This is Part 2 of a three-part series. See also Parts 1 and 3.
17 November 2008. Scientifically, the Dominantly Inherited Alzheimer Network (DIAN) will test three hypotheses. The first proposes that there is a long presymptomatic period in eFAD mutation carriers who will get AD, and that biochemistry and brain imaging can detect this outwardly quiescent phase. This hypothesis rests on ample prior data, and it can be cleanly tested by comparing how a range of potential biomarkers change, or do not change, in mutation carriers versus their non-carrying siblings.
The second hypothesis posits that the preclinical changes will show up in a given sequence in time. By that hypothesis, CSF Aβ42 would start edging down first, amyloid deposition would start appearing in the brain next, and then the person’s metabolic activity in brain would fall off and certain brain regions would atrophy. Somewhere in this sequence, subtle personality and cognitive changes would emerge. Here, too, prior evidence exists to support each of these markers; however, how many years before AD symptoms they begin and if indeed they appear in this order is unknown.
Third, DIAN proposes that symptomatic autosomal-dominant AD is largely the same as “sporadic” or late-onset AD. The test for this hypothesis will come from comparing DIAN’s results to those gathered in ADNI and also to data collected at the National Alzheimer’s Coordinating Center. ADNI follows 819 people diagnosed as cognitively normal, mildly impaired, or having early AD for three years with the same assessments to be used in DIAN. For its part, the NACC is an ongoing initiative to assemble clinical and cognitive assessments, at baseline and longitudinally, on tens of thousands of AD patients seen at the 29 federally funded U.S. Alzheimer’s Disease Research Centers in a unified way. To date, the centers have submitted uniform data on more than 12,500 patients to NACC, Morris wrote in DIAN’s grant application. In order to enable comparison of all these large datasets, the 10 DIAN centers will use the Uniform Data Set (UDS) of clinical and cognitive instruments that is also being used by the ADRCs, in ADNI, and by the ADCS.
Beyond the Unified Data Set, which uses paper and pencil, DIAN is considering eventually using Web-based cognitive tests, as well. This is necessary because many DIAN participants will be young and cognitively healthy, whereas the UDS was developed for elderly people and those whose cognition is already declining. Hence, the investigators expect that some study DIAN participants may ace the UDS tests, which may then miss the first signs of subtle change. At a planning meeting held last July in Chicago, neuropsychologists from the participating sites discussed what those additional tests should focus on. One early clue that might be exploited is that people who later develop dementia tend not to improve when they take the same test repeatedly. Cognitive domains such as problem solving, attention, visuospatial processing, and language hold promise as well, as do assessments of personality changes years prior to dementia. Ideally, these additional tests should be set up such that the participants can take them independently at study visits or even at home; they are not intended to supplant whatever protocols the participating sites already are using for their other studies, emphasized Martha Storandt of WashU. These computerized tests are still being developed, and will be incorporated into DIAN later, not when the study starts next January.
A planning and training meeting held in Chicago last July made evident that DIAN is being set up to cooperate closely with other large initiatives in AD research in terms of its structure, assessments, and data handling. In terms of structure, DIAN’s coordinating center at WashU will take advice from a steering committee that is the main decision-making body, plus from an external advisory group. The steering committee includes the leaders of the ADCS (Aisen) and ADNI (Michael Weiner of UCSF), besides WashU’s DIAN administrative and core leaders and the leaders at the participating sites. Representatives from the FDA and NIA, one family member each of an English-speaking and a Spanish-speaking family with eFAD, and Kimberly Quaid of Indiana University, a bioethicist and expert on genetic testing in neurodegenerative diseases (see article on genetic testing), serve on this committee. Like ADNI, DIAN has a resource allocation review committee. This group oversees requests by outside scientists to obtain fluid or tissue samples for their own analyses. It comprises research leaders who are active on this issue within the ADCS (Douglas Galasko, UCSD), ADNI (John Trojanowski, UPenn), and the Alzheimer’s Disease Genetics Consortium (Gerard Schellenberg, UPenn). Also like ADNI, DIAN has cores dedicated to the requisite sub-areas: clinical, biostatistics, neuropathology, biomarkers, genetics, imaging, and informatics, each headed by a scientist at WashU. The genetics core, headed by WashU’s Alison Goate, will collaborate with the National Cell Repository for AD in Indianapolis for the establishment of cell lines, as done in ADNI. For a detailed look at DIAN’s structure and people, see slides: DIAN structure, steering committee, other committees.
