You are on page 1of 12


Prevalence, risk factors, and neurobehavioral
comorbidities of epilepsy in Kenyan children
*Charles J. Kind, *†‡Charles R. J. C. Newton, †Symon M. Kariuki , and afor the
Neurodevelopment Disorders study group

Epilepsia Open, 2(4):388–399, 2017
doi: 10.1002/epi4.12069

Objective: To investigate the prevalence, risk factors, clinical features, and neurobe-
havioral comorbidities of epilepsy and acute symptomatic seizures in school-aged chil-
dren in Kilifi, Kenya.
Methods: Randomly selected children (N = 11,223) were screened for epilepsy and
other neurodevelopmental disorders. Those who screened positive were invited for
further clinical, electroencephalographic (EEG), and neuropsychological evaluations.
Prevalence was measured by dividing cases by screened population, providing Agresti–
Coull confidence intervals (CIs). Prevalence ratios were computed using log binomial
regression, and odds ratios (ORs) were computed using logistic regression; both were
implemented with generalized linear models. Attention-deficit hyperactivity disorder
(ADHD), autism spectrum disorder (ASD), and other neurodevelopmental impair-
ments were assessed in cases and controls.
Results: Prevalence of lifetime epilepsy was 20.9 per 1,000 (95% CI = 18.4–23.7), and
that of active epilepsy was 11.5 per 1,000 (95% CI = 9.7–13.6). Prevalence of acute
symptomatic seizures was 68.8 per 1,000 (95% CI = 64.2–73.6). Acute symptomatic sei-
zures preceded a diagnosis of epilepsy in 8% of children. Of 98 children diagnosed with
epilepsy, focal seizures were seen in 79%, abnormal EEG was seen in 39%, and 83% were
not receiving antiepileptic drugs. Childhood absence epilepsy and Lennox–Gastaut epi-
Charles J. Kind is a lepsy were the most easily identifiable epilepsy syndromes. Perinatal complications,
medical student at St. previous hospitalization, geophagia, and snoring were risk factors for epilepsy. Family
John’s College, history of seizures, abnormal pregnancy, previous hospitalization, and snoring were
Oxford, United risk factors for acute symptomatic seizures. Neurobehavioral comorbidities were pre-
Kingdom. sent in 54% of subjects with lifetime epilepsy and in 3% of controls, with associations for
individual comorbidities being statistically significant: ADHD (OR = 14.55, 95% CI =
7.54–28.06), ASD (OR = 36.83, 95% CI = 7.97–170.14), and cognitive impairments
(OR = 14.55, 95% CI = 3.52–60.14).
Significance: The burden of seizure disorders in this area is higher than in locations in
high-income countries, and can be reduced by preventing risk factors. A comprehen-
sive management plan for neurobehavioral comorbidities of epilepsy should be incor-
porated into standard epilepsy care.
KEY WORDS: Convulsive epilepsy, Prevalence, Neurobehavioral comorbidities,
Children, Africa.

Epilepsy is common in sub-Saharan Africa,1 where it carried out, with data for the most recent study col-
exerts an enormous burden on the health system,2 and lected more than a decade ago.5 Additionally, no study
is associated with substantial mortality relative to else- has attempted to describe epilepsy syndromes, and there
where in the world.3 Children bear the largest burden of are few studies of the comorbidity with other neurodis-
the disease.4 Despite this, only a small number of epi- abilities in African children, despite evidence for their
demiological studies of epilepsy in children have been existence.6,7


10 KHDSS is hosted at KEMRI-Wellcome Trust epilepsy in 8% of children. which is located about 60 km north tion with nontraumatic acute encephalopathy of Mombasa. Kariuki. Epilepsia Open published by Wiley Periodicals Inc. Research Programme. did not truly have a seizure disorder) for the NDST was excel- The copyright line for this article was changed on June 05 2018 after origi. lent (99.6–99. duration. and acute symptomatic seizures occurred in 69 of 1. does the child sometimes have fits..000. The negative Trust Research Programme. distribution and reproduction in assessors took a clinical history that included presence of any medium. Oxford.g.000 in Kenya. and Kilifi County is with epilepsy.5 Although wide variation in febrile seizures are uncommon in this group. Stage I involved screening of tomatic seizures should be systematically screened.3 per 1. which permits use. These data are required to inform for further clinical assessment in stage II. Additional risk factors for childhood mental that used the same age group. overesti.0%. KEMRI-Wellcome 95% (95% confidence interval [CI] = 94–96%). United Kingdom cal survey. so they would estimates may be attributed to differences in risk factors not confound epilepsy estimates. Neurodevelopmen- decisions on comprehensive preventative and therapeu. method. as epilepsy and other neurodevelopmental disorders in the com- they can easily be misdiagnosed as epilepsy. snoring and geophagia) have been screening about 15. on behalf of International League Against a Members of the NDD group are provided as Appendix S1. childhood absence and Lennox–Gas- taut epilepsy are the most easily identifiable epilepsy one of the poorest administrative regions in Kenya. particularly previous hospitaliza. Kilifi.12069 . 2017.5 We had calculated that health problems (e. which is located in Kilifi County. clinician still saw a proportion of those who screened nega- © 2017 The Authors.5%). tive in stage I to confirm the seizure status. and neurobehavioral comorbidities were described by Study design and procedures trained clinicians. Kenya. seizures. although in stage II the nal online publication. In stage II. United for epilepsy and acute symptomatic seizures in stage II. aimed at reducing the large treatment and was administered to the parents or close caregivers of the gap for epilepsy in children.1002/epi4. 389 Epilepsy Epidemiology in Kenyan Children Key Points Methods • Epilepsy occurred in 21 of 1.11 The questionnaire has one seizure question (Since birth. and the two conditions had shared risk factors. Kingdom. and types and use of Epilepsia Open. who form a total population of about children range between 7.000 randomly selected subjects from the documented in recent studies. We con. We recognized that acute symp. or lose consciousness?). The literacy levels are low. PO Box 230 (80108). nity. it may also be due to methodological ison with a previous study conducted around 10 years ago differences. seizures with a precision or margin of error of <1%. and to allow for compar- within the sites. the majority of whom are subsistence farmers. frequency. †KEMRI-Wellcome Trust Research Programme. The age group of 6–9 years was chosen because and 41 per 1. 2(4):388–399.000 defined area known as the Kilifi Health and Demographic • Acute symptomatic seizures preceded a diagnosis of Surveillance System (KHDSS). Kenya. University of Oxford. havioral comorbidities such as attention-deficit hyper- activity disorder and autism spectrum disorders Sampling This study involved children aged 6–9 years living Studies reporting on the burden of epilepsy in African within KHDSS. 95% CI = 98. University of Oxford. 2017 doi: 10. There is syndromes one epilepsy and neurodevelopmental clinic run by • Epilepsy in older children is associated with neurobe. provided the original work is properly cited. eligible child. for other neurodevelopmental questions were also assessed *St. trained clinicians and neuropsychological This is an open access article under the terms of the Creative Commons Attribution License. The main population in KHDSS is the Mijik- • Focal features occur in about 79% of Kenyan children enda ethnic group. Kilifi. graphic features of epilepsy were characterized.9 but it is unclear whether 28. the NDST detected epilepsy with a sensitivity of Address correspondence to Symon M. assum- ducted an epidemiological study that employed sensitive ing a prevalence of seizure disorders of 5% in the commu- screening methods to estimate the prevalence of epi.000 Kenyan children Study site and population This study was conducted on the Kenyan coast in a (aged 6–9 years). Clinical and electroencephalo. tal Screening Tool (NDST) was used for screening in stage I tic interventions. munity. Johns College.000 in South Africa8 28.000 children will detect epilepsy and acute symptomatic they are important in children with epilepsy. researchers at KEMRI-Wellcome Trust Research Pro- gramme and Kilifi County Hospital (KCH). but those screening positive Accepted July 6. According to a pilot study nested in this main epidemiologi- and ‡Department of Psychiatry. become rigid. The RAND() command of MySQL was used to select lepsy and its associated risk factors in children from a the eligible children using a simple random sampling rural area in Kenya. onset. E-mail: predictive value (the probability that those screening negative skariuki@kemri-wellcome. This was a two-stage design. Oxford. Those who screened positive in stage I were invited mating the burden.

