Phase 1 Research Plan

(Download Phase 1 Research Plan as PDF)

Summary -- Phase 1 Serious Adverse Event Research Plan

September 2007- October 2009


Although the Consortium’s scientific scope is broad in principle, it will focus initially on two projects. It will first identify DNA variants associated with drug-induced liver-disease (“DILI”) and Serious Skin Rash [Stevens-Johnson Syndrome (“SJS”) and Toxic Epidermal Necrolysis (“TEN”)]. These two projects, while important in their own right, will also allow the SAEC to generate initial results in a reasonable time frame [due to the availability of established case-control DNA sample collections]. It will afford the Consortium the chance to build its core operational processes and facilitate the timely development of the informatics and data analysis/coordinating center at Columbia University.

Simultaneous with the Phase 1 research focus described above, the SAEC will plan follow on, hypothesis driven studies [post whole genome association studies] for DILI and SJS and explore the feasibility of whole genome research on additional SAEs. As shown in Figure 1 below, the SAEC will work over time in two research dimensions [i.e. within existing SAEs and expand to additional SAEs].

SJS/TEN Phase 1 Project

The clinical and genotyping data from the SJS and TEN cohorts will be donated to the SAEC, by GSK and Illumina, for analysis and follow up study. [Illumina contributed the WGGT of all SJS cases and related controls.] SJS and TEN are related, rare, and severe mucocutaneous blistering disorders associated with over 200 medicines. Current estimates place the incidence rate of SJS at 1-6 cases per million person-years and TEN at 0.4-1.2 cases per million person-years. The SJS/TEN cases were gathered via clinical investigators in both the US and UK during the period 2001-2004. Each case was externally adjudicated by Prof. Robert Stern, MD of Harvard University. Over 200 potential cases were reviewed, netting the SJS/TEN cohort. The cohort consists of 37 SJS cases, 34 TEN cases, and two cases with SJS/TEN overlap. Cases were collected both retrospectively & prospectively. In addition, there are 140 controls matched for age, gender, and ethnicity. The mean age of subjects is 41.2 yrs. 69% are female and 31% are male. Their ethnic mix is as follows: White (79.5%), Black (9.6%), Hispanic (5.5%), and Asian (2.7%).

The inclusion criteria for enrollment in the SJS/TEN case cohort were as follows:

  • Males and females aged 18+ at enrollment (hospitalized)

  • Widespread exanthema with 1% or more detachment of epidermis; more than one blister, not only accral extension, with or without mucous membrane erosions in the spectrum of SJS-TEN

  • Able to complete adequate phenotyping including detailed clinical histories, photographs and biopsies, where available

Subjects were ineligible to participate in the study if they met any of the following exclusion criteria:

  • Any condition or disease, which, in the judgement of the investigator, would place the subject at undue risk, interfere with evaluation of the objectives of the study, or interfere with the ability of the subject to complete the study;

  • The subject is known to be HIV positive;

  • The subject has undergone bone marrow transplantation;

  • The subject is immuno-compromised; or

  • Subject is judged by the investigator as being mentally unfit to give informed consent.

The primary objective of the SJS/TEN phase 1 project is to identify genetic susceptibility factors for the development of severe rash. An important secondary objective is to assess the effectiveness of a population control study design versus case/control design. In support of this objective, GSK will also donate 1,500 population controls to be used in support of both the SJS/TEN and the DILI studies. Some of these data and the resulting informatics infrastructure will be leverage in the initial DILI study summarized below.

The genotyping data for the SJS/TEN cases will include 26 SJS/TEN candidate genes, classic HLA-A, B, C, DR, DQ; a high density (7,000) SNP panel manufactured by Illumina, as well as HSR associated markers (19 genes) identified in the Abacavir study. Whole genome association scans using both the Affymetrix 500K Mendel Chip and the Illumina 1M chip will also be completed.

