J Neurol Neurosurg Psychiatry 83:1139-1144 doi:10.1136/jnnp-2012-303042
  • Cognitive neurology
  • Research paper

Cognitive and mood effects of phenobarbital treatment in people with epilepsy in rural China: a prospective study

  1. Josemir W Sander8,11
  1. 1Institute of Neurology, Huashan Hospital, Fudan University, WHO Collaborating Centre for Research and Training in Neurosciences, Shanghai, China
  2. 2Ningxia Medical University affiliated Hospital, Ningxia, China
  3. 3Huaxi Hospital, Sichuan Province, China
  4. 4No 1 Hospital, Jilin University, Jilin Province, China
  5. 5Centre of Disease Control, Anhui Province, China
  6. 6Centre of Disease Control, Hebei Province, China
  7. 7Beijing Neurosurgical Institute, Beijing, China
  8. 8Department of Clinical and Experimental Epilepsy, UCL Institute of Neurology, London, UK
  9. 9Chinese University of Hong Kong, Hong Kong, China
  10. 10Departments of Medicine and Neurology, The University of Melbourne, Royal Melbourne Hospital, Melbourne, Australia
  11. 11SEIN – Epilepsy Institute in the Netherlands Foundation & WHO Collaborating Centre for Research, Training and Treatment of Epilepsy, Heemstede, 2103SW, The Netherlands
  12. 12China Association Against Epilepsy, Beijing, China
  1. Correspondence to Professor Ley Sander, Department of Clinical and Experimental Epilepsy, UCL Institute of Neurology, Queen Square, London WC1N 3BG, UK; l.sander{at}
  1. Contributors This work was conceptualised by JWS, ZH, WW and DD, and all authors approved the protocol. Data collection was done by QZ, DZ, WL, QW, JS, QZ and PY. Analysis was undertaken by DD, ZH, GSB and JWS. The first draft was prepared by DD, ZH, PK, PY and JWS and all authors contributed to and approved the final draft. WW, ZH, HMdB, SL and JWS organised the funding. ZH and JWS are the guarantors.

  • Received 21 April 2012
  • Revised 25 June 2012
  • Accepted 29 June 2012
  • Published Online First 31 July 2012


Background Phenobarbital is an effective treatment for epilepsy but concerns remain over its potential neurocognitive toxicity. This prospective study evaluated the effects of phenobarbital treatment on cognition and mood in people with epilepsy in rural China.

Methods We recruited 144 adults with convulsive seizures and 144 healthy controls from six sites in rural China. People with epilepsy were treated with phenobarbital monotherapy for 12 months. At baseline, and at 3, 6 and 12 months, cases and controls were evaluated with a battery of neuropsychological tests: the Mini-Mental State Examination, the Hamilton Depression Rating Scale, a digit span test, a verbal fluency test, an auditory verbal learning test and a digit cancellation test. Efficacy of phenobarbital treatment was evaluated at the end of follow-up for those with epilepsy.

Results Cognitive test scores and mood ratings were available for 136 (94%) people with epilepsy and 137 (95%) controls at the 12 month follow-up. Both groups showed slightly improved performance on a number of neuropsychological measures. The people with epilepsy showed greater performance gains (p=0.012) in verbal fluency. Nine people with epilepsy complained of memory problems during the treatment period.

Conclusion In this study, phenobarbital was not found to have a major negative impact on cognitive function of people with convulsive seizures and some cognitive gains were observed, possibly due to improved seizure control.


Up to 1% of people in the world have epilepsy, of whom over 80% live in low and middle income countries (LAMIC).1 It is estimated that 9 000 000 people in China have epilepsy, of whom up to two-thirds living in rural areas do not receive regular treatment; a similar large ‘treatment gap’ is present in many resource poor settings.2

From the public health perspective, use of drugs that are effective, affordable and simple to use at the primary care level is of paramount importance in reducing this treatment gap. As such, phenobarbital is recommended by the WHO as a firstline drug for focal and generalised tonic–clonic seizures in LAMIC.3 Consistent with experience in other resource poor settings, an observational study in rural China demonstrated that phenobarbital was efficacious and well tolerated when used at the primary healthcare level.4 The use of phenobarbital as a firstline agent, however, continues to raise concern; anecdotal evidence from resource poor countries suggests that people who have been taking phenobarbital for many years with no apparent side effects may gain marked cognitive improvement on stopping it.5 We performed a prospective study to evaluate the neurocognitive effects of phenobarbital treatment in people with epilepsy in rural China.


