes 53.6; No 30.5; Unaware 15.9 |
Contagious by what | Hereditary 41.5; Saliva 21.5; Sharing food plate 17.6%; Sexual route 7.8; Rubbing shoulder/shaking hands 5.8; Gas emission 1.9%; Others 3.9% |
Epilepsy is curable | 84.3, non-curable 15.6 |
Curable by what | Traditional 2.4; modern 91.5; combination 6.1 |
Origin of epilepsy | Unaware 56.3; Natural 42.7; Supernatural 0, |
Attitude of those around | Highly protective 98.9; No response 1.1 |
Education of epilepsy child is possible | 68.7 |
Marriage of epilepsy child is possible | 33.3 |
Why not | Hereditary 53.9; Fear of contamination 23.8; Avoid bad luck 15.8; Fear of seizure 6.5 |
Taboos exist | 22 |
Examples of taboo | Fear of contamination 52.3; Avoid bad luck 47.6 |
Seizures interrupt work | 63.5 |
Those around permit you to travel | 47.9 |
Others DO NOT prevent you from using same smoke pipe or food plate | 75 |
Seizure interrupt practicing sports | 30.3 |
Others DO NOT prevent you from attending functions | 79.1 |
Others allow you to talk about your disease | 57.2 |
All cases were enquired about their treatment history. Traditional treatment was being used by 18.7% (3.1% of cases had changed from drug therapy to traditional treatment) and was mainly prescribed by subject himself (40.0%) or by traditional healer (53.3%) or someone else (6.6%). Traditional treatment was mainly herbal (60%) and nature of treatment for rest was mineral, mixed or unknown. None had animal product based treatment. Only 4 cases (26.6%) were taking traditional treatment regularly and its efficacy (subject's view) was good (6.6%), moderate (13.3%), bad (46.6%) or indeterminable (33.3%). Modern treatment was being used by 34.3% cases (60.6% on modern monotherapy, 36.3% on modern plus traditional treatment, 3.0% changed from traditional to carbamazepine). Out of these, 42.2% were taking modern treatment regularly and its efficacy (subject's view) was good (21.2%), moderate (54.6%), bad (18.1%), or none and unknown in 6.0% cases. Forty-three percent (43%) were not on any treatment and treatment type was unknown in 4.1% cases. The treatment gap was 65.8%.
The QOL using QOLIE-31 is measured on a scale from 0 to 10, with 10 representing the best possible quality of life and 0 representing worst possible quality of life. The median overall QOL was 5.0 (SD±1.4),
We conducted a first-ever large representative population-based matched case-control study in Cambodia that had epilepsy as a primary objective and used robust case-finding methods. All epilepsy cases were confirmed by epileptologists and each case had two matched controls randomly selected from the same source population. We aimed to identify factors that might have been associated with epilepsy in this population and in multivariate analysis, most of our risk factors were maternal-child-health related signifying the importance of this period for risk of having epilepsy. These factors are not different with results from other populations that had predominant generalised seizures (i.e. predominant absence of localised lesions) and epilepsy onset at a young age
Factor that was most significantly (at least statistically) found was «other problems beside seizures» and this mainly included mental retardation and hyperthermia. Mental retardation is independently associated with the risk of having epilepsy
Other factors that we obtained in our results were also pre- and perinatal such as eventful pregnancy of the subject's mother and long or difficult birth. These factors are independently and causally related to the development of lifetime and active epilepsy
A positive family history increases the risk to have epilepsy by two- to three-fold
Even though most seizures are considered to occur spontaneously, they can be precipitated by several endogenous or exogenous (seizure inducing and seizure triggering) factors. We enquired about presence of seizure precipitants and many cases reported a presence of seizure precipitants. Our frequency of precipitating factors is much higher than other studies
Stigma is often associated with epilepsy
Epilepsy since long is considered to be a contagious disorder and in our study nearly 54% cases reported epilepsy as a contagious disorder, mainly through hereditary route (41.5%), saliva (21.5%), sharing food plate (17.6%), sexual route (7.8%), rubbing shoulder/shaking hands (5.8%). However, some of the contagions reportedly observed in other populations were not reported by our population for instance reporting of breath, urine, faeces, blood, sperm and genital secretions as contagions
Median overall QOL in our population was 5.0±1.4,
The score for seizure worry was (36.3±22.1) lowest and highest for medication effects (median score of 83.3±24.0), a pattern that is also seen in nearby Asian populations
Treatment gap that is observed in our population is not different from some populations although much lower than some other populations such as Pakistan, Laos etc
Strengths of this study lies in its population-based design and use of standard procedures for epilepsy cases and their controls. Results from all major aspects of epilepsy are addressed in this study. Weakness of our study is similar to other case-control studies conducted in such settings due to the possibility of recall bias.
This study identified challenges and opportunities for Cambodia in relation to epilepsy,
|
Lack of treatment opportunities by modern anti epilepsy treatment |
Frequent seizure-precipitating factors |
Poor QOL possibly related to various precipitating factors and lack of treatment opportunities or absence of alternative compounds or severity of epilepsy itself |
Moderately high level of stigma. 46.0% cases with highest stigma particularly males |
Epilepsy is a contagious disorder for about 54.0% cases (mainly hereditarily) |
Origin of epilepsy is unknown for 56.0% cases |
Can't freely practice sports or work |
|
Major prevention opportunities exist. Strengthening of usual maternal-child health service can reduce number of new epilepsy cases by at least 41.0% |
Epilepsy is a treatable disorder for 84.0% cases |
Epilepsy is considered to be treatable by modern medicines rather than traditional |
Epilepsy is not a supernatural disorder for all cases |
Stigmatization of whole family and village is inexistent |
Extensive social support (others are protective) for 99.0% cases |
No limitations for education in 69.0% cases |
Can freely share smoke pipe and food plate or attend social functions |
More females than males are absent of any stigma |
Our results show that in at least 56.0% of cases the causative or risk factor can be evidently determined in Cambodia. The epilepsy-associated factors (pregnancy events, long and/or difficult delivery, mental retardation, febrile seizures) obtained in this population are preventable in nature. These factors thus provide opportunity for prevention by at least 41.0% should exposure to these risk factors is eliminated in the general population of Cambodia. Prevention of epilepsy can be aimed through strengthening of usual maternal and child health service (MCH). This prevention proportion matches with the results from other resource-poor populations through control of similar factors
Consistent and sufficient treatment remains one major challenge particularly in many resource-poor populations. Since traditional healers are usually approached for treatment needs, by minority of cases in this population though, it might be necessary to engage them in case identification, ascertainment and their treatment in future studies. They might be essential to bridge the gap between treatment and the target population since it is likely that neither traditional nor modern treatment can sufficiently address all epilepsy individuals alone.