Standard statistical genetic analysis will be completed at the end of the genotyping phase of this project. In addition, the Data Analysis and Coordinating Center will coordinate analyses using novel whole genome analysis techniques suggested both by the members and academic advisors to the project.


Figure 2 below summarizes the SJS/TENS research operating plan and associated milestones, as of August 2007.



DILI Phase 1 Project

The Drug Induced Liver Injury (DILI) case cohorts will be furnished to the SAEC in collaboration with two initial European DILI research networks (i.e. DILIGEN and EUDRAGENE). As quoted in the DILIGEN proposal to the SAEC, The vast majority of DILI is due to idiosyncratic reactions not predictable from drug dosage/concentration. These reactions place a considerable burden on healthcare services worldwide and pharmaceutical industry during drug development. The fact that most of the agents commonly associated with DILI are very useful medicines means that simply limiting their use is not a feasible option. Instead, it is important that we understand the genetic and environmental factors associated with these adverse events, with the aim of developing strategies to identify susceptible individuals prior to prescription allowing the option of substituting alternative therapies. Many different drugs can cause DILI, with the precise pattern of injury varying between drugs. Typically, DILI reactions are classified as “hepatocellular” when the injury is focused on the hepatocyte and “cholestatic” when the damage occurs at the hepatocyte canalicular membrane or further downstream in the biliary tree. The precise mechanisms of either type of reaction remain unclear; however, there is emerging evidence that both drug metabolites and immune factors play a role. Consequently, common variations (polymorphisms) in genes encoding proteins involved in either the production or clearance of drug metabolites, or in immune regulation offer a potential explanation for inter-individual susceptibility to DILI.”

There are approximately 180 DILI cases presently available to the SAEC through its collaboration with DILIGEN and EUDRAGENE. Both groups have entered into sponsored research agreements with the SAEC to continue to yield additional cases through out the first year of the SAEC’s research activities. It is expected these efforts will yield an additional 130 DILI cases. In addition, the Consortium will continue to cultivate collaborative relationships [post formation] with the NIDDK/DILIN (US), Spanish DILI (Spain), and the Japanese Biobank (Japan). Were all of these parties to collaborate with the SAEC, the total available DILI case cohort could approach 1,000 patients.

In general terms, the DILI inclusion criteria are defined as patients with either (a) clinically apparent jaundice or bilirubin > 40 mol/l (after exclusion of cases due to hemolysis), or (b) an ALT >5x ULN (upper limit of normal) or (c) an ALP >2x ULN plus any raised bilirubin above ULN. These cases were be recruited both prospectively and retrospectively via the two networks. 47% of the DILI cohort is male and 53% female, with a mean age of ~53. The cases are primarily Caucasian (98.0%), with the remainder Black (0.5%), and Asian (1.5%). Over 30 different drugs have been cited as being associated with onset of DILI [including a variety of different drug combinations]. The 1,500 population controls for this study will be matched as closely as possible on ethnicity, age and gender. Appendix A presents a high level comparison of the clinical report format [CRFs] from three major DILI case aggregation efforts, i.e. EUDRAGENE, DILIGEN, and DILIN.

The genotyping data for the DILI cases will be generated from the 1+ Million SNP panel manufactured by Illumina. This panel includes 1,072,820 SNPs, with a mean spacing between SNPs of 1.5kb, 565,000 SNPs in genes, 24,000 nsSNPs, 15,486 SNPs in ADME genes, and 10,073 SNPs in MHC regions. The exact genomic spacing is profiled in the table below:

As part of its genotyping approach, the SAEC will accommodate a “candidate gene component” of 50 -100 hepatotoxity related genes, with SNP coverage of at least 80% genomic information across three populations.

