The lacking evidence for the heritability of drug response – and why nobody cares about it

Seizure disorders Hughlings Jackson, one of the pioneers of epileptology, commented that “man is built to seize”.  Seizures, pathological synchronisation of network activity in the Central Nervous System, are the final common pathway of different pathogenic processes within the brain.  Whereas fever is the most common provoking factor in children, provoking factors in adults are metabolic derangements, head trauma or brain tumours. Epilepsy, in contrast to seizures, is defined as the occurence of more than one unprovoked seizure.  This reflects an inherent hyperexcitability of the Central Nervous System.  Seizure disorders are common with up to 2% of the population affected and have a strong genetic component.

Drug response in epilepsy A significant proportion of seizure disorders can be controlled with antiepileptic medication and up to 30% of all patients are seizure free on medication.  Novel antiepileptic medications are constantly developed which harbour the potential of providing a broad spectrum for the physician and the patient to choose from.  However, many medications carry common and rare, potentially lethal side effects.  Finding the medication that provides the best seizure control while having the fewest side effects is the ultimate goal in each patient. Furthermore, up to one third of all patients in specialised epilepsy centre are pharmacoresistant [1].  Accordingly, there is an urgent need to understand factors that influence drug response in patients with epilepsy.

Pharmacogenetics Identifying genetic variations that predict response to particular antiepileptic medications has received much attention in recent years [2, 3].  This research is mainly motivated by particular drug responses in rare genetic epilepsy syndromes such as Dravet syndrome [4] and animal studies on selective inbreeding [5].  Several candidate genes have been suggested that determine drug response in patients with epilepsy, but most studies so far lack replication and draw upon a biased patient population [2].  Accordingly, the potential for pharmacogenetics is difficult to estimate.  Epidemiological methods such as the analysis of familial aggregation are inadequate to answer the question as to how much of the response to antiepileptic medication is due to genetics factors and whether the field of pharmacogenetics is built upon a reasonable hypothesis.

Response to medication in twins with epilepsy Twins have traditionally been the flagship of epilepsy genetics.  Ever since the first report of a concordant twin pair with Temporal Lobe Epilepsy by Wolfson in 1929 and the historical studies by William Lennox in the 1950’s [6],  twins have been crucial in demonstrating a genetic impact on a variety of epilepsy syndromes [7].  Particular the Idiopathic Generalised Epilepsies are considered to be almost essentially genetic nowadays with a heritability of more than 80% [8].  Constance and Kathryn, an identical (monozygotic) twin pair with Childhood Absence Epilepsy, have become the well-recognized face of twin research in epilepsy genetics.  Constance and Kathryn both suffered from Childhood Absence Epilepsy and the EEG recordings showed virtually identical traces.  However, it is commonly not known that Constance and Kathryn were discordant for drug response [9].  Whereas seizures could easily be controlled in one twins, the other twin required multiple medications to gain control of the epilepsy.  Apparently, factors other than genetics influenced the response to medication in this pair. However, there is also evidence for striking similarities in response to medication between twins, such as identical concordant twin pairs with severe Childhood Absence Epilepsy and an concordant unusual response to steroid treatment. In summary, little is known about similarities and differences between drug response in identical twins except for anecdotal evidence and the heritability of drug response is not known. What is the consequence of this lack of information on the genetics of drug response? As strange as it might sound at first glance, researchers are not particularly concerned about the lack of clinical genetic evidence and even look for genetic factors in the absence of clinical evidence. In fact, common genetic variants might actually be prevalent in conditions that do not seem to be heritable at first glance [10], a topic that will be addressed in a future blog.


Please quote this blog (APA format) as Helbig, I. The lacking evidence for the heritability of drug response – and why nobody cares about it. Retrieved [enter date], from

1.           Kwan, P. and M.J. Brodie, Early identification of refractory epilepsy. N Engl J Med, 2000. 342(5): p. 314-9.

2.                 Szoeke, C.E., et al., Update on pharmacogenetics in epilepsy: a brief review. Lancet Neurol, 2006. 5(2): p. 189-96.

3.                 Sisodiya, S.M., Genetics of drug resistance. Epilepsia, 2005. 46 Suppl 10: p. 33-8.

4.                 Mulley, J.C., et al., Channelopathies as a genetic cause of epilepsy. Curr Opin Neurol, 2003. 16(2): p. 171-6.

5.                 Cramer, S., U. Ebert, and W. Loscher, Characterization of phenytoin-resistant kindled rats, a new model of drug-resistant partial epilepsy: comparison of inbred strains. Epilepsia, 1998. 39(10): p. 1046-53.

6.                 Lennox, W.G., The heredity of epilepsy as told by relatives and twins. J Am Med Assoc, 1951. 146(6): p. 529-36.

7.                 Berkovic, S.F., et al., Epilepsies in twins: genetics of the major epilepsy syndromes. Ann Neurol, 1998. 43(4): p. 435-45.

8.                 Berkovic, S.F., et al., Human epilepsies: interaction of genetic and acquired factors. Trends Neurosci, 2006. 29(7): p. 391-7.