Study setting

Five sites, located in rural areas of Ningxia, Sichuan, Jilin, Anhui and Hebei provinces in China, were selected (figure 1). These areas were chosen to represent the different geographical areas and different economic levels of China.

Figure 1

Map showing the five study sites in China. 1, Ningxia; 2, Sichuan; 3, Jilin; 4, Anhui; and 5, Hebei.

Training of village doctors

Prior to study initiation, rural physicians at each study site underwent standard training for subject recruitment, diagnosis and treatment of seizures, administration of neuropsychological tests and follow-up.

Study subjects

Participants were consecutively recruited from 77 village health centres in the five provinces by trained rural physicians. At either the town hospital or the individual's home, local neurologists reviewed the medical history and performed physical examinations in those potentially suitable as cases to confirm the diagnosis of epilepsy. The number of cases in the area during the study period determined the sample size. People were enrolled as cases if they consented to participate and met the following inclusion criteria: (1) at least 16 years old (as the neuropsychological battery had been previously validated for use in a Chinese population aged over 16 years); (2) presenting with primary or secondary generalised tonic–clonic seizures with the diagnosis confirmed by a neurologist; (3) never treated, or inadequately treated, with a non-standard antiepileptic treatment protocol, or non-adherent to antiepileptic drugs (AEDs), or taking Chinese herbal medicine; (4) inadequately controlled seizures (at least one seizure in the previous year); (5) Mini-Mental State Examination (MMSE) score >226 and Hamilton Depression Rating Scale (HDRS) <7.7 People were excluded if they had: (1) a progressive neurological condition; (2) only non-convulsive seizures; (3) seizures well controlled by standard antiepileptic treatment; (4) cardiac, hepatic or renal disorders, or severe hypertension; (5) a history of status epilepticus alone; or (6) an active psychiatric condition.

When cognitive tests are performed over time, a learning effect may cause the scores to improve in the absence of any other change in treatment. It is unethical to deprive a group of people with convulsive epilepsy of effective treatment, so we included a control group without epilepsy or any other neurological or chronic disease and without phenobarbital treatment to try to balance out any learning effect. These healthy volunteers were matched with the people with epilepsy for age (±2 years), sex, education level and the village in which they lived. They all had MMSE scores >226 and HDRS scores <7.7

Phenobarbital treatment schedule

People with epilepsy entering the study were treated with phenobarbital monotherapy. The starting dose was at least 30 mg, taken once daily at night. For each individual, dosage was adjusted to a maintenance dose as clinically dictated based on seizure control and tolerability. The usual daily maintenance dose was 60–180 mg.8

Evaluation of the efficacy of phenobarbital treatment

The efficacy of phenobarbital was evaluated after 12 months of treatment for the group of people with epilepsy. Those defined as seizure free were free of seizures for the entire 12 months of follow-up.

Cognitive assessment

The cognitive tests were selected to measure functions that have been shown to be sensitive to adverse drug effects.9 The neuropsychological battery comprised: (1) the MMSE (a screening test for global cognition); (2) the HDRS (rating the severity of depressive symptoms, and used to monitor mood changes); (3) a digit span test (DST—providing a measure of attentional capacity. The subject repeats digit strings of increasing length, first in the same order and then in the reverse order.10 The performance indicator was the total number of digits recited (forwards + backwards)); (4) a verbal fluency test (VFT—employed as a measure of working and semantic memory. The subject was asked to name as many items as possible relating to specific categories (animals or home items) in 1 min.11 The total number of items for each category was summed to provide the performance indicator); (5) a digit cancellation task (DCT—used as a measure of mental speed. Subjects were asked to cross out a target digit from a set of numbers. The score used was the total number of correct cancellations per minute12); and (6) an auditory verbal learning test (AVLT—used as a measure of episodic memory. The subject was presented with 12 words to learn over five trials.13 The score was the sum of words correctly recalled in the learning, delayed recall and recognition phase). All tests were conducted in Chinese and the test battery had been validated in the Chinese language.14–16

Study procedure and schedule

At the baseline interview, demographic data, such as age, sex, height, weight, body mass index and education were collected for all participants, and clinical characteristics (onset age, duration, number of seizures in the previous 12 months and medical history) were recorded for people with epilepsy.

