Premium
O1‐04‐02: An assessment of variability in CSF biomarkers in clinical experimental models: a meta‐analysis
Author(s) -
Waring Jeffrey,
Slats Diane,
Gonzales Celedon,
Dean Robert,
Lee David,
Siemers Eric,
Claassen Jurgen,
Li Jinhe,
LeBlond David,
Slemmon Randy,
Bateman Randall,
Tong Gary,
Fox Gerard
Publication year - 2011
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2011.05.249
Subject(s) - medicine , placebo , clinical trial , morning , meta analysis , lumbar , pharmacodynamics , lumbar puncture , pathology , cerebrospinal fluid , surgery , pharmacokinetics , alternative medicine
all biomarkers considered for this review. Six percent of studies included more than 500 patients. It was unclear whether scan/test interpretation was conducted blind to knowledge of conversion to dementia in 62% of reports. Similarly, it was unclear whether decisions about conversion to dementia were made without knowledge of scan/test results in 63% of reports. Uninterpretable or intermediate test results were either not reported, or reported unclearly in 79% of cases. It was rarely possible to make an assessment, on the basis of the report, that the spectrum of patients in the cohort was representative of the patients whowould receive the test in practice. 21 studies of CSF or plasma Abeta reported conversion in 1128 patients. 26 studies of CSF tau reported conversion in 1054 patients. 22 structural MRI studies reported conversion in 1343 patients. 11 studies of FDG-PET reported conversion in 270 patients. 5 studies of PET based amyloid imaging reported conversion in 27 patients.Conclusions: The quality of reporting of longitudinal studies of biomarkers raises the possibility that substantial biases may be present. Multiple publication is commonplace in studies of diagnostic accuracy in dementia. Whilst this reflects developing techniques and continuing monitoring of subjects for outcomes, greater transparency is required so as to allow easier compilation of these data. A more formal process with cohort registration, reporting of ’intention to diagnose’ data, and clear independence of clinical and test evaluations is required. PET-based tests have the smallest evidence base. The small sample size of most studies limits their overall generalizeabilty. Funders, including those seeking licences for new products, should support large-scale, representative evaluations in that part of the population who would be most likely to utilize the test. Critical evaluation of the evidence base for diagnostic biomarkers is of major importance to the field of dementia. Without it, there is a risk that future clinical care and research will be built on assumptions about diagnostic validity that are wrong.