390 C. so that each child could be categorized for more than one seizure type. Focused probability that is close to nominal confidence level. 129 of whom had active epilepsy.5%) were seen by a clinician who assessed them as hav- bulletin. and group discussions and in-depth interviews were conducted does not produce narrow intervals as the exact or Wald with adults in the community to elicit phrases and idioms to intervals approach would with large sample sizes such as in be used in the translated version. lepsy was 98 (Fig. which has a coverage standardized forward and back translation process. Generalized linear models age-matched controls. Prevalence was computed by dividing vation Schedule. types. or Fisher exact test where observations tors for neurodevelopment.13 children gave informed consent for their children to Parents were asked to describe the three most recent seizure participate. Kiswahili.14 ing epilepsy. Stata Corp. acute symptomatic seizures. Pearson chi-square. antiepileptic drugs (AEDs). College Station. for those without watches. and put on clothes if of NDST developmental disorders. Status Results epilepticus was defined as seizures lasting for 30 min or A total of 11. Kind et al. generalized. 2(4):388–399. of whom 2. including randomly selected estimates by sex and age group. the factors associated with either epilepsy or acute symp- The questionnaire was designed for use in neurodevelopmen. following up to three invitations) between stage I and trained neuropsychological assessors supervised by a II.1002/epi4.367) and sex (p = 0. those who fail to turn up in stage II for further clinical evalu- vey.18 The prevalence accounted for sensitivity of the their appropriateness in assessing neurodevelopment and screening tools at stage I. tomatic seizures. Log binomial regres- A risk factor questionnaire was given to the parents of sion was used to compute prevalence ratios for prevalence every child assessed at stage II. clinical evaluation of those who screened positive for other and unable to sit. stand upright. medical history. attention-deficit hyperactivity disorder Data were entered using MySQL. and child habits.640 mate duration such as boiling a pot of maize. The risk factor questionnaire had items with a logit were used to compute the odds ratios (ORs) of on socioeconomic status.17 Cognition was considered impaired if the erated from those who screened positive in the NDST speci- child had a standardized z score of <2. Scientific and Ethics Review unit of KEMRI. Of these. The distribution conducted. U. or neither. The tools were administered by experienced and ation. weekly. All the tools including NDST were trans. whereas Of the 1.640 children seen by the clinician in stage II.4%) were gen- described. 1. Schizophrenia. 69 (70.9%) were positive for the specific seizure question in zures occurring during the entire life of the child. a radio news (69.223 children aged 6–9 years were initially intermittent seizures for a period of 30 min as timed by a screened.15 Those with unprovoked seizures in the past collected at the clinical assessment stage. Bonferroni correction was used to performed according to the international 10–20 system.087). if appropriate. Of the 98 cases of epilepsy. 12 months were deemed to have active epilepsy. previously described. Those who did not attend the assessment were similar to those who Diagnosis and definitions did in terms of age (p = 0. Parents of tion and Terminology recently published in Epilepsia. and analyzed using (ADHD) with Kiddie Schedule for Affective Disorders and STATA software (version 13. 2017 doi: 10.16. active epilepsy was seen in zures provoked by a febrile illness. 556 lifetime epilepsy included a history of unprovoked sei. Seizure frequency was cate- gorized into daily. CIs were based on lated into the local language.361 (21. of these.0%) screened positive accord- watch or. observed cases by the total population. Prevalence was operationally defined as tion between stages I and II. Discrete variables were compared using tal studies following a thorough review for possible risk fac. and autism with Autism Diagnostic Obser.12 We classified seizures as focal. or milking a cow.12069 . monthly. whereas the remainder were from were present if a child was unable to hold implements. or others as Ethical approval recommended by the position paper of the International Permission to conduct this study was obtained from the League against Epilepsy (ILAE) commission for Classifica. After accounting for attri- appropriate age. and yearly. Motor deficits fic seizure question. (33. The total number of children diagnosed with epi- 9 years. S1). by dividing crude prevalence with sensitivity and attri- psychologist. through a the Agresti–Coull interval approach. of epilepsy cases in those who screened positive and in those Epilepsia Open. and for attrition (proportion of adapted accordingly before use in the epidemiological sur.). Epilepsy was defined as a history of two or more Sociodemographic comparisons between the attendance unprovoked seizures according to recommendations by groups cannot be made because this information was only the ILAE.S. the adjusted total number of children with epilepsy of epilepsy as of the day the epidemiological survey was was 234. Cognition was measured with Statistical analysis Ravens Matrix Test. up to stage I. J. as previously 54 (55%). Acute symptomatic seizures were defined as sei.1. They were piloted to test this study. and sensitivity of the screening the proportion of children aged 6–9 years with a history tools. Electroencephalography was in a cell were sparse (<5). as adjust for multiple testing in risk factor analysis. ing to the NDST in stage I screening. events of that approxi. TX. tion (both expressed as proportions).A.