Rural physicians followed-up study subjects at 3, 6 and 12 months. At each follow-up visit, physicians counted the number of tablets remaining for each person with epilepsy; people were defined as having ‘good adherence’ if they appeared to have taken at least 80% of the number of tablets predicted. Seizure numbers were checked and the phenobarbital dosage was adjusted if necessary. The neuropsychological battery was administered to both groups at baseline and at 3, 6 and 12 months by rural physicians who were aware of group assignment. The control group was assessed at the same time as the people with epilepsy. The village doctors were assessed for their competence in test administration.

Statistical analysis

Higher scores on the MMSE, AVLT, DST, VFT and DCT indicate better function, while a higher score on the HDRS indicates lower mood.

Comparisons for continuous variables were evaluated with the Student's t test, or the Mann–Whitney U test when the data were not normally distributed. The general linear model was used for the repeated measures analysis of neuropsychological tests at baseline and at each follow-up to determine whether scores changed with ‘time’ and to examine the interaction of ‘time and group’.17 All p values were estimated in a two tailed fashion. Differences were considered to be statistically significant at p<0.05. Data were analysed using SPSS V.13.0 (SPSS Inc). CIs for proportions were calculated using CI analysis software (version 2.1.2, Southampton, UK).


Characteristics of study participants

We recruited 144 people with epilepsy and 144 matched controls (47 pairs of subjects from Ningxia, 23 pairs from Sichuan, 28 pairs from Jilin, 27 pairs from Anhui and 19 pairs from Hebei provinces). In each group, 93 (65%) were men. The average age of participants was 36 years and the average time in education was 7.4 years. There was no statistically significant difference in body mass index between the case and control groups (mean 22.2 (SD 2.7) vs mean 22.7 (SD 2.7); p=0.104).

Most (103, 72%) people with epilepsy were taking non-standard antiepileptic treatment, particularly traditional Chinese medicine. Clinical details of epilepsy factors and of risk factors for epilepsy are shown in table 1.

Table 1

Demographic data and clinical characteristics of people with epilepsy (n=144)


One hundred and thirty-six (94%) people with epilepsy and 137 (95%) controls completed the study and underwent a cognitive assessment at 12 months. Figure 2 provides participant data available for analysis at different time points.

Figure 2

Flow diagram of participant data available for analysis at different time points.

Phenobarbital dosage, adherence and efficacy

The median phenobarbital dosage used by people with epilepsy was 60 mg (range 30–240 mg, mean 75 mg, SD 30 mg) per day during the first 3 months, increasing to 90 mg (range 30–210 mg, mean 88.2 mg, SD 36.8 mg) per day by the end of follow-up. Almost all (98%) people with epilepsy were evaluated as having ‘good adherence’. Compared with pretreatment baseline, seizure frequency was reduced by at least 50% during the 1 year study period in 78% (106/136) of people with epilepsy, and 40% (55/136) were seizure free for the full year. Twenty-five (45%) of those who became seizure free had had 1–3 seizures in the previous (baseline) year, and the other 30 (55%) had had eight or more seizures in the baseline year.

Adverse events

Adverse events, including fatigue, tiredness, headache, dizziness and vomiting, were recorded by 32 (23%) people with epilepsy at 3 months and by 25 (18%) at 6 months and at 12 months. Complaints of memory impairment or decline were reported by nine people with epilepsy during the study period.

Assessment of cognitive function

Means and CIs of scores for each test at each follow-up period are given in table 2. For most tests, the changes in scores over time were small, and ran in parallel between the two groups. In people with epilepsy, the MMSE, VFT and DCT scores increased significantly while DST score decreased. There was no significant change in HDRS or AVLT scores. For controls, MMSE, AVLT and DCT scores improved significantly and mood ratings on the HDRS declined while there was no significant change in DST and VFT scores. The only significant difference in change in scores (time–group interaction) between the two groups was on the measure of VFT, with the epilepsy group showing slightly greater gains (p=0.012).

Table 2

Means (95% CI) for neuropsychological scores over follow-up times by study group

An exploratory subgroup analysis stratified for phenobarbital efficacy was carried out for the VFT score over time (table 3). The VFT scores increased from baseline in all groups (those with <75% seizure reduction, those with 75 to <100% seizure reduction and those seizure free) but there was no significant difference in the changes (time–group interaction) among the three groups (p=0.399).

