After 100 years of constant use, phenobarbital is still the most widely prescribed antiepileptic drug (AED) in the developing world (Kwan & Brodie, 2004) and remains a cost-effective choice for people with epilepsy in some industrialized countries (Savica et al., 2007). In others, however, its use is declining because of the perception of substantial tolerability issues, particularly in children and adults (Nicholas et al., 2012). This contribution focuses on articles published in the English language since January 1, 2000. We have emphasized clinical studies exploring its efficacy for adult and childhood epilepsies and neonatal seizures, and its use in status epilepticus and for other potential indications outside epilepsy. In addition, we have reviewed studies that describe its side-effect profile and teratogenicity. The last section considers the likely role of phenobarbital into the next century of its clinical use.
This article reviews the current position of phenobarbital using articles published since 2000 and speculates on its likely future contribution to epilepsy care. Over the last decade there have been no major double-blind randomized placebo-controlled or comparative trials with phenobarbital. Previous studies have suggested that phenobarbital is as effective in monotherapy as phenytoin and carbamazepine. Several observational studies undertaken in developing countries over the last decade have confirmed its efficacy and safety for the common epilepsies. This was particularly so in the substantial demonstration project undertaken in rural China under the auspices of the World Health Organization in partnership with the International League Against Epilepsy and International Bureau for Epilepsy. Phenobarbital is still widely used for neonatal and childhood seizures and for drug-resistant convulsive and nonconvulsive status epilepticus. Recent data have confirmed in a prospective cohort of women taking phenobarbital as monotherapy that the drug can be associated with a range of congenital defects in exposed infants. Much effort has gone into exploring the apparent contradiction of higher withdrawal rates due to cognitive and behavioral side effects in studies undertaken in developed countries but not in those sited in the developing world. A raft of data over the last 10 years, including a systematic review, showed no important differences between the tolerability of phenobarbital compared to that with other antiepileptic drugs. Finally, cognitive test scores and mood ratings in 136 people with epilepsy receiving phenobarbital for a year were similar to those in 137 age-, sex-, and education-matched controls in a number of Chinese villages. Indeed, there were some cognitive gains in the patients possibly due to improved seizure control. Phenobarbital is still the most cost-effective pharmacologic treatment for epilepsy. All these data predict a healthy future for phenobarbital, particularly in helping to close the treatment gap in low- and middle-income countries during its second century of clinical use.
Phenobarbital is effective for partial and generalized tonic–clonic seizures; it shows similar efficacy for “time to 12-month remission” or “time to first seizure”: as phenytoin (PHT) and carbamazepine (CBZ) (Taylor et al., 2001; Tudur Smith et al., 2003). All of the important comparative studies in adults and children were undertaken 15 or more years ago (Kwan & Brodie, 2004). More recently 108 children aged 2–15 years in Bangladesh were randomized to phenobarbital or CBZ in a single center (Banu et al., 2007). The focus of the study was on tolerability, particularly on behavioral side effects. However, there were no important differences in efficacy between the drugs over the 12-month follow-up period. In a recent small prospective three-arm parallel-group, case control trial, 95 patients with Alzheimer’s disease and epileptic seizures were randomized to monotherapy with levetiracetam (LEV), lamotrigine (LTG), or phenobarbital (Cumbo & Ligori, 2010). A 4-week dose adjustment was followed by a 12-month evaluation period. There were no statistically significant differences in efficacy between these three AEDs. Four of the patients taking phenobarbital dropped out because of side effects in this nonblinded study.