3 (25. and it was similar between active epilepsy and inactive epilepsy.8 (3.2–10.9 (27.9) 9.9–11.64 (0.520).8%. Based on the hospital records for these 124 cases treated for acute symptomatic sei- Prevalence active zures.6) 4.07 (0.2) 1.008). The adjusted prevalence was 68.646 5. screened nontreatment was not significantly more frequent in status Number 11. p = 0. Table 2). Table 2). p < 0.b of 98 (7.6–40.4 (18.5%).3–27.01 (0. or weekly seizures (p = 0.71–1. Self-reported Includes lifetime epilepsy only. As would be expected. 15.62–2. daily seizures in 17 (17.2–73.577 those without these features.9 per 1.0–6. S1.4) 4. with fever in 28 (22. The distribution of prevalence by age and sex is shown in Table 1.1% were focal.3% vs.27 (0. seizures Prevalence ratio computed with log binomial regression.87–1.4 (6.1–7.7) epilepsy per 16.0 (6.58) 1.223 11.8) 3.1 (3.3) 5.3–33.59) for epilepsya.8 (64.004) were significantly 234 10 29 36 23 51 47 98 more frequent in active than inactive epilepsy. 6 7 8 9 Sex a b c daily seizures (p = 0.9 (2.4–23. Generalized seizures were more common in seizures active (81.223 epilepticus (p = 0.324) than 1.5%).2 (2.2) 1. self-reported AED use was lifetime epilepsy reported in 13 (17.00 — — cant differences in prevalence of epilepsy between age groups or sex.0 (19. About one-quarter (24.00 1.00 — — acute symptomatic seizures.5%) than inactive epilepsy (56. Female Male Age. and generalized tonic–clonic seizures were the most com- CI.2) 25. S1). yr p = 0. Of all the focal seizure types.45) 1. Baseline categories for age and sex are 6 years and male sex.4%).6) epilepsy per 6.7 (21.1 (3.8 per 1.7%) than inac- Number with tive epilepsy (59.016) and weekly seizures (42.8 (7.5–7. Age at onset of seizures was inversely Crude total correlated with self-reported use of AED (q = 0. confidence interval.000 (95% CI = 64.000 (95% CI = 9.013). the 42 98 99 69 most common was focal impaired awareness (76.7) 9.5%.465). respectively.5%) children with epilepsy.4–23.6) seizures was 308 (Fig.8–9.7–34.00 1.295) or daily seizures (p = 0.07 (0.636 2. daily (25.3) 27. There were no signifi- 1.9% vs.2%) were treated at KCH in the last 6–9 years.5–36. there active epilepsy were significantly more of these in active (90.4 (24.8%.11–1. Table 1).3–11. 2(4):388–399.9%).6) 68.8) 31. The age at onset of seizures among those with epilepsy was 3 years (interquartile Extrapolated totalc range = 1–6).6%) and generalized seizures (70.5 (3.1–16. status epilepticus in 26 (26. Other significant differ- ences in seizure types between active and inactive epilepsy 129 are shown in Table 2.651 3.9%. the most important diagnoses associated with acute 11.0001).53) 1. 2017 doi: 10.2–10.000 (CI) 25. p = 0. according to the computed prevalence ratios. focal (78.1 (3.6.7).39 (1.85–3.7 (7. 391 Epilepsy Epidemiology in Kenyan Children who screened negative is shown in Fig.9) 8.5 per seizures 1. p = 0.2–6. which was based on the Agresti–Coull interval approach.75–1.1–13.2%) of acute symptomatic seizures treated at the hospital were repetitive. but prevalence was significantly greater in males than females. 6.12069 .7–13. 180 128 308 772 mon (57.1) 32.8) 23.74) CI = 18. The adjusted Prevalence ratio for acute symptomatic prevalence of lifetime epilepsy was 20. whereas Prevalence lifetime 20.28.000 (CI) 6. 10 19 14 11 29 25 54 Of those with epilepsy.7–6. 204 (66.6 (5.2) 27. Prevalence of epilepsy and acute symptomatic seizures in the year 2015.6–30.26 (0. and Number with Adjusted for attrition and screening sensitivity.2 (22.6).81–3.9 (7.4%) described any of acute symptomatic the three most recent episodes of seizures (not mutually Number with exclusive). and gastroenteritis in 16 (12.7%.2–73.1–11. and that of active epilepsy was 11.2) 7.3%) treated for 1. symptomatic seizures The total number of children with acute symptomatic Prevalence acute per 1.5 (9.72–1.1002/epi4. There were no significant differences in prevalence of acute symp- Table 1.7–13.6) Clinical features of epilepsy Of 98 children diagnosed with epilepsy in stage II. respiratory tract infections in 38 (30. but not with status epilepticus (p = 0.1%).803 3.2) 8. with 124 (40.1%. Acute symp- tomatic seizure preceded the diagnosis of epilepsy in 23 1. Epilepsia Open.000 (CI) 5.74) 1. 10.133 5.684).7%).000 (95% 1.9 (18.78 (0.3) symptomatic seizures were falciparum malaria in 62 (50%).1–11.0 (3.5) 5.1–31. Of the 308 cases of acute symptomatic seizures.29) Prevalence ratio 1.86) 1. by sex and age tomatic seizures between age groups.

