Somewhere in a lab in Gothenburg, an envelope arrives. Inside is a small card, spotted with a few drops of blood that dried hours or days earlier at a clinic in Barcelona, Brescia, Copenhagen or Exeter. No dry ice, no centrifuge, no minus 80 freezer, no courier chasing a cold chain. That card is the whole experiment: the DROP-AD project wanted to know whether Alzheimer's blood biomarkers survive being pricked out of a fingertip, blotted onto paper and posted. The answer, published in Nature Medicine on 5 January 2026, is a qualified yes, and the qualifications matter as much as the yes.
The study enrolled 337 participants across seven European centers, spanning cognitively unimpaired adults, people with mild cognitive impairment, people with Alzheimer's dementia, people with non-Alzheimer's dementias and adults with Down syndrome. Of those, 304 gave paired samples: a finger prick onto a dried plasma or dried blood spot card, and a conventional venous draw processed the old careful way. Three markers were measured on the Simoa platform: p-tau217, the phosphorylated tau fragment that has become the most reliable blood flag for amyloid pathology; GFAP, released by reactive astrocytes; and neurofilament light, a general signal of nerve fiber damage. Capillary p-tau217 tracked venous p-tau217 with a Spearman correlation of 0.74, rose stepwise with disease severity, and predicted cerebrospinal fluid biomarker positivity with an area under the curve of 0.864. GFAP and NfL also correlated strongly with their venous counterparts. This is a cross-sectional method comparison, not a trial and not a follow-up study, and the paper's own conclusion is that the approach is ready for research, not for managing patients.
The part that makes this more than another assay story is who else was in it. Adults with Down syndrome carry three copies of chromosome 21 and therefore three copies of the amyloid precursor protein gene, which is why Alzheimer's pathology is near universal with age in this population; a separate Nature Communications study of 2,329 samples from Barcelona cohorts, published in November 2025, showed plasma p-tau217 separating cognitively stable adults with Down syndrome from those with Alzheimer's dementia with an area under the curve of 0.96. The obstacle has rarely been the biology. It is the tourniquet. Venipuncture can be genuinely difficult, distressing or repeatedly unsuccessful, and a test that requires a good vein quietly excludes the people who need it most. In DROP-AD, 31 participants with Down syndrome gave dried spot data, with a mean age around 48, and capillary p-tau217 correlated with venous p-tau217 at r = 0.875, GFAP at r = 0.629. Capillary levels of both were higher in those with dementia than in those without Alzheimer's related cognitive impairment. No NfL results came from this group, because the card type needed for that marker was not collected there.
The second scalability question was whether anyone has to be in the room at all. Exeter was the only site to test self-collection, recruiting from the PROTECT-UK study, and the paper reports high concordance between supervised and self-collected samples. A follow-up in Nature Communications on 6 May 2026, from the same PROTECT cohort, took that further in 174 people who were cognitively normal or had mild cognitive impairment or Alzheimer's disease, reporting that self-administered finger prick biomarkers correlated with venous measures and with cognition and function. For a family in a rural county, or a carer who cannot get a relative into a phlebotomy chair without a bad afternoon, that is the difference between a research study existing on paper and existing for them.
Now the qualifications. To get a workable volume from a finger prick, the plasma had to be diluted tenfold, and capillary concentrations run far below venous ones, which is why numbers from a card cannot be read against thresholds built for venous plasma. Commentators on the study have pressed exactly this point: a correlation of 0.74 is a strong relationship, not interchangeability, and it says little about people sitting close to a decision cut-off. Diagnostic performance illustrates the gap, since the discrimination that reaches roughly 0.98 with venous plasma fell to around 0.86 with the dried spot. The authors and independent commentators converge on the same near-term use: population research, epidemiology and trial prescreening, where Henrik Zetterberg suggested cards combined with remote cognitive testing could "dramatically reduce screen failure rates". The authors also note that a capillary NfL measure, if refined, would matter well beyond Alzheimer's, in frontotemporal dementia, atypical parkinsonian syndromes and other neurological conditions.
What a card in an envelope actually changes, for now, is who gets counted. Every large biomarker cohort assembled so far has been shaped by proximity to a well equipped hospital, and every population excluded from that geography, or from venipuncture itself, has been excluded from what we know. This study does not put a diagnosis in the letterbox. It makes the case that the letterbox is a plausible piece of research infrastructure, and that adults with Down syndrome need not be the last to be studied because their veins are the hardest to find.