One interesting recent study involved phenobarbital use in rural China. This was a demonstration project of epilepsy management at primary health level under the auspices of the World Health Organization (WHO) in partnership with the International League Against Epilepsy (ILAE) and the International Bureau for Epilepsy (IBE). The project represents the largest study of phenobarbital, and the results have been published in a series of reports (Wang et al., 2003, 2006, 2008; Kwan et al., 2012). The project started with a door-to-door survey in 2000, estimating a treatment gap (proportion of people with active epilepsy not receiving treatment) of >60% (Wang et al., 2003). In parallel with an educational program, a pragmatic intervention study was then conducted that recruited patients older than 2 years of age, with at least two convulsive seizures in the previous 12 months and not receiving adequate medical treatment (Wang et al., 2006). Trained local primary care physicians made the initial diagnosis using a specially designed questionnaire to screen possible cases of convulsive epilepsy. A supervising neurologist then assessed screen-positive patients to confirm or refute the diagnosis and to establish baseline seizure frequency in those with a confirmed diagnosis. Patients entering the study took phenobarbital monotherapy as the first option. A total of 2,455 patients were treated between December 2000 and June 2004. At 24 months of treatment, 71% of patients showed significant benefit, with 26% free from convulsive seizure during the entire treatment period and another 45% having >50% reduction in seizure frequency (Table 1). A second survey 6 months after the end of the intervention noted that the treatment gap had significantly decreased to 49.8% (Wang et al., 2008). Therefore, the intervention measures were effective and evidently feasible in rural China, contributing to decrease in the treatment gap of epilepsy.
|Number of patients (%)|
|6 Months (n = 2,217)||12 Months (n = 1,897)||24 Months (n = 1,324)|
|Seizure-free for the whole period||919 (41)||644 (34)||347 (26)|
|Reduced by >75%||305 (14)||415 (22)||415 (31)|
|Reduced by 51 – 75%||245 (11)||230 (12)||185 (14)|
|Reduced by 26 – 50%||162 (7)||146 (8)||99 (8)|
|≤25% change||217 (10)||156 (8)||91 (7)|
|Increased by at least 25%||369 (17)||306 (16)||187 (14)|
Because epilepsy is a chronic disorder, it is important to determine the long-term impact of any intervention to assess its clinical effectiveness and to inform health care planning. At the study team’s recent revisit of the original study cohort, phenobarbital appeared to have maintained its benefits in the long-term, with an estimated probability of 0.53 for patients remaining on treatment at 6 years (Fig. 1; Kwan et al., 2012). This long-term retention rate compares favorably with newer AEDs used in similar settings. Based on the success of the demonstration project the government has established nationwide epilepsy control programs across rural China (World Health Organization, 2009).
Other observational studies
Over the last decade or so, phenobarbital has been investigated for treating epilepsy in several observational studies, largely undertaken in developing countries. Their conduct was often driven by practical considerations in attempting to establish a feasible clinical service and a constant supply of medication. They tended to include unselected, often untreated, patients with a range of seizure types across all age groups. There have been several published reports, which we shall now consider separately.
To evaluate the availability and accessibility of AEDs in two health districts in Cameroon, Preux et al. (2000) interviewed 33 patients with epilepsy, 26 physicians, 13 private and 8 hospital pharmacists, 8 traditional healers, and 3 drug distributors. They used structured questionnaires to assess knowledge of epilepsy, treatment accessibility, prescription methods, and availability and frequency of delivery of drugs. Only 1 of 33 patients did not take modern treatment; 91% of the patients were followed by a traditional healer, and 78% by a hospital physician. Phenobarbital was the most frequently prescribed drug by 69% of the doctors and was delivered regularly.
The Yelandur study, conducted in rural south India, enrolled only patients with generalized tonic–clonic seizures (Mani et al., 2001). Just 15 (11%) of the 135 recruited patients were lost to follow-up during the 5-year study period. Seventy-five patients (55%) took phenobarbital, and mostly (n = 68) as monotherapy. More than 50% of patients taking phenobarbital were seizure-free at 1 year, but this dropped to 20% over the next 4 years. There were major adverse events in only three patients taking phenobarbital.
In a hospital in urban Nigeria where 90% of 344 children with epilepsy were treated with phenobarbital, 50.6% achieved complete seizure control and only 2 patients stopped the drug because of intolerable side effects (Sykes, 2002). However, 94 (27%) patients were lost to follow-up after one or two visits to the hospital, making it difficult to draw valid conclusions.