9.                 Lennox, W.G.L., M. A., Epilepsy and Related Disorders. 1960, Boston: Little, Brown & Co. 548-574.

10.                 Visscher, P.M., W.G. Hill, and N.R. Wray, Heritability in the genomics era–concepts and misconceptions. Nat Rev Genet, 2008. 9(4): p. 255-66.

Can microattribution be applied to epilepsy genetics?

Genetic studies in seizure disorders require at least two components, (a) recruitment/phenotyping and (b) genotyping.  While genotyping technologies slowly turn into standardized industry-like analysis pipelines that allow for fast processing of large sample volumes, recruitment and phenotyping still requires significant human input and is unlikely to become more efficient in the future.  Bluntly said, it is difficult to imaging that any form of high-throughput recruitment or phenotyping can ever be realized.  On the contrary, the effort and time required for recruiting and sufficiently characterizing patients for epilepsy genetics studies is probably likely to increase, given the complexity of consenting patients for whole exome or whole genome sequencing projects [1].

In summary, large cohorts or patients consented for high-throughput sequencing technologies require an asymmetrically large time investment through referring clinicians.  These investments might already be insufficiently represented in publications, authorships and referenceable attributions of the scientific work.  Needless to say that much of the work required for recruitment and phenotyping is insufficiently covered by research funding and is often unpaid.

Nevertheless, motivated patient recruitment, thorough annotation, clinical data collection and database entry is likely to become a key issue in correlation genotypic and phenotypic data.  How this time-consuming work can be reflection in attribution is yet unsolved.

A recent string of papers in Nature Genetics has highlighted a novel avenue for crediting collaborating researcher for small contributions to a larger scientific project [2-5].  These microattributions consist of crediting authors for database entries by publicly displaying the contributor (e.g. entry created by W.G. Lennox) and agreeing to publishing the work jointly at defined stages of the research.  Applying this microattribution strategy to hemoglobinopathies has resulted in a sharp increase in reported variants [4].

Similar strategies for documenting phenotypic data could be used for collecting phenotypic data for epilepsy databases in large collaborative projects such as EuroEPINOMICS.  This strategy might enable individual researchers to reference their contribution to large collaborative projects, which usually require years until joint papers are written, submitted, reviewed and published.


1. Kaye, J., et al., Ethical implications of the use of whole genome methods in medical research. Eur J Hum Genet, 2010. 18(4): p. 398-403.

2. What is the human variome project? Nat Genet, 2007. 39(4): p. 423.

3. Crowdsourcing human mutations. Nat Genet, 2011. 43(4): p. 279.

4. Giardine, B., et al., Systematic documentation and analysis of human genetic variation in hemoglobinopathies using the microattribution approach. Nat Genet, 2011. 43(4): p. 295-301.

5. Mons, B., et al., The value of data. Nat Genet, 2011. 43(4): p. 281-3.

Please quote this blog (APA format) as

Helbig, I. Can microattribution be applied to epilepsy genetics?. Retrieved [enter date], from

Does the 15q13.3 microduplication protect against epilepsy?

The publication of recent papers by Mefford et al. [1] and Cooper et al. [2] brought up an interesting issue again concerning the role of the 15q13.3 microduplication.  While the 15q13.3 microdeletion is a well-established risk factor for epilepsy, particularly, idiopathic generalised (genetic generalised epilepsy), the role of the reciprocal microduplication is not clear. In two epilepsy cohorts this microduplication is absent in 1223 IGE patients [3], 517 epilepsy patients [4] and 315 patients with epileptic encephalopathies [1].

This variant is seen in two control cohorts at a frequency of 0.5% (2/3699, 3/8329) [2, 3] and in a mixed cohort of patients with intellectual disability at 0.13% (20/15,767) [2], the absence in IGE and other pure epilepsy cohorts is intriguing (Table 1).

Table 1
15q13.3 dup n %
Intellectual disability 20 15,767 0.13
controls 25 27,795 0.09
IGE/GGE, epilepsy 0 2055 0.00

Neither comparison between two groups using a Fisher’s exact test is significant, given that it is conceptually difficult to arrive at statistical significant results with very small frequencies.  However, comparing these three frequencies using a Chi square test, the differences are significant at p=0.02. Accordingly, this might indicate that the 15q13.3 microduplication is a protective factor against epilepsy or the differences are due to ascertainment bias.


1. Mefford, H., Rare copy number variants are an important cause of epileptic encephalopathies. 2011.

2. Cooper, G.M., et al., A copy number variation morbidity map of developmental delay. Nat Genet, 2011. 43(9): p. 838-46.

3. Helbig, I., et al., 15q13.3 microdeletions increase risk of idiopathic generalized epilepsy. Nat Genet, 2009. 41(2): p. 160-2.

4. Mefford, H.C., et al., Genome-wide copy number variation in epilepsy: novel susceptibility loci in idiopathic generalized and focal epilepsies. PLoS Genet, 2010. 6(5): p. e1000962.

Please quote this blog (APA format) as

Helbig, I. Does the 15q13.3 microduplication protect against epilepsy?. Retrieved [enter date], from