This study was carried during 6 months from February to July 2017 in patients with epilepsy over than 18 years old followed at neurology department of Yalgado Ouedraogo University Teaching Hospital in Burkina Faso. Burkina Faso is a French speaking country located in West Africa which occupies an area of 274.300 K msq for a population of 16.248.558 according to the 2006 Burkina population census. The prevalence of epilepsy was 10.6/1000 in 1993 ( Debouverie et al., 1993 ). Youth is the one of fourth main hospitals of the country which received patients referred from peripheral health centers of Burkina Faso.

Patients whose the duration of following up is less than 3 months, and whose duration of epilepsy is less than 1 year and non-consenting patients were not included in the study.

Sample size and sampling procedures

The recruitment of patients was systematic and non-randomized. Sample size was calculated using single population proportion formula [n = 16 patients ((z_/2)2 × p × (1 − p))/d2]. By considering an assumption of 1.6% (0.01) proportion of epilepsy in Burkina Faso, the prevalence of depression in patients with epilepsy is unknown in our country, Zα/2 at 95% CI (1.96), and tolerable margin of error (0.05), the minimum sample size was 16.

We evaluate the presence of depression symptoms by using the Hamilton Depression Rating Scale. The Hamilton Depression Rating Scale (HDRS) was administered to evaluate the level of depression in epilepsy patients ( Hamilton, 1960 ). HDRS consisted of 21 questions but only the first 17 were scored. Eight items are scored on a scale of 0 (not present) to 4 (severe) and nine are scored as 0 - 2. Patients with score 8 or more were classified as being depressed. The severity of depression was graded for both measures using the acceptable standard scoring system: for the HDRS-17: normal―1 - 9 score; mild mood disturbance―10 - 13 score; borderline clinical depression―7 - 20 score; moderate depression―21 - 30 score; severe depression―31 - 40 score; and extreme depression. The definition of epilepsy was the 2017 International League against Epilepsy definition ( Scheffer et al., 2017 ). The duration of epilepsy was calculated as the interval between the date of the first attack ever (corroborated by a close family member) and the first contact with the patient in our neurology clinic. The duration of treatment was calculated as the interval between the first treatment and the first contact with the patient in our neurology clinic. Medication adherence was measured using the self-reported 6-item of Girerd questionnaire, a commonly used and validated instrument ( Girerd et al., 2011 ). The frequency of seizures was defined as described into: 1) very frequent (occurring several times a day or at intervals shorter than 7 d; 2) frequent (at intervals longer than 7 d but shorter than 30 d); 3) occasional (at intervals longer than 30 days but shorter than one year; and 4) rare (at intervals longer than one year. Regarding the degree of control on AEDs, patients were considered controlled on AEDs treatment, when seizure free for ≥1 year and were uncontrolled when seizures were frequent, very frequent, occasional or rare. All patients underwent neuroimaging [such as computed tomography (CT) or magnetic resonance imaging (MRI)].

The data were collected by an end-of-life medical student trained in administering the Hamilton Scale, Gererd questionnaire and analyzed using of the Epi-info version 7 software. In each patient, we collected sociodemographic characteristic (age, sex, profession, martital status, education, residence), characteristics of epilepsy (duration of disease, age at disease onset, EEG, neuroimaging, etiologies, treatment of epilepsy, frequency of seizures and outcome of seizures), draw back social consequences (psychological support, financial support, perceived stigma, perceived of disease) and depression (history of psychiatric disorders, antipsychotic treatments, family history of psychiatric disorders, results of HDRS). The qualitative variables were compared using the Pearson Khi 2 test with a 95% confidence interval, and the averages were performed using the Student’s test. The probability was significant if p < 0.05.

The study proposal was initially approved by the ethical review board of University of Ouagadougou. A formal letter of permission was obtained from the hospital and submitted to the respective outpatient department. The information about the study was given to the participants. Verbal and then written informed consent was sought for each participant who agreed to participate in the study and full filled the inclusion criteria. Only anonymous Patient informed consent was required prior to inclusion in the study. Patients diagnosed with severe depression were referred to psychiatry department for appropriate management.