Parenthetically, power calculations and sample size considerations are an important factor to consider in any study design to ensure the quality of the proposed scientific experiment and to provide one factor into our expectations of the likely results.   In reality, the weight that can be given to the power calculation depends upon the certainty of the assumptions made in the underlying calculation. In many genetic and non-genetic situations these are not known and so we estimate based on what we know and in the case of drug development based on what would be a meaningful result e.g. we power many studies to detect a clinically meaningful benefit from placebo.  The study may be over or underpowered to detect the true benefit but a decision is made that it has to be at least a minimally important clinical benefit over placebo.   Power calculations in genetic studies have the additional complexity of not knowing the magnitude of the effect of the polymorphism. We also do not know how common it is (not to mention the fact that we are usually measuring a surrogate linked to the causative variant by LD).  
The SAEC SMC reached the following consensus relative to powering considerations in SAE WG Association studies: 

  1. There are clear examples in the literature of pharmacogenetic effect sizes in SAEs of significantly greater than 3 and detectable with sample sizes well below those identified by the consortium.
  2. Formal power calculations are of limited value here as the range of values for the assumptions is large i.e. true genetic effect could be as low as 2 or 3 or as high as 50 and the frequency of the genetic variant could range from 1-5% all the way up to 50%.
  3. To be clinically useful the size of the genetic effect needs to be at least moderate and so we are looking for polymorphisms with effect sizes of greater than 3.
  4. The SAEC has sufficient samples to detect a clinically significant pharmacogenetic effect which is common across DILI reactions; but additional are always beneficial,
  5. The study proposed by the SAEC should answer the question of are there polymorphisms which have sufficient predictive power to be used as a meaningful clinical test and guide prescribing.

A number of fine PGx studies have generated important results, powered to the level of our first two studies.   [i.e. For Gefitinib (Diarrhea), RR = 5, number of cases required for detection = ~30. For Abacavir,  RR = 36, number of cases required for detection = ~15. For Carbemazipine (SJS), RR = 1000, number of cases required for detection = ~ 6-9.]

Standard statistical genetic analysis will be completed at the end of the genotyping phase of this project. In addition, the Data Analysis and Coordinating Center [DACC] will coordinate all analyses using novel whole genome analysis techniques suggested both by the members and academic advisors to the project. Those data (clinical and genotype) able to be released as dictated by informed consent and reasonable ethical practice, will be released (via a dedicated web portal) to qualified researchers within 12 months after the completion of the genotyping phase. This will allow amble time for the preparation of publications. No individual researcher working within the Consortium will have preferential access to the raw data set. All analysis requests will be brokered through SMC, and will be executed by the DACC staff. The results of these analyses will be presented to and discussed by participating researchers via the SMC’s deliberations and meetings.

Figure 3 below summarizes the DILI research operating plan and associated milestones, as of August 07.


Summary -- SAEC Research Policies

As note above, the SAE Consortium is organized as a private 501 c 3 organization which functions in the “public good”; thus, tax law influences its research/operational policies. In addition, its dues paying members are the world largest pharmaceutical companies which represent a significant percentage of the global [ethical] pharmaceutical industry; thus, considerations under international anti-trust law have an important influence on the SAEC’s research policies.

In light of these factors, as well as the requirements of SAEC’s clinical collaborators, their institutional requirements, and the current genomic/personal data handling policies by international research organizational such as the Wellcome Trust, the following policies currently guide the Consortium’s translational research programs:

  • Public Research Data Release: To promote the public welfare and enable the broadest beneficial use of the results of the Consortium’s research efforts, all research data, except those specifically excluded by patient informed consent or institutional IRB policies, will be made available on a non-discriminatory basis to all qualified researchers, at the same time, at no charge. These data will be made available to the public no later than twelve (12) months following the QC approval of the complete genotyping data set by the DACC at Columbia University. This will provide participating researchers with the requisite time to analyze the data and generate publications. In order to limit the privacy risks and to comply with any other data use limitations (e.g., informed consents), specific Program Data will not be classified as Public Data unless and until it has been so designated by the SMC.