In rural Mali, children and adults with epilepsy were treated at home with phenobarbital at dosages ranging from 50 to 200 mg daily (Nimaga et al., 2002). Advice on compliance was provided and an uninterrupted supply of phenobarbital was guaranteed. After a year of treatment, 80.2% had been seizure-free for the previous 5 months. There were improvements in physical and mental health and social status in some patients with few reported side-effects. The authors concluded that low-dose phenobarbital was very effective in this setting with the appropriate patient support.
A 2-year community-based phenobarbital program in rural Laos offered adult patients phenobarbital, 100 mg daily, if they reported two or more generalized seizures during the previous 2 years (Tran et al., 2008). There were limited available data from 46 retained cases (Fig. 2). Twenty percent of the patients were mentally retarded. Some patients dropped out of treatment and few demonstrated optimal compliance. Six patients died of various causes. The 18 people who completed the program deemed it “efficient” and reported an improved working capacity and quality of life. This study illustrates many of the problems in treating epilepsy in a rural environment (Fig. 2). Particular problems occur with compliance (Asawavichienjinda et al., 2003) and in obtaining a constant, high quality supply of phenobarbital (Laroche et al., 2005; Odermatt et al., 2007). There seems little doubt that phenobarbital was effective, well-tolerated, and safe in this setting.
There are few modern data regarding phenobarbital’s efficacy and tolerability in high-income societies. Its everyday use is likely to be largely confined to nonspecialist clinical settings, such as nursing homes (Galimberti et al., 2006). In a recent study of successful combination therapies from Glasgow, United Kingdom, phenobarbital together with phenytoin and carbamazepine were the third and ninth most common successful duotherapy, respectively (Stephen et al., 2012).
Like other AEDs, phenobarbital is associated with a range of dose-dependent and idiosyncratic drug reactions (Kwan & Brodie, 2004). Those of most interest are phenobarbital’s propensity to produce sedative, behavioral, and mood effects especially in children, and in causing teratogenicity in exposed infants.
Phenobarbital’s reputation for poor tolerability was ignited by results from the double-blind Veteran’s Administration study comparing phenobarbital, primidone, PHT, and CBZ monotherapy (Mattson et al., 1985). Approximately 50% of patients randomized to phenobarbital dropped out of the study mainly due to side effects. However, PB doses were very high, reflecting the therapeutic strategy of “more is always better” prevalent at that time. The wide range of CNS-related side effects included sedation, behavioral problems (particularly hypersensitivity in children), impaired cognition, and depressed affect (Brodie & Kwan, 2004). They predominate in several subsequent unblinded randomized clinical trials from developed countries, whereas high degrees of efficacy and tolerability have been reported consistently in observational studies in developing countries without a similar burden of toxicity. Indeed, a recent meta-analysis found no evidence of an association between phenobarbital and a higher rate of adverse effects (Zhang et al., 2011). However, there was a greater likelihood for phenobarbital to be withdrawn compared with other AEDs, possibly due to higher relative drug doses and concerns about its potential toxicity (Brodie & Kwan, 2004).
In the recent randomized comparison between phenobarbital and CBZ in 108 children in Bangladesh with partial and/or generalized tonic–clonic seizures, follow-up was for 12 months (Banu et al., 2007). Ten children developed behavioral problems, which were unacceptable in four (one on phenobarbital, three on CBZ). The authors concluded that there was no excess in behavioral side effects associated with phenobarbital administration in children with epilepsy in a country with limited resources. In a systematic review of phenobarbital for childhood epilepsy, the author concluded that there was no evidence of adverse behavioral effects with this drug compared to other AEDs (Pal, 2006). There were no apparent differences in IQ or Wechsler Intelligence Scales in 22 seizure-free children in whom phenobarbital was withdrawn in Taiwan (Chen et al., 2001), whereas in an Iranian study the authors reported a small improvement in neuropsychiatric performance, mainly in nonverbal rather than verbal items, in 24 seizure-free children who stopped phenobarbital (Tonekaboni et al., 2006).
