Blood Markers Promising in Alzheimer's

by John Gever for MedPage Today:Blood tests may soon be a realistic screening tool for Alzheimer's disease, with two research groups reporting they had found multi-marker panels that distinguished individuals with the disease from healthy people.In one of the two reports, both published online in Archives of Neurology, Ralph N. Martins, PhD, of Edith Cowan University in Joondalup, Australia, and colleagues found that a set of 18 markers in blood had sensitivity and specificity of more than 80% for distinguishing patients with Alzheimer's disease from healthy controls.The other report, by Holly Soares, PhD, of Bristol-Myers Squibb in Wallingford, Conn., and colleagues, indicated that a mostly different set of markers were increased in patients with Alzheimer's disease and mild cognitive impairment.Also, the second study found distinct biomarker profiles that depended on individuals' APOEgenotype status "irrespective of diagnosis," Soares and colleagues wrote.Both groups concluded that their findings were a significant step toward a screening test for Alzheimer's disease using only a blood sample.Currently, Alzheimer's disease is diagnosed on the basis of clinical symptoms, which can be backed up with PET brain scans or biomarkers in cerebrospinal fluid (CSF). Clinical symptoms are notoriously nonspecific and the confirmatory tests are expensive and, in the case of CSF markers, painfully invasive.Although there is currently no treatment available to stop or slow Alzheimer's disease progression, a blood test may still be useful for differential diagnosis and perhaps for monitoring treatment response in drug trials.The study by Martins and colleagues began with a discovery phase, involving blood samples from 207 patients diagnosed with Alzheimer's disease and 754 healthy controls in the Australian Imaging Biomarker and Lifestyle cohort.Starting with 27 potential biomarkers identified in earlier studies, the researchers used data from the discovery cohort to pick out 18 that produced the best specificity and sensitivity for distinguishing patients from controls.The 18 markers included some that were increased and others that were decreased in patients relative to controls. Most were proteins, such as pancreatic polypeptide, apolipoprotein E (theAPOE gene product), and interleukin-17, but the panel also included calcium and zinc.Martins and colleagues then sought to validate the panel in a different cohort, consisting of 112 Alzheimer's disease patients and 58 healthy controls in the Alzheimer's Disease Neuroimaging Initiative (ADNI).To evaluate the results' robustness, the researchers used two different validation procedures. Both of them yielded similar results, with both specificity and sensitivity in the range of 80% to 84%. Areas under the receiver-operating characteristic curves ranged from 0.84 to 0.87 -- generally acceptable for a screening test to be used in high-risk populations.Interestingly, the best accuracy in the validation sets was obtained when Martins and colleagues used data only on eight of the markers. The researchers indicated that the optimal mix of markers -- which could include others not examined in their study -- remains to be determined.Soares and colleagues started with data from the same ADNI cohort of 112 patients and 58 controls, but ended up with disease-associated markers that mostly did not overlap with those identified by Martins and colleagues.The only markers in common were pancreatic polypeptide and apolipoprotein E. Others identified by Soares and colleagues included eotaxin 3, N-terminal pro-brain natriuretic peptide, tenascin C, immunoglobulin M, which did not figure in the other group's panel.Depending on cutoffs used, sensitivity ranged from 80% to 90% with specificities up to 80%, Soares and colleagues indicated.The researchers also examined marker levels across diagnostic groups among individuals with different APOE genotypes. Not only did apolipoprotein E levels differ -- highest in those with theAPOE2/3 genotype and lowest in those homozygous for genotype 4 -- but so did levels of C-reactive protein, interleukin-13, apolipoprotein B, and interferon-gamma.Soares and colleagues suggested that this finding could help explain "the biological variability of blood-based biomarkers of Alzheimer's disease described in the literature," which has frustrated earlier efforts to develop Alzheimer's disease blood tests."Clearly, examination in much larger cohorts is needed," they wrote.They cautioned that the findings might be confounded by cardiovascular disease, which was not evaluated in the study, insofar as several of the markers may reflect vascular pathologies. They also did not perform extensive sensitivity analyses on individual markers and noted that some may be redundant.Also, a limitation of both studies is that neither examined biomarker levels in patients with other types of dementia or neurodegenerative conditions that may be confused with Alzheimer's disease.Source: http://www.medpagetoday.com/Neurology/AlzheimersDisease/33769