In terms of assessments, DIAN’s imaging, clinical, and biomarker cores will collaborate especially closely with ADNI and ADCS, Morris said at the July meeting. Led by WashU’s Mark Mintun, the imaging core contracted with Clifford Jack at the Mayo Clinic, Rochester, Minnesota, to perform quality control and preprocessing for MRI scans taken at the 10 sites, and with Robert Koeppe of the University of Michigan, Ann Arbor, to do the same for PET scans. DIAN has retained Chet Mathis at University of Pittsburgh, who heads ADNI’s amyloid imaging sub-study, to certify that the participating DIAN sites synthesize 11C-PIB correctly. The participating sites will upload their scans to WashU’s DIAN central archive, from where the quality control groups will pull it, check it, resolve with the local sites any problems that might crop up, and provide some pre-processing. They will then release the clean imaging data to DIAN investigators for further processing and analysis. In a departure from ADNI, DIAN’s raw scans will be placed under quarantine until quality control is finished, whereas ADNI scans are available instantly to the world’s scientific community (see ADNI Snafus). DIAN will use the ADNI imaging protocols so that data will be comparable, but details about the precise MRI sequences will be updated slightly to reflect progress in neuroimaging since 2004 when ADNI was designed. All DIAN participants will receive MRI and PET scans so scientists can measure brain atrophy, search for small hemorrhages, and image the buildup of amyloid deposition, the brain’s glucose metabolism, and some network function.
DIAN’s clinical core will work closely with the ADCS. This consortium was founded in 1991 by the National Institute on Aging; Neil Buckholtz of the NIA is that agency’s representative of the ADCS, ADNI, and now also DIAN. The late Leon Thal of UCSD pioneered the ADCS, and he supported a registry for observational science preceding prevention trials until his untimely death (see Thal quotes in eFAD prevention story and eFAD clinical trials essay). While the ADCS is best known for running clinical trials, it focuses intensely on markers as well. “The ADCS is involved in imaging and biomarker work. That is critical for finding disease-modifying therapies, because they are most likely to succeed at the earliest, even preclinical stages. DIAN is a clear extension of that interest,” Aisen said at the planning meeting in Chicago. The ADCS has much to offer DIAN because its research staff has spent the past decade establishing clinical infrastructure and standard operating procedures. This may not be flashy, career-building science, but it is essential for large, multicenter clinical studies to succeed.
The ADCS provides clinical and cognitive assessment monitoring for ADNI, and will do the same for DIAN. In practice, this means that ADCS will create a DIAN site. Its staff will support DIAN’s clinical, biomarker, and genetics cores. For example, the 10 distributed sites will enter their raw clinical and psychometric data to a DIAN site at ADCS, where ADCS staff will perform quality control, among other functions. Once those data are “clean,” ADCS staff will upload it to the DIAN central archive at WashU—an informatics resource that will host a unified database of all DIAN data. This arrangement makes available to DIAN ADCS’s innovations around real-time, online data entry from distributed sites. For example, data entry for ADCS studies resembles online ordering, where quality assurance is built into the system so that it prompts the person entering clinical data to fix problems with given fields before allowing the data into the system. The online, real-time case report form then serves as the source document, not the original paper record. “The key about that is that only clean data will be allowed into the system,” Jennifer Emond of the ADCS said at the Chicago DIAN meeting. ADCS also will deploy its clinical operations group to help coordinate DIAN. They develop and distribute procedures manuals, train site personnel in these procedures, and make sure sites adhere to them. The ADCS has nine permanent clinical monitors, who evaluate sites for their capacity to carry out DIAN and review HIPAA, regulatory, and other records to ensure sites follow the protocol procedures. Both Devon Gessert and Mario Schittini from the ADCS noted that their team will keep DIAN moving apace by insisting that sites submit their data within three days of a participant’s visit, but they softened their muscle-flexing with offers to provide extensive training in study procedures. “All this serves to reduce site variance,” Schittini said.
The DIAN center at ADCS will also support DIAN’s fluid biomarker work. Like in ADNI, DIAN study participants will be asked to donate blood and CSF at each visit, though there are some small differences in how they will be analyzed. (For example, DIAN will not analyze homocysteine, isoprostane, and sulfatide levels, as the investigators believe that these factors are less important in autosomal-dominant AD than in sporadic AD. It will also not collect urine samples.) The 10 sites will send most samples to the DIAN biomarker core at WashU (led by Anne Fagan) for analysis. Here, too, the ADCS will coordinate sample collections at the sites and information management, before transferring clean data to the DIAN central archive.—Gabrielle Strobel.
This is Part 2 of a three-part series. See also Parts 1 and 3.