Finally, Ding et al. (2012) compared a range of cognitive tests in 144 adult patients treated with phenobarbital with 144 healthy controls matched for age, gender, and education level, who lived in the same Chinese villages. Cognitive test scores and mood ratings were available for 136 people (94%) with epilepsy and 137 controls (95%) at 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. In this study, phenobarbital was not found to have a major negative impact on cognitive function in people with convulsive seizures, and some cognitive gains were observed, possibly due to improved seizure control.
Until recently, there have been no prospective studies of AED teratogenicity and so much of the data in the literature focused on anomalies in small numbers of retrospective cases. For instance, Dessens et al. (2001) interviewed men exposed to phenobarbital in utero at age 19–35 years and reported that 15% of this cohort compared with matched controls had undescended testes at birth. Outcomes from the International Registry of Antiepileptic Drugs in Pregnancy (EURAP) prospective pregnancy register recently demonstrated a dose-dependent risk of malformations with a range of AEDs, including LTG, CBZ, and valproic acid (VPA) (Tomson et al., 2011). A total of 217 women taking phenobarbital monotherapy were recruited. At 1 year, congenital malformations were detected in 5.4% of 166 exposed infants whose mothers took <150 mg phenobarbital daily compared with 13.7% in the 51 infants whose mothers took 150 mg or more daily (p < 0.05). These figures were higher than those occurring in babies exposed to LTG and CBZ, but lower than those reported in infants whose mothers took VPA. The range of defects in the phenobarbital-exposed infants included cardiac anomalies, neural tube defects, hypospadias, renal malformations, and polydactyly. There was no one obvious specific dominant anomaly.
Neonatal and Other Seizures
Many pediatricians still regard phenobarbital as the drug of choice for neonatal seizures (Bartha et al., 2007; Vento et al., 2010). Standardized approaches to this indication, however, remain undeveloped (Bartha et al., 2007). Outcomes are also variable, and phenobarbital may be ineffective when the background electroencephalography is abnormal (Boylan et al., 2002; Bartha et al., 2007). In addition, phenobarbital was consistently ineffective in preventing febrile convulsions across 11 placebo-controlled or comparative trials, including a total of 2,159 participants in a systematic review (Pal, 2006). Again, however, there was no convincing evidence for an excess of behavioral effects over other AEDs.
In a small randomized placebo-controlled trial, phenobarbital given within 6 h of birth to term and near-term asphyxiated neonates significantly decreased the incidence of neonatal seizures, but did not alter the mortality or neurologic outcomes at discharge (Singh et al., 2005). In a second placebo-controlled trial, antenatal phenobarbital did not decrease the risk of intracranial hemorrhage or early death in premature infants. Neither did it adversely affect neurodevelopmental outcomes at 18–22 months of age (Shankaran et al., 2002). In another study of 23 infants born to mothers established on phenobarbital, around half demonstrated acute sedation or withdrawal symptoms (Zuppa et al., 2011).
We could identify no randomized or observational studies undertaken with phenobarbital in patients with status epilepticus from 2000 onward. However, phenobarbital use is cited in a number of guidelines usually for established and/or drug-resistant convulsive and nonconvulsive status in adults and children (Appleton et al., 2008; Meierkord et al., 2010; Shorvon, 2011).
There have been a small number of studies published since 2000 exploring the efficacy of phenobarbital for a number of different potential pharmacologic indications. Crawley et al. (2000) undertook a placebo-controlled intervention study with phenobarbital in childhood cerebral malaria. A single intramuscular dose of 20 mg/kg provided highly effective seizure prophylaxis, but was associated with increased mortality. Respiratory arrest was more likely in the phenobarbital group, and mortality was greatly increased in children who also received three or more doses of diazepam. In another study, maternal phenobarbital was administered in the hope of reducing the rate of exchange transfusion in neonates with hemolytic disease of the fetus and newborn by increasing bilirubin glucuronidation and thereby reducing subsequent encephalopathy (Trevett et al., 2008). Of interest, after controlling for confounding variables, the relative risk for exchange transfusion after phenobarbital administration was 0.23 (95% confidence interval [CI] 0.06–0.76). Phenobarbital and gabapentin were equally effective in treating alcohol withdrawal in a small randomized comparative trial (Mariani et al., 2006). There were no significant differences in the proportion of treatment completers in each group requiring rescue medication for breakthrough signs and symptoms of alcohol withdrawal. There were no significant differences in withdrawal symptoms or psychological distress, and there were no serious adverse events. However, the study did not include placebo control. The most recently published study with phenobarbital was a prospective randomized comparison with lorazepam in the treatment of mild to moderate alcohol withdrawal in the emergency department and 48 h later (Hendey et al., 2011). There were no differences in efficacy between the drugs for either end-point.