8) 0. 2017 doi: 10. and motor impair- (0.918). n (%). N = 44 v2 pa Generalized seizure types Any generalized seizureb 69 (70. Electroencephalographic features centrotemporal spikes. status epilepticus symptomatic seizures were compared to 471 controls who or self-reported AED use. focal motor seizures comprise automa- tisms and clonic. There were no differences between those who Risk factors for epilepsy and acute symptomatic seizures attended and those who did not attend the EEG recording in The risk factors for lifetime and active epilepsy and acute terms of age (p = 0.3) 12 (22.6) 24 (44.54–28. and snoring as p = 0.55.1) 34 (63.5) 35 (63. Focal epileptiform discharges were observed in 13% of more frequent in those with epilepsy (53 of 98.1) 2 (3.83. General.5) 3 (6.5) 49 (90.173 Yearly 42 (42. ciation was particularly great for autism spectrum disorder Some EEG traces displayed patterns characteristic of (ASD. parental marital status screened negative for seizures in stage I.680).06).7) 26 (59. There were no significant differences between cases Comorbidities associated with epilepsy of active and inactive epilepsy in terms of EEG abnormali.6) 50 (92. a Comparisons are between active and inactive epilepsy. previous hospital admissions. The ORs Table 2.7) 3 (6. Three chil. generalized nonmotor/absence seizures comprise typical. tonic. Among those with lifetime epilepsy.501 Focal impaired awareness 75 (76. but the occur.001 Focal sensory 15 (15. Kind et al. ized epileptiform discharges were the most common abnor- mality. n (%). which includes all those that are not included in the other three categories. People with snoring as risk factors for epilepsy.6) 7 (15. eating soil.02%) had temporal Sixty-four out of the 98 (65.654d Impaired awareness 27 (27. atonic.4) 44 (81. OR = 36.5) 9 (20.5) 0.008 Generalized tonic–clonic 56 (57. community controls without neurodisability (22 of 643.77–1.9) 0. The same model active epilepsy were more likely to attend the clinic for the (Table 4) identified a family history of seizures.9) 25 (46. 95% CI = 7.03%) had benign epilepsy of childhood with ments (OR = 34. because a child can have different seizure types across multiple episodes of seizures. 3%.6) 27 (61. b Not including focal to bilateral tonic–clonic seizures.197 Status epilepticus 26 (26. 95% CI = 3.0) 22 (50.138). Epilepsia Open.9) 3 (6. e Seizure frequency categories are not mutually exclusive.05%) had childhood absence syndrome. n (%). cations.14). rence was not different from those without these seizures p < 0. acute symptomatic seizures. pregnancy.7) 0.617 AED. ADHD well-described epilepsy syndromes (Fig.1) 8 (19.934 Generalized nonmotor/absence 20 (20. two (OR = 14.004 Monthly 57 (57. myoclonic.0001). the asso- types (11%.797).4) 17 (1.12069 . hyperkinetic.0) 0.5) 17 (31. sex (p = 0.97–170. S2). myoclonic.8) 0.016d Weekly 30 (30. cognition dren (0.8) 0. antiepileptic drug.2) 3 (6. atonic.6) 22 (50.0) 19 (43. maternal eco. and epileptic spasms.3%) children with epilepsy lobe epilepsy. All percentages are derived from observed numbers divided by the totals in the column heading. d Fisher exact test used when any value for this variable was <5.0) 0. paternal schooling (p = 0. 50%.4) 14 (25.107).6) <0. abnormal EEG than those with inactive epilepsy (76% vs. Table 2).51. atypical. risk factors for acute symptomatic seizures.634.003d Focal seizure types Any focal seizure 77 (78. and one (0.5) 25 (56.6) 23 (42. The father hav- Abnormal EEGs were recorded in more than one-third of ing attended school was significantly protective against both lifetime and active epilepsy cases (Table S1). 54%) than in those with a generalized seizure semiology. N = 54 Inactive epilepsy. and epileptic spasms. maternal schooling (p = 0. and and paternal economic activity (p = 0. attended clinic for an electroencephalographic (EEG) recording.001d Other clinical features Seizure frequencye Daily 17 (17.392 C.219 Self-reported AED use 13 (17. 95% CI = 3.2) 0. 2(4):388–399.1002/epi4.1) <0. apart from yearly.006. N = 98 Active epilepsy. Neurobehavioral and neurological comorbidities were ties. and eyelid myoclonia.8) 0.3) 20 (45.4) 3 (6.324).0) 0.1) 5 (14.197 Generalized other motorc 45 (45. tic regression analysis (Table 3) identified perinatal compli- nomic activity (p = 0. Seizure types are not mutually exclusive. Multivariate logis- (p = 0.048d Focal to bilateral tonic–clonic 5 (5. tonic.731). p = 0.9) 27 (50.001d Focal motor 46 (46.5) 0.55. c Generalized other motor seizures comprise clonic.9) 20 (37. myoclonic.8) 0. 95% CI = 7.990).14). myoclonic–atonic. (OR = 14. Clinical features of seizures in those diagnosed with epilepsy Dominant seizure types Lifetime epilepsy. previous hospital admissions. J.0) 22 (50.8) <0.62).52–60.

n = 471 Epilepsy.8) 9 (9.14–1.16 (0.33 (2.0) 66 (68.98 (0.2) 82 (89.88–2.0) 69 (71.01) — Mother’s marital statusd Never married/single 13 (2.9) 1.3) 12 (12.81) Incompletely constructed 75 (16.8) 1.91) Needs no or minor repairs 297 (63.95 (0.00 Yes 296 (64.14) 1. 2017 doi: 10.9) 52 (53.00 Yes 414 (93.19) Pregnancy Normal 418 (90.9) 79 (81.75 (0.77–43.27 (1. 393 Epilepsy Epidemiology in Kenyan Children Table 3.00 Divorced/separated 48 (10.13–1.4) 1. n = 98a Univariate OR(95% CI)b Multivariate OR (95% CI)c Median (range) age of child.24–4.75–1.97–3.2) 5.24 (0.4) 23 (23.59) 2. Uni.20–6.8) 38 (38.1) 13 (13.1) 1.1) 18 (18.0) 0.11 (1.00 Yes 398 (93.9) 1.70 (1.8) Yes 17 (3. 2(4):388–399.79) 1.81–4.64 (0.65–4.71–2.00 Home 339 (74.64) Married 376 (81.1002/epi4.67) Place of delivery Hospital/clinic 119 (25.8) 1.11)* — Child eats soil No 448 (96.28) Child has previous head injury No 440 (96.00 Yes 215 (45.34) — Father earns cash No 29 (6.00 Yes 459 (98.87–1.9) 82 (84.24–74.96 (0.6) 47 (47.0) 1.00 Yes 373 (80.27 (1.24 (3.5) 78 (86.35) 4.69–2.0) 1.68) Male sex No 243 (51.6) 2.28–10.2) 0.18 (0.0) 2.4) 32 (33.8) 10 (10.5) 1.61–38.59–1.56–1.53)* — Pets in the householdf No 253 (54.98–1.1) 0.62) 0.00 Yes 18 (4.81 (3.17–1.53)* 1.31) 0.01–9.4) 1.95–1.25–1.76 (0.25 (0. mo 24 (6–36) 24 (2–38) 1.6) 84 (86.00 Abnormal 42 (9.7) 10 (10.87 (0.7) 0.6) 12 (13.28 (0.5) 1.3) 1.1) 94 (95.46 (0.58–2.53 (0.97–1.58 (0. yr 27 (15–46) 26 (13–41) 0.15) 0.32) Continued Epilepsia Open.6) 65 (67.5) 9.0) 1.83 (0.73–1.37) 0.21 (0.10–0.00 Yes 72 (15.4) 4.9) 28 (28.34)* Yes 31 (6.4) 1.6) 1.60 (0.04) Widowed 27 (5.6) 61 (62.95–1.27) — Mother earns cash No 88 (19. yr 8 (6–9) 8 (6–9) 1.46 (0.2) 2.25) Family history of seizures No 395 (84.6) 4.28 (0.38) — Father attended school No 29 (6.61–11.and multivariate ORs of risk factors for epilepsy Risk factor Controls.42) 1.9) 1.22–0.6 (1.93) 0.00 Yes 228 (48.1) 53 (54.04) — Child has previous hospital admission(s) No 395 (86.00 Yes 16 (3.45) 5.9) 4 (4.96–8.59) Median (range) number of children ever born 6 (1–12) 6 (1–11) 0.28–1.1) 45 (46.2) — House status Completely dilapidated shack 28 (6) 1 (1.1) 15 (15.88 (1.00 Needs major repairs 67 (14.4) 51 (52.1) 0.3) 90 (92.9) 5.85 (0.25–8.7) 7 (7.99)* — Child eats cassava No 9 (1.3) 4.41) — Perinatal complicationse No 433 (96.3) 1.3) 88 (89.0) 82 (85.01 (0.8) 10 (10.75–3.36 (0.1) 1.36 (0.2) 60 (61.0) 45 (45.0) 14 (14.12069 .92) Median (range) duration of breast-feeding.5) 13 (13.03)* — Father dead No 431 (93.9) 0.03) Median (range) age of mother at birth.53–1.01 (0.6) 1.16) Mother attended school No 165 (35.88 (0.00 Yes 64 (14.