  • Restrictions on Public Data Access: The DACC will make the Public Data available through a controlled-access database to all researchers worldwide who have agreed (and whose institutions have agreed) to comply with certain restrictions determined by the SMC. Such restrictions will include the following: (i) not to share the Public Data with any person who is not employed by a Public Researcher, (ii) not to attempt to identify individual subjects represented by genotype or phenotype data, (iii) not to use the Public Data for other than legitimate pharmaceutical or biological research purposes, and (iv) such other reasonable restrictions consistent with this Policy as determined by the SMC. Public Researchers shall be required to agree to comply with the foregoing restrictions through a signed agreement or certification and/or a “click-wrap” document that must be accepted prior to accessing online Public Data. Any Public Researcher using Public Data shall acknowledge the Consortium in any resulting oral or written presentation, disclosure or publication relying on such Public Data.

  • Genotyping Data: SAEC Genotyping Vendors will perform the genotypic analyses (including quality control) required by the Consortium on the Samples received from Clinical Sources. At the Consortium’s request, Genotypic Data generated by a Genotyping Vendor may be submitted to one or more QC Vendors for quality control purposes. Such QC Vendor will receive only sufficient quantities of Sample and Genotypic Data to assess the quality of such Genotypic Data. The SMC will not have direct access to Genotypic Data. Unused portions of the Samples will be returned to the supplying Clinical Source in a timely manner.

  • Data Analysis: The Data Analysis Committee [DAC], in cooperation with the DACC, will determine suitable strategies for analysis of the Clinical Data and Genotypic Data. The DACC will perform such analyses in accordance with such strategies and will report any Analytic Results to the DAC. The DAC, in consultation with the SMC, will evaluate the Analytic Results and may request additional analyses from the DACC. Data Analysis Committee and SMC members who have access to Analytic Results shall use such Analytic Results solely in connection with the Consortium’s Research Program, and not for the individual benefit of any institution or company, whether or not a Member of the Consortium. Each such individual will be bound by confidentiality agreements associated with Consortium membership, and will be prohibited from sharing Analytic Results with any individual who is not a member of the SMC or DAC.

  • No Preferential Data Access for Members: The DAC or any other member of the Consortium will not have direct access to Clinical Data or Genotypic Data and shall only be permitted to access the Public Data on the same terms and conditions as all researchers.

  • Treatment of Program Data: Each recipient of Program Data (whether a member of the SMC or Data Analysis Committee, a Genotyping Vendor or the DACC) shall be responsible for complying with all applicable national, state and local laws, rules, regulations, enactments, directives, orders and standards and relevant institutional policies and requirements, and shall be required to maintain the strict confidentiality of any Individual Data. All persons and organizations having access to Program Data (other than Public Data) shall be required to take reasonable security measures to ensure that such Program Data is not compromised, improperly disclosed or misappropriated. Individual Data shall be protected to the greatest extent practicable. Program Data (including Individual Data) other than Public Data shall be used solely in support of the Consortium’s Research Program.

  • Publication Preparation: Following the receipt of the final Analytic Results, members of the Data Analysis Committee and SMC may prepare scientific papers based on such Analytic Results for publication in peer-reviewed scientific journals. Each such paper must be provided to the SMC prior to its first submission to a journal. The SMC shall have a period of 45 days in which to provide any comments and to determine whether all appropriate authors have been credited in such paper, and the author shall comply with the SMC’s recommendation as to any additional authors. All such papers shall acknowledge the support of the Consortium.

  • Intellectual Property: It is the goal of the Consortium to maximize the public benefit of research supported by the Consortium and, accordingly, to make freely available DNA markers for susceptibility to drug-induced severe adverse events and related data and analyses to all parties. The Consortium intends to release all Public Data as early as possible so as to place it in the public domain and reduce the likelihood that use of the Public Data will be encumbered by patents. In some cases, the Consortium may determine that the most effective way to ensure that Public Data is placed in the public domain, with the earliest available priority date, is to file a provisional patent application covering all novel discoveries made prior to the filing. Such filings shall include one or more claims directed toward the Program Data (including genetic markers and genotype/haplotype-phenotype associations). Subsequent to this provisional application, the Consortium may file additional utility applications to further validate or expand on its initial utility. At the end of the Research Program Activities, with respect to a given indication, patent counsel may file a final utility application, with the intention that such application either be abandoned following publication or converted to a statutory invention registration in the U.S. Researchers who develop intellectual property funded by the Consortium will be required to assist the Consortium in any such filings or other procedures deemed necessary by the Consortium to ensure the contribution of such intellectual property to the public.