One hundred and two patients with a mean age of 41.47 ± 16.67 years (Range 19; 88 years) were included in this study. The sample consisted of 59 (57.8%) male patients and 43 (42.1%) female patients with a sex ratio of 1.37. According to the working status, 22 (21.5%) patients were housewives, 18 (17.6%) civil servants, 17 (16.6%) farmers and 17 (16.6%) students. According to educational level, 65 (36.2%) patients were illiterate, 20 (33.3%) had secondary level of education. According to marital status, 51.9% were married, 40.2% were single. Fifty neigh (57.8%) patients had living in couple. Sixty five patients (68.6%) were coming from urban area, 11 (10.7%) from semi-urban and 26 (25.4%) patients from rural area.

The mean duration of disease was 7.01 years (Range 3 - 34 years). The majority of patients (n = 67%; 65.6%) had past history of generalized tonic clonic convulsions, while 35 (34.3%) had focal seizures and one patient had focal and generalized seizures. None patient had acute seizure and history of psychiatric history. Laboratory investigations revealed that 42 (41.1%) patients had done electroencephalography. Of whom, 26% had focal parttern and 43% generalized pattern. EEG was normal in 31% of patients. CT scan was performed in 64 (62.7%) patients and was abnormal in 48 (75%) patients. Patients were treated with one AED in 85.2% [PB (n = 62% or 60.7%), CBZ (n = 24% or 23.5%, VPA (n = 16% or 15.6%)] and bitherapy in 15 (86.6%) patients [(n = 22% or 20.95%)]. CBZ was the most AED prescribed in biotherapy (n = 12% or 80%). The mean duration on AED was 4.08 years (3 months - 25 years). The majority of patients (n = 65% or 63.7%) were uncontrolled (had frequent or very frequent seizures). Almost all of (99%) patients had a good tolerance of the treatment. According to Girerd questionnaire, adherence of AED was good in 83 (81.3%) patients, average in 13 (12.7%) patients and poor in 6 (5.8%) patients. According to social consequences, ninety eight (96%) patients had psychological support and 75 (73.5%) patients had financial support from their family. According to the perception of disease, sixty six (64.7%) patients had perceived shamed of epilepsy, 44 (43.1%) had high perceived stigma and 25 (24.5%) had mystic and diabolic perception of epilepsy.

Sixty eight patients (67.33%) with epilepsy had depression (HDRS > 10). The mean age of patients with depression was 42.59 ± 17.78 years (Range 19 - 88 years). All of them were undiagnosed and untreated before for depression. The depression occurred in the post ictal period in 50 (78.13%) patients. We identified depression symptoms in 57.6% of male patients and in 79% of female gender. According to severity, depression was mild in 24 (35%) patients, moderate in 23 (34%) and severe in 21 (31%). patients. Moderate to severe depression was seen in 65% of patients with epilepsy. The average depression scale was 15.62 ± 4.26 points (ranges 10 - 28 points). Table 1 presents the distribution of subtypes of depressive disorders among patients with epilepsy according to HDRS. Hypochondriasis (97%) was the most symptoms of HDRS-17 followed by work and interest (95.5%) and anxiety-somatic (94.1%). Higher scores were found for work and interests, anxiety―psychic, hypochondriasis, depressed mood and anxiety―somatic symptoms. Low scores were obtained for the following items: insight, agitation and genital symptoms. In conclusion, insomnia, anxiety and somatic symptoms were observed respectively in 43 (42.16%), 88 (86.27%) and 48 (47.06%) patients.

On bivariate analysis, the sociodemographic factors associated with depression (p-value < 0.05) were high perceived stigma and female gender (p = 0.02). Other variables were not found to be significant risk factors of depression among epileptic patients including marital status (p > 0.05), profession (p = 0.32), residence (p = 0.55), antiepileptic treatment (p > 0.05), duration of illness (p = 0.29), level of education (p > 0.05). Table 2 presents the distribution of study subjects by demographic factors in patients with epilepsy. Social risk factors of depression was respectively financial support (p = 0.004) and perceived stigma (p = 0.00000075). Table 3 presents the social risk factors of depression among patients with epilepsy. Epileptic factors associated to depression were the occurrence of seizures during the last 3 months (p = 0.0014) and uncontrolled seizures (p = 0.002). Table 4 presents epilepsy risk factors of depression.