zures likely began in infancy).37–7. Acute analysis.5 third of these seizures were not treated in a hospital.30 The Tanzanian study was conducted acute symptomatic seizures. which factors for epilepsy in several studies from Africa. J.8) 61 (62. c Age and sex were added a priori to the multivariate model irrespective of the significance level.9%) is higher than those in high- of these seizures. the burden of neurodisabilities (including epilepsy) the prevalence of epilepsy and acute symptomatic seizures would be expected to increase with improved childhood sur- in school-aged children.7) 3. p ≤ 0. Table 3.000) (which can result in underestimation). methodological differences may partially explain the varia- bly. income countries (2–5%)29 and in a population-based study talization. and breast-feeding. and cognitive impairments.23 We demon- all comorbidities except cognitive impairments (Table 5). geophagia.0–2. Control Our study included identification of acute symptomatic sei.16 (2. Risk factor Controls. EEG was burden of epilepsy.9) 1.4 (2. a Includes lifetime and active epilepsy. e Includes difficulties in child’s breathing. Continued.2%) had neurobe.1) 1. median age at onset of epilepsy of 3 years.A. confidence interval.20. and snoring. and epilepsy educa- is similar to a study from the U.394 C.31) Childs uses bed net No 36 (7.1 and may Epilepsia Open. 71%). Significant multivariate associations are indicated by bold italics. demonstrated by increased admissions of neonates26 and reduced home deliveries compared to the past decade (76% vs. these two rural this age group. Kind et al.64 (2.21 Notably. Additionally.12069 . and abnormalities were strongly associated with epilepsy. f Includes dogs and cats.00 (0.3) 36 (37.10 is indicated by bold. whereas we did ours in a demographic commonest epilepsy syndromes. they conducted theirs in an absence epilepsy and Lennox–Gastaut epilepsy were the administrative area. More than two-thirds of acute symptomatic seizures were The prevalence of acute symptomatic seizures in these admitted to a hospital. focal seizures were very common (79%). The prevalence may fail to capture epi- abnormal in 39%.000).5 may have reduced perinatal complications. which may have improved identification of epilepsy.1002/epi4.0–4. previous hospitalization. OR was computed using logistic regression (generalized lin- ear model with logit link).9) 1. of malaria may reduce the burden of acute symptomatic sei- zures. n = 98a Univariate OR(95% CI)b Multivariate OR (95% CI)c Child snores >3 nights per week No 316 (68.3) 94 (95.99–1..S.75) 4.25 Improved obstetric services and neonatal care.27 but massive gains in reduction of infectious etiolo- in 1. A number of neurobehavioral in a similar malaria-endemic region such as ours. ADHD.000 and 69 vival. and snoring were independently associated with in Tanzania (2.68–5. crying.31–5.1%). respectively. strated that decline in malaria reduces the incidence of acute Of the 98 children with epilepsy. fur- Discussion ther influencing the current prevalence estimate.65)* — Median (range) birth weight. and 83% were not receiving an AED. and added to the multivariate model.000 of the population.25 and should be addressed. symptomatic seizures. the present study estimates are half symptomatic seizures are associated with the risk of epi- what was reported in a previous study in the same area (41 lepsy in this area.00 Yes 147 (31.25 is indicated by asterisk. previous hospi. n = 471 Epilepsy. dence of the community members to report seizures. settings may have different levels of social desirability bias The prevalence of epilepsy in these children (21 in 1. as would be expected for area under active surveillance. 12 (12. and falciparum malaria caused half school-aged children (6. d The three categories of mother’s marital status are compared to the baseline category of never married/single. For variables with multiple categories. In those with epi. as discharges compatible with Lennox–Gastaut syndrome.24 The widely recognized decline in malaria22 since the previ.17.24 which are associated with epi- havioral comorbidities and generalized EEG epileptiform lepsy. for example. Con- This epidemiological study provides important data on versely.7) 4 (4.4 (1. given that one- in 1.19 but higher than tional interventions in Kilifi31 may have improved the confi- the worldwide prevalence estimated in a recent meta. perhaps explaining the late tions. OR.96 (0. 2017 doi: 10. gies of epilepsy such as falciparum malaria28 may offset the lepsy.1) Yes 431 (92. 2(4):388–399. zures attributable to malaria. b p ≤ 0.01) — Values are n (%) unless otherwise indicated.9) 3. CI. odds ratio. the OR represents all categories combined. Childhood tion in estimates. The lepsy in the children who died before the survey (whose sei- important risk factors for epilepsy were perinatal complica. for is associated with severe falciparum malaria. kg 3. were greater in active epilepsy than in lifetime epilepsy. for which data were collected >10 years ago. nota. Perinatal and neonatal complications are established risk ous study5 may have reduced the burden of epilepsy. Family history of seizures. ASD. placing them at 21 in 1.

6) 1.96–1.79–2.33) 2.5) 1.98) 7.64 (1.5) 268 (90.2) 1.00 Continued Epilepsia Open.80) Child has previous head injury No 440 (96.1) 1.0) 171 (55.00 Yes 17 (3.41–2.45 (1.1) 1.62–1.4) 54 (17.43 (0.94–1.41 (1.41) Male sex No 243 (51.36–3.0) 2.7) 19 (6.1) 0.59 (0.00 Yes 228 (48.9) 0.9) 259 (84. yr 27 (15–46) 26 (15–46) 0.3) 1.00 Yes 72 (15.88–3.05 (1.3) 2.8) 1.78–1.6) 3.0) 27 (9.8) 33 (11.55) Place of delivery Hospital/clinic 119 (25.40–1.84) Median (range) duration of breast-feeding.1) 1.12069 .7) 27 (8.4) 91 (29.0) 1.35)* — Child eats cassava No 9 (1.0) 2.00 Yes 414 (93.30 (0.8) 10 (3.00 Divorced/separated 48 (10.0) 241 (78.05–5.6) 27 (9.1002/epi4.72) 1.95) Needs major repairs 75 (16.41 (0. 2017 doi: 10.48–10. yr 8 (6–9) 8 (6–9) 1.05) — Father dead No 431 (93.36–1.1) 136 (44.2) 265 (88.33 (0.39) 1.71 (0.5) 0.15) Widowed 27 (5.28–4.73 (0.6) 182 (59.00 Yes 215 (45.00 Yes 296 (64.37 (1.07 (0.3) 287 (93.2) 1.2) 1.and multivariate ORs of risk factors for acute symptomatic seizures Acute symptomatic Risk factor Controls.2) 1.33–11.1) 303 (99.04) Median (range) age of mother at birth.12–2.94) — Mother attended school No 165 (35.23 (1.78–3. mo 24 (6–36) 24 (6–41) 1.98 (0.75) Mother employed No 88 (19.9) 65 (21.12 (0.22–0.00 Yes 18 (4.1) 64(21.6) 128 (41. 2(4):388–399.6) 1.04) — Child has previous hospital admission(s) No 395 (86.3) 1.9) 127 (41.31) — House status Needs no or minor repairs 28 (6) 7 (2.00 Yes 459 (98.1) 1.99 (0.9) 3 (1.12)* — Child eats soil No 448 (96.2) 200 (65.92–1.01) — Mother’s marital statusc Never married/single 13 (2.70 (0.84 (0.31–7.97–1.6) 213 (70.9) 1.53–7.01 (0. 395 Epilepsy Epidemiology in Kenyan Children Table 4.9) 0.9) 3. Uni.0) 1.99) 0.09 (1.76 (5.84)* — Perinatal complicationsd No 433 (96.42) — Father attended school No 29 (6. n = 308 Univariate OR(95% CI)a Multivariate OR(95% CI)b Median (range) age of child.35–0.00 Yes 64 (14.3) 1.1) 64 (20.35 (0.96)* — Father employed No 29 (6.50 (1.33–2. n = 471 seizures.5) 1.25) 1.11 (0.00) 1.7) 7.74–4.6) 291 (95.93 (5.00 Yes 398 (93.17 (0.99–2.1) 0.0) 242 (79.99–1.00 Yes 373 (80.00 Home 339 (74.00 Yes 31 (6.0) 1.1) 179 (58.84 (0.35) Family history of seizures No 395 (84.02) 2.23–3.8) 1.0) 274 (91.9) 1.92–1.4) 180 (58.9) 2.0) 1.92–1.4) 1.65–3.90–1.3) 278 (91.9) 241 (79.3) 31 (10.00 Abnormal 42 (9.05 (0.4) 1.83 (0.40–8.34 (1.20)* — Median (range) number of children ever born 6 (1–12) 6 (1–13) 0.8) 106 (34.99 (0.99) Married 376 (81.34) Completely dilapidated shack 297 (63.00 Incompletely constructed 67 (14.38) — Pets in the householde No 253 (54.8) 23 (7.09) Pregnancy Normal 418 (90.1) 46 (15.