Table 3

Means (95% CI) for verbal fluency test scores over follow-up times by phenobarbital efficacy in people with epilepsy


Phenobarbital treatment was not observed to have any adverse impact on cognitive test performance or mood ratings of adults with convulsive seizures in this multicentre, case control, parallel, follow-up design study, although there was a small decrease on a measure of attention. Both groups showed small but positive changes over time on a number of measures, with gains in verbal fluency being greater for the epilepsy group taking phenobarbital. Most people taking phenobarbital had improved seizure control and it is possible that the cognitive gains brought about by this were greater than the impairments potentially caused by the treatment itself. Subjective complaints of memory problems were made by only 6% of the study participants.

On a global scale, phenobarbital remains one of the most important AEDs. It has many advantages, including ease of use, affordable cost (less than US$5/year), once daily dosing, broad spectrum of action and comparable efficacy with other established AEDs.18 ,19 Its drawbacks include its potential for teratogenicity; children born to women with epilepsy are at increased risk of somatic malformations and behavioural impairments.20 Phenobarbital also has great interaction potential. It is an enzyme inducer and can be the target or the effector in interactions involving AEDs and other lipid soluble drugs19; this may be a problem with both Western and traditional medicines. The other major problem, however, is its potential adverse effects on cognition, mood and behaviour. Reports of this are conflicting, and most data are derived from paediatric populations. In a double blind, counterbalanced, crossover study, 21 children were treated with phenobarbital and valproic acid, each for 6 months. Children performed significantly less well on the neuropsychological tests, and had worse behaviour while receiving phenobarbital than when receiving valproic acid.21 A randomised controlled study observed 217 children aged between 8 and 36 months who had had at least one febrile seizure and were at increased risk of further seizures. After 2 years, mean IQ was 7 points lower in the group assigned to phenobarbital than in the placebo group. Six months later, after the medication had been discontinued, mean IQ was 4 points lower in the phenobarbital group.22 In another study, 73 children were randomly assigned to receive phenobarbital, carbamazepine or valproate. No significant differences were found in IQ or scores on the Bender Gestalt test between groups after 6 and 12 months of treatment, but auditory event related potentials (P300) were prolonged in the phenobarbital group compared with the others.23 In a randomised controlled trial in childhood epilepsy that used a validated behaviour rating scale, no significant differences were found between children taking phenobarbital and those taking phenytoin.24 Some studies found reversible IQ, P300, attentional and memory deficits in children after discontinuation of phenobarbital.25–27

The situation is less clear in the elderly; there are few data available, but it is known that the elderly may be sensitive to the CNS effects of drugs.28 A prospective, randomised, parallel study observed 95 people with seizures and Alzheimer's disease taking levetiracetam (n=38), phenobarbital (n=28) and lamotrigine (n=29), and a control group (n=68). Phenobarbital was found to have persistent negative cognitive side effects during the 12 month evaluation period.29

A study involving healthy adult volunteers reported significantly slowed decision making speed associated with low blood levels of phenobarbital compared with mid-range blood levels of phenytoin or valproate. No differences were observed between the groups for working or episodic memory.30 In another study, 27 seizure free patients were assessed before and at intervals following the withdrawal of phenobarbital using tests of intelligence, vigilance, attention, memory and visuomotor performance. The patient group performed less well than controls on tests of visuomotor performance and immediate spatial memory. No differences were found following phenobarbital dose reductions and up to 1 year after complete discontinuation.31

Studies in LAMIC in general have not shown neuropsychological toxicity of phenobarbital compared with alternative AEDs.4 ,32–34 These studies may have underestimated the cognitive effects of phenobarbital, as neuropsychological testing was not undertaken. This study is one of the first to use a neuropsychological battery to assess prospectively the cognitive effect of phenobarbital monotherapy in adults with epilepsy. Our results support a lack of major adverse effects on cognitive functions as well as good efficacy of the drug.

There is currently no standard neuropsychological battery to monitor the cognitive effects of AEDs. Some drug sensitive cognitive tests have been proposed.35 A recent study assessed cognitive function of 147 people with newly diagnosed epilepsy before they started treatment (with carbamazepine, gabapentin, lamotrigine, oxcarbazepine, topiramate, valproate) and after 12 months. After controlling for statistically confounding factors, epilepsy had a different cognitive trajectory compared with healthy volunteers from the general population. Memory, psychomotor speed and higher executive functioning were the domains most vulnerable to change over a 12 month period.36 On the basis of past research and clinical experience, we selected neuropsychological tests assessing key cognitive domains and mood that would be expected to be maximally sensitive to AED effects, drawing on measures that had been validated in an adult Chinese population. The battery took 30 min to complete and was easily conducted by local health workers. A more extensive battery runs the risk of reduced engagement and performance declines over time. We consider our battery a good tool to monitor the cognitive effects of AEDs in adults with epilepsy in rural areas of China.