Our study had showed a high prevalence of depression among patients with epilepsy (63.3%). Similar prevalence was observed in several studies from Asia: 60% in India and Paskistan ( Pandian & Mohan, 2015 ; Yousafzai et al., 2009 ) and 63% in Korea ( Lee et al., 2010 ). In comparison with others studies in Sub Africa, the prevalence of depression was low than in the study of Onwuekwe ( Onwuekwe et al., 2012 ) in Nigeria (85%), but higher than others studies ( Owolabi et al., 2016 ; Tsegabrhan et al., 2014 ) in which the prevalence was ranging from 20.4% to 49.3%. According to literature ( Barry, 2003 ), high prevalence of depression is probably due to the lack of acceptance of the diagnosis of epilepsy. In our study 24.5% of patients had mystic and diabolic perception of epilepsy for many patients; epilepsy is due to sociocultural features. The differences of prevalence between the studies are possibly due to the different methodology in different studies including method of patient selection (e.g. hospital or community based), the method of evaluation for depression (different scales) and sociocultural characteristics of population study.

Depression was moderate to severe in the majority of depressed patients in our study (65%), in line with the studies of Hamed in Egypt with 78.43% ( Hamed et al., 2012b ). and Tsegabrhan et al. in Ethiopia with 60.1% ( Tsegabrhan et al., 2014 ) but in contrast with Nidhinandana et al. in India ( Nidhinandana et al., 2007 ) who had found a majority of patients with mild depression (65.2%). Hypochondriasis is the most symptom of depression seen in 97% of depressed patients. Suicidal ideation was observed in only 22% of this patients, in line with studies from Ethiopia with 29.8% ( Haile et al., 2018 ) and Egypt with 23.5% ( Hamed et al., 2012a ) but higher than in literature ( Jones et al., 2003 ), Andrijić in Bosnia ( Andrijić et al., 2014 ) with 14% and in USA ( Hecimovic et al., 2012 ) with 11.9%. In opposition, highest prevalence of suicidal ideation was observed in studies from Cuba Havana with 45.2% ( Espinosa et al., 2010 ) and Egypt with 55% ( Hamed et al., 2012b ). Furthermore, in Cuba Havana, study subjects were patients with temporal lobe epilepsy, but our study included all people living with epilepsy followed in neurology department. Thus, the presence of suicidal ideation should prompt immediate referral ( Kanner, 2003 ).

According to the perception of disease, about forty three percent of our study population (43.1%) had high perceived stigma. High perceived stigma was associated with the development of depression, in line with several studies in African setting, as showed in the study of Bifftu ( Biffu et al., 2015 ) and Adewuya in Nigeria ( Adewuya & Ola, 2005 ). Social stigma is due to the danger and unpredictability of the illness, and the psychotropic effects of antiepileptic drugs ( Mula & Schmitz, 2009 ). Female gender was associated to depression symptoms in our study, in line with literature ( Hamed et al., 2012b ; Thomsonand & Brennenstuhl, 2009 ). In contrast, male gender was significally depression than women in the study of Yousafzai ( Yousafzai et al., 2009 ). As in several studies, we did not found differences between marital status and depression ( Thomsonand & Brennenstuhl, 2009 ; Zahiroddin et al., 2008 ), age and depression ( Nidhinandana et al., 2007 ; Asadi-Pooya & Sperling, 2011 ), level of education and depression ( Thomsonand & Brennenstuhl, 2009 ; Kutlu et al., 2010 ). Epilepsy-related factors associated with the development of depression in PWE were occurring of recent seizures in the last 3 months and uncontrolled seizures were associated to depression (p < 0.05). These features are in line with several researchers according to uncontrolled seizures ( Hamed et al., 2012a ; Ogunrin & Obiabo, 2010 ). Burkina Faso lacks the person power and infrastructure to manage epilepsy well for the majority of patients in need. For example, for a country of almost 16 million people, Burkina Faso has only 9 neurologists, 3 neurosurgeons and approximately 13 psychiatrists. The most of these heath workers are concentrated in the capital city, Ouagadougou. According to seizures frequency, having recent seizures is a risk factor of depression ( Owolabi et al., 2016 ; Tsegabrhan et al., 2014 ; Biffu et al., 2015 ). Seizure frequency of less than 1 per month predicted postictal depression ( Andrijić et al., 2014 ). In our study, AEDs were not associated with depression. According to literature, phenobarbital and other barbiturates have long been associated with increased risk of depression and suicide ( Barry, 2003 ; Miller et al., 2008 ; Salzberg & Vajda, 2001 ). In our study, phenobarbital was the most prescribed drug (60.7%).