47) Childs uses bed net No 36 (7.8) 164 (53.35 Acute symptomatic seizures preceded the onset of tions.57 (10. This rules out a bidirectional relationship (Table S2).1 The association between Focal with impaired awareness seizures.7) 14.5–5. Table 5.39–37. Table 4. Neurobehavioral comorbidities associated with epilepsy Lifetime epilepsy. Adjusted OR.09–520.25–1458.01–255. ASD. b Age and sex were added a priori to the multivariate model irrespective of the significance level.77–1634. e Includes dogs and cats. N = 643 epilepsy (95% CI)a epilepsy (95% CI)b ADHD 28 (28.63 (0. consistent with previous obser- but more studies are needed. Active epilepsy. previously known Epilepsia Open.70–2. risk of epilepsy in Africa. OR was computed using logistic regression (generalized lin- ear model with logit link).9) 1. which.25 and mercury. in particular lead seizures.32 and is probably related to upper airway sents febrile infections that are important causes of acute obstruction causing hypoxia. showed no statistically significant associa. autism spectrum disorder.66 (8. Kind et al. Continued. attention-deficit hyperactivity disorder. in similar propor- investigated.66)c ADHD.96) — Child snores >3 nights per week No 316 (68.3) 141 (46. lifetime Adjusted OR.00) — a p ≤ 0.02)* — Median (range) birth weight. and breast-feeding.51 (3.3) 17 (2.17 Family history of seizures is widely for seizures.2) 7 (13.62)c 37.00 (0.5) 1.3) 14.37) 3.1) 2 (3. n = 471 seizures.89 (0.38 (0.54–28.7) 15 (4. n (%).00 Yes 431 (92.22) ASD 11 (11.396 C. OR. kg 3. crying.50 (1. n = 308 Univariate OR(95% CI)a Multivariate OR(95% CI)b Yes 16 (3.33.7) 3. epilepsy in 8% of children. Previous hospitalization was associated with epilepsy this area. CI.55 (7. which geophagia and neurocysticercosis was not supported by pre- corresponded to the 25th percentile of age at onset of epi.1) 1. c Unadjusted OR computed using the immediate command cci in Stata. the OR represents all categories combined. p ≤ 0. c The three categories of mother’s marital status are compared to the baseline category of never married/single.7) 2 (0.83 (7.08 (2. which was associated with an increased identifiable and preventable underlying causes of epilepsy. p ≤ 0.05 is indicated by bold italics. repre- South Africa.82 (1.07–43.9) 1.1) 3 (5.52–60.38) Motor impairments 5 (5.87–3.01)c Hearing impairments 1 (1.00 Yes 147 (31. zures.7) 0 (0) 13. active Comorbidity n (%).69–790. and should be Focal seizures semiology was common. and suggests that soil in this area acute symptomatic seizures compared with those without may be contaminated with heavy metals.0) 2 (3.4 (2. The association The risk for acute symptomatic seizures was increased by between snoring and epilepsy was reported previously in previous hospitalization. Acute symptomatic Risk factor Controls. and acute symptomatic seizures and epilepsy possible genetic susceptibility to acute symptomatic sei- as risk factors. A multivariate reanalysis of snoring as an out.8) 2.2) 1.0–4. human immunodeficiency virus (HIV).16)c 24.6) 18 (33. b Association of comorbidities with active epilepsy adjusted for age and sex in a logistic regression model.14) 50.9) 0 (0) 6. confidence interval. assuming at least one person with comorbidity in controls. and added to the multivariate model.3) 36.29)c 11.5 (1.25 is indicated by asterisk.76 (2. J.06) 17. N = 54 Controls.0) 2 (0. which may damage the brain. recognized as a risk factor for febrile seizures and suggests come variable. For variables with multiple categories.0) 1 (1. a Association of comorbidities with lifetime epilepsy adjusted for age and sex in a logistic regression model.55 (3. and likely represent Toxocara infection.13–4. N = 98 n (%).1002/epi4.15–934.10 is indicated by bold.90 (0.49) Cognitive impairments 6 (6.99–1.93–1985.69 (1.86–3. An association of epilepsy with vations that the risk for epilepsy is increased in those with eating soil is interesting.14) 6. intellectual disability represents probably because it is a proxy for acute infections that are risk factors for geophagia and should be taken into account known to cause epilepsy such as cerebral malaria and in analyses such as ours.6) 0 (0) 34.97–170.5) 15 (4.44 (0. be associated with seizures in the first year of life. 2(4):388–399. 2017 doi: 10.12069 . vious computed tomography scan and serology studies in lepsy. d Includes difficulties in child’s breathing.3) 292 (95. odds ratio. Geophagia is also strongly associated with tions with previous studies in Africa.34 However.54)c Visual impairments 2 (2. which may lower the threshold symptomatic seizures. as explained above.

