For Physicians

Here is a more detailed and technical description of the PGBD study for physicians or scientists interested in the study design.

Though a number of studies of genes and mood stabilizer response have been reported, to date they have suffered from several limitations that have hindered gene detection. First, most are studies have used relatively small samples of subjects numbering in the low hundreds. It is clear from more recent genome wide association studies (GWAS) that larger samples are necessary in order to detect what are likely small gene effects. Secondly, most have examined only a limited number of candidate genes selected based on theories of lithium’s mechanism of action. To date only the STEP-BD GWAS by Perlis et al. has examined genes on a comprehensive genomic scale. Thirdly, most studies have employed a retrospective chart review approach to assessing lithium response. Though this approach is easier and may allow for larger sample sizes, these studies suffer from the limitations of subject memory and medical records. Response is also generally confounded by polypharmacy making it difficult to unequivocally rate lithium response. Lastly, almost all published studies focus on lithium, and little data exists regarding other drugs, thereby limiting the ability to develop a predictor of comparative response.

The PGBD is designed to address these limitations. The PGBD employs a state of the art prospective monotherapy open label relapse prevention design. The design is illustrated in the figure below.

Subjects are enrolled who have bipolar I disorder and are either stable on lithium monotherapy or unstable on another medication regimen, such that a change of medications and trial of lithium is clinically appropriate. Subjects cannot have previously had a clear failure on lithium or intolerance of lithium. Subjects stable on lithium and another medication may enter the study if discontinuation of the other psychotropic medication is clinically appropriate.

Subjects first undergo a 1-2 week screening period when diagnostic and medical evaluations take place. Once it is determined that the subject has bipolar I disorder and is clinically and medically appropriate for lithium treatment, the subject is started on lithium and stabilized over a 4 month period. Typically, other psychotropic medications are gradually tapered as lithium is titrated up in dose during this period. Once patients has been stabilized and other medications discontinued, they enter a one month observation phase, when they are observed to be sure they are in remission and don’t relapse after other medications are discontinued. Hypnotics and anxiolytics are allowed during the study. Following observation, subjects enter the maintenance phase when they are assessed at two month intervals. The primary outcome measure of the study is the time to relapse for subjects who have successfully achieved remission. Survival analysis methods will then be used to determine if variation in genotype is associated with longer time to relapse.

Subjects who fail lithium in either the stabilization, observation or maintenance phases will be crossed over to an identically designed valproate arm. The study has a number of other features designed to maximize the information obtained and statistical power. Detailed information regarding course of illness, premorbid history, and temperament is obtained for use as covariates in the analyses. Subjects are also assessed for major life events at each visit in order to include environmental factors in the analysis. It is a common clinical observation that patients relapse in the context of life events both positive and negative. Blood will be obtained for DNA, which will be used to examine over 1 million DNA markers using genome wide association methods. Should prices continue to drop, it may be possible to instead sequence the entire genomes of all subjects. A subset of subjects will undergo skin biopsy for the production of induced pluripotent stem cells (iPS). In this technology, skin fibroblasts are treated with virally transferred transcription factors that reprogram them into stem cells. These stem cells can then be differentiated into neurons allowing examination of the response of neurons from clinically documented lithium responders and non-responders to lithium in vitro.

Ten clinical sites are participating in the study, which will enroll subjects over 4 years. It is anticipated that a total sample of over 700 subjects will be studied.

This study is supported by the NIGMS and the NIMH. It is a project in the Pharmacogenomics Research Network (PGRN) administered by NIGMS.