Approximately 50% of patients with newly diagnosed epilepsy become seizure-free with use of their first AED (Brodie et al., 2012a), often at modest or moderate dose (Kwan & Brodie, 2001; Brodie et al., 2007). There is no doubt that phenobarbital is an effective drug in the setting and that the chance of side effects will depend on the dose (Kwan & Brodie, 2004). Recent comparative trials suggest equivalent efficacy and tolerability to other first-line AEDs for the common epilepsies in adults and children (Taylor et al., 2001; Tudur Smith et al., 2003; Appleton et al., 2008; Zhang et al., 2011), although there may be substantial bias inherent in combining different studies using different doses over a 30-year period (Rheims, 2011). As with all AEDs, some patients will not be able to tolerate phenobarbital at low dose due to sedation and others will develop behavioral side effects. The efficacy of low-dose phenobarbital alone and in combination is likely to be high, given its unique γ-aminobutyric acid (GABA)ergic mechanism of action (Brodie et al., 2011). On the negative side, is its propensity to produce drug interactions and other potentially injurious problems due to broad-spectrum enzyme induction (Brodie et al., 2012b). What cannot be disputed is phenobarbital’s cost-effectiveness in the developing world (Chisholm, 2005).
In view of its affordable cost, ease of use with once-daily dosing, reliable supply, broad spectrum, unique mechanism of action, and comparative efficacy to other established AEDs, it is not surprising that phenobarbital is the most prescribed AED worldwide. However, its purported behavioral and cognitive side effects, identified largely in randomized controlled trials in high income countries, have raised genuine ethical concerns. Observational studies from developing countries have reported a high degree of effectiveness, but some of these had serious methodologic deficiencies. In particular, availability of a constant supply of phenobarbital and a clinical setting appropriate to a good clinical service were often lacking. The lowest price quoted by the WHO is $0.008 per 100-mg tablet. However, in a recent study it was not possible to obtain appropriate information of the availability and price of PB in 46 countries (Cameron et al., 2012).
A pragmatic response to this dilemma would be to conduct a modern, comprehensive, prospective evaluation program with the primary goal of optimizing phenobarbital use as monotherapy for untreated epilepsy. This project could help solve the pressing need to devise cost-effective and sustainable treatment programs for epilepsy, while providing long-term outcome data and evaluating factors that influence treatment response (Kwan & Brodie, 2004). The increasing impressive results from the WHO/ILAE/IBE China project suggests that this is likely be a successful enterprise. It is arguable that phenobarbital’s negative reputation regarding tolerability comes more from its lack of a commercial sponsor than from a critical analysis of the available literature (Kale & Perucca, 2004). In the meantime, phenobarbital still has a major role to play in narrowing the treatment gap in low- and middle-income countries into the second century of its clinical use (Mbuba et al., 2008).
In this sense, the continuous popularity of phenobarbital arguably parallels more of the socioeconomic status and organization of the health care setting in which it is used than mere clinical considerations. It also reflects the disappointing situation that the symptomatic nature of AED therapy has not fundamentally changed in the last 100 years. It might be amusing to ponder whether Alfred Hauptmann would be pleasantly surprised if he were to know that phenobarbital remains the most popular AED 100 years after his serendipitous discovery of its antiseizure properties. However, it is certainly our hope that, when as a result of socioeconomic advancement, the resource-poor countries are no longer so, and future epilepsy treatment is no longer symptomatic but disease-modifying or even preventive, phenobarbital would have served its historic role long before its next centennial anniversary.
Neither author has any conflict of interest related to this review. We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.