identified 8% with possible Lennox–Gastauat syndrome.1002/epi4.39 suggesting improved concordance between the cation in the study. classification that may help to optimize treatment for these which was unavailable in our setting. and the causal role of the investigated absence and Lennox–Gastaut epilepsies. Residual confounding of unmeasured factors dromes.7 suggesting that frequent poral sclerosis in more than one-third of patients with epi. seizures may be responsible. this This study shows that the burden of seizure disorders in could be a reflection of the preponderance of focal seizures this rural area of Kenya is higher than in many high-income in the entire epilepsy sample rather than a bias. which we think is taken together with generalized epileptiform EEG patterns related to febrile status epilepticus. These results are only generalizable in other settings passive eye-closure EEG. was attrition between stages I and II. as they were more frequent recording. The reduction in malaria witnessed in pre- The overall prevalence of neurobehavioral problems of vious studies may have reduced the burden of epilepsy. Comprehensive management this area. although we accounted ties may be related to damage to regions supplied by middle for this in the adjusted prevalence. as even interaction within the family or parenting processes. We have observed unilateral and bilateral mesial tem. The ties are likely to be more complex. and might have pro.36 The underlying symp. assessment of both epilepsy and acute symptomatic seizures suggesting identifiable underlying symptomatic causes. comes. Most epilepsy was associated strongly associated with ADHD (OR = 15). particularly in those with active epilepsy. The ensured that these two conditions were well differentiated. or clonic. often presented together. atonic. sleep EEG studies. Neurobehavioral problems lepsy using magnetic resonance imaging. sleep-induced and study.7 Therefore. ited identification of epilepsy syndromes.5 but comparable to African esti.38 and epilepsy in children life of the child than seizures. and telemetry may have lim- in perinatal complication or nontraumatic acute encephalo. Compared snoring as risk factors in this and other recent studies war- to community controls. simi. The associations with neurobehavioral comorbidi- vided insights into treatment and prognosis of epilepsy. which disorders in this area.12 Focal abnormalities were particularly common. This study did not pathies. or genetic paroxysms. comorbidities should be assessed in children with epilepsy.41 in epilepsy syndromes of older children. HIV status may have confounded the neurobehavioral out- Lennox–Gastaut syndrome.40 which rules out benign epilepsy with occipital perform additional neuroimaging. for example. cognitive impairments (OR = 15). country settings. Systematic mates. All tools are standardized tures was 13%. Adherence to treatment may control seizure limited settings. it is still needs of children with epilepsy. The was informed by a thorough search of the literature. and motor an indication that temporal lobe epilepsy is common in this deficits (OR = 35).2 Use of AEDs It is thought that neurobehavioral comorbidities of epi- and adherence in children is decided by parents. adherence to treatment in childhood epilepsy explains the through training of clinicians and development of locally high rates of status epilepticus. proportion of generalized seizure semiology with focal fea. the study. Focal seizures with similar risk factors.5 Although the treatment comorbidities should be set up in low and middle income gap across all age groups has been substantially reduced in countries to ensure comprehensive management of the this area31. The non. The emergence of eating soil and may have overlapped with behavioral problems. and myoclonic. 83% of children were not receiving AEDs. 397 Epilepsy Epidemiology in Kenyan Children as complex partial seizures. ASD with focal features. such as tonic. A detailed investigation of risk factors EEG and reported seizure types in the present study. may be helpful. were more common in those who had an EEG recorded. avoiding overestimation of either. stage II after positive screening in stage I. but this was only helpful in identifying childhood cannot be ruled out. The associations were greatest in active area. There characteristic posterior parietal–occipital EEG abnormali. Peri- 54% in children with epilepsy is slightly higher than for pre. investigations. significantly greater in children than adults. epilepsy. household environment. a tomatic causes can be identified through neuroimaging. 2017 doi: 10. This is perhaps (OR = 37).39 Daily seizure frequency was commonly programs for children with epilepsy that address neurobe- reported. but virology tests were not routinely performed in Overall. natal complications and previous hospital admissions with vious similar studies of behavioral problems in the area febrile infections remain important risk factors for seizure (49%)7. 2(4):388–399. were common. implying that risk factors was difficult to establish in a cross-sectional future detailed EEG studies. This is a population-based study in which detailed clini- EEG was abnormal in 39%. Health services for assessing and addressing these lar to previous estimates (89%). Lack of prolonged EEG and posterior cerebral arteries. as would havioral comorbidity should be developed in resource- be expected. this study did not include ASD and ADHD. We hoped to use EEG to identify epilepsy syn. who choose lepsy may have more detrimental impact on the quality of treatment options for children. pointing to identifiable symptomatic Epilepsia Open.12069 . frequency and perhaps improve outcome. tonic– networks of the maturing brain. higher than that reported in a cal and neuropsychological evaluations were performed in previous study (20%). perhaps related to the commonest generalized seizure types. children. neurobehavioral may go untreated if mistaken for febrile seizures.37 following educational interventions. metabolic. for example. epilepsy was several times more rants further investigation. as in a previous study. less than the previous rates of 29% in all age and underwent adaptation and validation before their appli- groups. similar in previous studies in appropriate assessment tools.