Studies in LAMIC are frequently observational and hence may have low follow-up rates due to loss to follow-up or death, and may lack both systematically collected data on adverse effects and formal testing for cognitive side effects.19 The current study has significant advantages of a high follow-up rate (over 94%) for collection of data on phenobarbital efficacy and side effects, and formal neuropsychological assessment.

The study has limitations. First, people with epilepsy enrolled in the study had mild to moderate epilepsy (average seizure frequency 15/year) and were prescribed relatively low phenobarbital dosages (median dosage 60–90 mg/day). Thus phenobarbital effects on the cognitive function of people with more severe epilepsy prescribed higher dosages were not addressed. It is also difficult to evaluate the potential effect of phenobarbital dose on the cognitive measures without blood levels. Second, at baseline, most individuals were taking non-standard antiepileptic treatment, particularly Chinese herbal medicine. The switch from these could have obscured the effects of phenobarbital, although the cognitive side effects of Chinese herbal medicine are currently not fully known. Third, follow-up was only 1 year, and it is not possible to comment on longer term effects on cognitive function and mood. Depression in particular may be a consequence of long term administration. Fourth, the neuropsychological measures were chosen for their likely sensitivity to adverse drug effects but we cannot rule out that failure to detect negative cognitive changes is due to the tests selected. For example, the cut-off value for the MMSE might be considered low and to reflect a degree of pre-existing cognitive impairment with less possibility for further decline or to demonstrate practice effects. Furthermore, inclusion of a subjective side effects rating scale would have provided a fuller assessment of the impact of phenobarbital. Finally, the sample size, while large for such a study in a resource poor setting, may have been too small to identify significant group–time interactions between the two groups. Further studies with a larger sample size are needed to evaluate the long term cognitive effect of phenobarbital in people taking a higher dosage, and in different age groups.

Phenobarbital was not found to have a major negative impact on cognitive function and thus vindicates its place in the WHO essential list as a treatment option for epilepsy, particularly in LAMIC. Indeed, in view of its cost effectiveness, phenobarbital has retained a unique position in the therapeutic armamentarium and can be widely used for control of epilepsy in resource poor settings.19 ,37


We thank all of the participants for their cooperation. We also wish to thank the local health workers in the different study sites of China for their help and support.


  • Funding The project was funded by the Chinese Ministry of Health and Stichting Epilepsie Instellingen Nederland. The sponsors were not directly involved in the study design, in the collection, analysis and interpretation of data, in the writing of the report or in the decision to submit the paper for publication. GSB, PJT and JWS are based at University College London Hospitals/University College London, which receives a proportion of funding from the Department of Health's National Institute for Health Research Biomedical Research Centres funding scheme. JWS is supported by the Dr Marvin Weil Epilepsy Research Fund.

  • Competing interests DZ serves on the editorial advisory boards of Neural Regeneration Research and Chinese Journal of Neurology; has received research support from the Chinese Ministry of Health, the Chinese Medical Association and the China Association Against Epilepsy; and has been a commissioner of the neurology branch of the Chinese Medical Association. GSB's husband works for, and has shares in, GlaxoSmithKline. PK serves on the scientific advisory board of Pfizer and receives research grants from Eisai, Johnson & Johnson, Pfizer and UCB Pharma. ZH serves on the editorial advisory boards of the Chinese Journal of Neurology; has received research support from the Chinese Ministry of health and the China Association Against Epilepsy; and has been a commissioner of the neurology branch of the Chinese Medical Association. JWS served on scientific advisory boards for GlaxoSmithKline, Eisai and UCB; has received funding for travel from UCB and Janssen; serves on the editorial boards of Lancet Neurology and Epilepsia; serves on the speaker's bureaus of UCB and GlaxosmithKline; and has received research support from UCB, Eisai, the NIH, the European Union Seventh Framework Programme, the Wellcome Trust, WHO, the National Epilepsy Funds of the Netherlands and the Epilepsy Society.

  • Ethics approval The study was approved by the medical ethics committee of Huashan Hospital, Fudan University, Shanghai (process No 2008-005).

  • Provenance and peer review Not commissioned; externally peer reviewed.


Visit the full archive of podcasts for JNNP here >>

Free sample
This recent issue is free to all users to allow everyone the opportunity to see the full scope and typical content of JNNP.
View free sample issue >>

Don't forget to sign up for content alerts so you keep up to date with all the articles as they are published.