  • Inventions Based on Public Data: Each person accessing or using SAEC’s Public Data, and his/her institution/organization, must agree not to file or support any patent application claiming any DNA marker(s), genotype/haplotype-phenotype association or other attribute disclosed as part of, or derived from, the Public Data or that would prevent or block access to, or use of, any element of the Public Data, or conclusions drawn directly from the Public Data.

  • No Limitation of Downstream Protection: Subject to the previous point, the Consortium acknowledges that intellectual property protection may be appropriate for inventions and discoveries made by Members and/or Public Researchers, where such inventions and/or discoveries have been enabled by the Public Data and/or the Research Program Activities, but are not directly derived from the public data. Such “downstream” inventions may include novel assays, drug targets, therapeutics and diagnostics developed using DNA markers discovered through analysis of the Public Data, but whose utility is not solely derived from the associations or other information contained in, or generated directly from, the Public Data. The SAEC acknowledges such developments does not provide the Consortium with any ownership or control of such downstream intellectual property and, accordingly, Members are not prohibited hereunder from filing or supporting any patent application claiming such “downstream” inventions. P

Please refer to The SAEC’s “Data Release and IP Policy for more details on these policies.

In addition, there are a number of “values” which influence the SAEC’s approach to collaborative research and the expectations it holds for its research collaborators and partners. These are summarized below:

  • Business Objectivity and Prudence: SAEC projects will be selected solely on the basis of the ability to achieve results in a given timeframe, with acceptable risks, and a clear understanding of the investment required and the return expected. The consortium will function in a “business prudent manner” in all of its research dealings. It will select the best available collaborators as partners without bias. Projects will only be undertaken with strong, experienced collaborators.

  • Strong, Focused Management: The affairs of the SAEC will be directed and lead by dedicated management, with strong professional management experience across the life sciences research continuum. This will included leadership experience specifically in the definition, development and execution of international research consortia.

  • Strong Project Management – All SAEC projects will have an agreed upon project strategy, including firm milestones deliverables, and actions to deal with “under-performance”.

  • Innovation: The Consortium will look to create and deploy innovative research and organizational methods, which leverage the skills of parties that heretofore have not exploited their ability to work together to generate research outcomes desirable to the public.

  • Pro-Competitive Activities: All SAEC projects will be “pre-competitive” [i.e. not serving directly the competitive interest of SAEC members], which result in data of a broad scientific utility that are “pro-competitive [i.e. increase the ability of all parties to develop more and better products and/or services].

  • Strong Member and External Expert Involvement: The effectiveness and efficiency of the SAEC’s program depend entirely on the volunteerism of its members and top academic to serve on the consortium’s various administrative and scientific committees. The SAEC will strive to get the best talent involved in its scientific affairs.

  • Regulatory Involvement: The research agenda of the SAEC will be developed in-conjunction with the FDA and other international drug regulatory bodies. As such, the SAEC will mindful of these organizations’ priorities and strive to include the relevant clinical and scientific talent from these agencies. Where appropriate, the SAEC will apply GLP, GCP, etc. practices to maximize the utility [in drug regulatory submissions] of the research data generated from its SAE research projects.

  • Global Orientation: The SAEC research orientation is global. It will strive to build diverse research coalitions across geographies, without bias to any one market.

Drafted by:

Arthur L. Holden
Chairman and CEO
SAE Consortium, Ltd.