15:180. Storz C. Epilepsia clinical characteristics of febrile seizures in Tanzania. Psychopathology in chil- 15. BMC Pediatr 2008. Prevalence and risk factors for zanian children with epilepsy: a community-based case–control study.1:112–120. J Child Neurol 2007. Epilepsy in rural South impairment in Kenya. Wagner RG. Fisher RS. Camfield P. et al. Abubakar A. Matuja W. Kariuki SM. J. Rodenburg R. Dev Med Child Neurol 2012. Kariuki SM. Ikumi M. 2017. 2017 doi: 10.17:396–404. risk factors. Hunter E. BMC Neurol 2015. epilepsy.53:1135–1142. Epilepsy Behav 2012.39:643–650. aetiology and out- come of acute seizures in children admitted to a rural Kenyan district renewed efforts in reducing the treatment gap in children to hospital. Lancet 3. Idro R. et al. Exposure to multiple para- Afr Med J 2000. 123. et al. Lancet Psychiatry 2017. 2(4):388–399. Mageto J. Gwer S. Camfield C. Ojal J. et al. Neurology 2015. Begley CE. Kariuki SM. et al. Operational classification of sei. et al. tures of convulsive epilepsy in Africa: a multicentre study of preva. White S.14:133–152.89:751–756. 39. Kariuki SM. Lan- 11.52:2–26. Traditional healers and lence. et al.55:76–85. J Pediatr Psychol 2005. zures in African children with falciparum malaria. Disclosure 23. Khan NZ.82:582–589. Kleinschmidt I. Mung’ala-Odera V. (083744) and S. convulsive epilepsy in rural Kenya: causes and associated factors. Wiebe S. Kariuki SM. 2015. Kariuki SM. Epilepsia Epilepsia 2014. et al.7:145–150. associated disability and management. et al. et al. Seaton C. More than 80% were not receiving AEDs.55:344–352.398 C. Incidence. Shilpi AB. Chengo E. Electroencephalographic fea. Incidence and outcome of 657. Prevalence and incidence of epi- clinicians for taking care of the sick children.6:151. et al.54:1072. Ibinda F. Ngugi AK.4:136–145. causes. Epilepsia 1973. et al. comprehensive clinical services. Moisi JC.30: studies and surveillance of epilepsy. et al.22:26–32. Manga P. O’Meara WP. The decline in paediatric malaria admissions on the coast of Kenya. et al.127: epilepsy treatment on the Kenyan coast. Standards for epidemiologic dren with epilepsy: a meta-analysis. 40. Kathomi C. Gikandi PW.90:262–266. Waruiru C. Crawley J. Holding PA. parum malaria in an endemic area on the Kenyan coast. Kenya. Behavioral problems in 33. Penry JK.134:1519–1528. Child Care Health Dev 2013. Prevalence and and clinical features of cerebral malaria. 32. Burden of epilepsy in rural 30.41:650– 36. 5. et al. prevalence and aetiology of The Wellcome Trust supported C. Kendall-Taylor N. Rogathe J. et al. Ngugi AK. Epilepsia Open 2016. Kariuki SM. 22. Bejon P. incidence 31. and sub-Saharan Africa. Kakooza-Mwesige A. of people with epilepsy admitted to a rural hospital in Kenya. Clinical features. et al. Bauni E. tomography scan of the brain in a community study of neurological 8. Agresti A. Bertram MY. Kenya: a cross-sectional study.58:522–530.R. Mbuba CK. Risk factors associated with the Bangladesh. Kahindi M. Twine R.49: 1099–1107. Mungala-Odera V.84:1838–1845. Premature mortality in active 2008. et al. et al. Meijer AM. BMC Public Health 2010.8:5. proximate zure types by the International League against Epilepsy: position paper causes. This article was published lepsy: a systematic review and meta-analysis of international studies.8:e2908. Muslima H.56:577–584. Effect of a fall in malaria 2015. The genetic risk of acute sei- dren: a population-based study. Smith S. 7. Malar J 2007. 2004.1002/epi4. 19.12069 . et al. 1638–1639.108:782–791.12: 19 years. Lancet 12. Fegan G. Christianson AL.54: 10. convulsive status epilepticus in Kenyan children: a cohort study. et al.K. Burden. African children–prevalence. Coull BA. transmission on morbidity and mortality in Kilifi. Mbuba CK. Chong WK. Incidence. Zwane ME. et al.55:1626–1633.84: factors associated with convulsive status epilepticus in Africans with 247–253. Abubakar A. Epilepsia 2013. Prevalence of epilepsy in chil- Acknowledgments dren. Mwaniki MK. Hay SI. Pediatr Res 2014. Ibinda F. Approximate is better than “exact” for inter- val estimation of binomial proportions. PLoS Negl Trop Dis 2014. Increased prevalence of epi- lepsy associated with severe falciparum malaria in children. Validation of a home-based cet Neurol 2008. Lancet Neurol 2013. 1. Wagner RG. causes and pheno- havioral comorbidities that should be addressed through types of acute seizures in Kenyan children post the malaria-decline per- iod. causes. The authors would like to thank the par. et al. Prevalence of active 26.11:688–696. Meindl M. Clark TG. active convulsive epilepsy in rural northeast South Africa. Burton K. read the Journal’s position on issues involved in ethical publication and 24. Newton CRJC.C. Bottomley C. Sadarangani M.and middle- 2. Int J Epidemiol 2012. Seizure education programme: a randomized controlled trial. Computerized children with epilepsy in rural Kenya. 20. Beghi E. graphic Surveillance System (KHDSS). Mturi FN. Epilepsy Dev Med Child Neurol 2011. 35. Mwangi TW.M. Bistervels IM. Serem GK. et al. Scott JA. Prevalence. improve outcome.10:591. et al.77:591–596. Matuja W. 2014. Res 2014. Epilepsia 28. White S. Cross JH. et al.45:978–981. Evaluation of Kilifi epilepsy and risk factors of epilepsy in older children in rural Kenya. et al. Wagner RG. neurodevelopmental screening tool for under 2-year-old children in 37. Epilepsia Open. Epilepsia 2008. Kariuki SM. sites is associated with the prevalence of active convulsive epilepsy in 9. Brain 2011. Newton CR. Appleton R. et al. Sauro KM. seizures and epilepsy in children. et al.372:1555–1562. Kariuki SM. Makrush RE.88:296–303. Epilepsia 2008. et al. Rose SW. et al. Bottomley C. 453–468. Arch Dis Child Neurology 2014. Neurodevelopmental disorders in low. of the ILAE commission for classification and terminology. Epilepsy was associated with neurobe. Scott JA. Epilepsia 2011. Stams GJ. 990–1001. Am Stat 1998. Njuguna PW. Kariuki SM. White S. Electroencephalographic 14. Kombe M. 27. The incidence. 4. Behavioural comorbidity in Tan. Arch Dis Child 2001. 17. with permission of the director of KEMRI. Chengo E. Febrile seizures: an update. S 34. We confirm that we have 2004. Kamuyu G. Ngugi AK. et al. Thurman DJ. 253–263. Okiro EA. An increase in the burden convulsive epilepsy in sub-Saharan Africa and associated risk factors: of neonatal admissions to a rural district hospital in Kenya over cross-sectional and case-control studies. Clin Neurophysiol 2016. Neurol 2012. Epilepsia The authors declare no conflicts of interest.17:117– (099782/Z/12/Z) during the study. Gatakaa HW. 6. et al. 25. and outcomes income countries. Burden. epilepsy treatment gap in Kilifi. Rockett K. Neville BGR. Fiest KM. Newton CR. and consequences of active convulsive epilepsy in Africa. Meehan R.N. and the nurse and 21. Neurology 2017. Epileptic Disord 2015. Rimba K. 18. Muthinji P. pattern and associated factors. 38. ents of children in this study for agreeing to take part. Kind et al.23:41–46. Community-based prevalence and Kenya measured in disability-adjusted life years.J. Akpalu A. et al. Profile: the Kilifi Health and Demo.52:119–126. 13. Fegan G. comorbidities of behavioural and emotional problems in Kenyan chil. 41. Carter JA. Ibinda F. Bottomley C. French JA. Kariuki SM. urging 16. 29. Acute seizures attributable to falci- affirm that this report is consistent with those guidelines. Risk of convulsive epilepsy References following acute seizures in Kenyan children. et al. Ngugi AK.

Electroencephalographic features in those childhood absence epilepsy. Additional Supporting Information may be found in the Figure S2. Epilepsia Open. Examining whether acute seizures or epilepsy are a risk factor for snoring. Electroencephalogram of a 7-year-old female online version of this article: shows a clear 3-Hz spike and slow wave pattern typical of Table S1. Raw prevalence of epilepsy and acute seizures Supporting Information in 2015.1002/epi4.12069 . Members of the NDD study group. Appendix S1. 2017 doi: 10. 399 Epilepsy Epidemiology in Kenyan Children Figure S1. diagnosed with epilepsy. 2(4):388–399. Table S2.