Newsletter · · Ashutosh Agarwal

The Alzheimer's Blood Test Just Walked Into Primary Care - The Neuro & Alzheimer's Pipeline - Week of August 16, 2026

The Neuro & Alzheimer's Pipeline for the week of August 10–16, 2026. A practicing family physician's CME talk becomes the week's one substantive signal: blood biomarkers lift primary-care diagnostic accuracy from roughly 40% to 90%, cracking the demand-side wall for symptomatic patients while screening the worried-well stays off the table.

The Neuro & Alzheimer's Pipeline

Week of August 16, 2026: The Alzheimer's Blood Test Just Walked Into Primary Care

For the second week running, the specialist neuro and biotech podcasts went quiet. No fresh Kisunla launch numbers, no Leqembi ramp update, nothing new on the muscarinic psychiatry story that dominated last week, and not a single sell-side desk episode on the space. But one episode this week is worth the whole issue on its own, and it lands directly on the single most important open question in this newsletter: is the Alzheimer's blood test actually being used, or is it a great science stuck behind a wall of doctors who won't order it?

Two weeks ago a practicing neurologist said, flatly, that he refuses to order the blood test in most patients. This week a front-line family doctor in a safety-net clinic said the opposite: he orders it routinely, and it has transformed how he diagnoses the disease. That is the tension the diagnostics thesis lives or dies on, and this week it got its clearest real-world data point yet, from the exam room, not a conference stage.

(Quick vocabulary, used throughout: amyloid and tau are the two toxic proteins that pile up in an Alzheimer's brain and are the target of today's approved drugs. pTau-217 and pTau-181 are two forms of tau you can now measure from an ordinary blood draw, the leading blood tests for the disease. MCI = mild cognitive impairment, the early, still-independent stage before full dementia. Amyloid-targeted therapy (ATT) is the class of drugs (Lilly's Kisunla and Eisai/Biogen's Leqembi) that clear amyloid from the brain and modestly slow decline; both are given by infusion and both carry a brain-swelling/bleeding risk called ARIA. A CME program is a certified continuing-medical-education talk doctors listen to for credit: teaching, not company news, which matters for how you weigh it.)

TL;DR

  • The blood test has arrived in primary care: the demand-side wall just cracked, at least for sick patients. A family physician says adding pTau-217 lifted the accuracy of spotting cognitive impairment in primary care from roughly 40% to 90%, and that he now orders these tests routinely in an ordinary safety-net clinic. This is the direct counterpoint to the neurologist two weeks ago who wouldn't order it at all.
  • But the wall didn't fall, it moved. The same doctor is emphatic that the test is only for people who already have symptoms, that you "probably won't get paid" unless you document cognitive impairment on the order form, and that the mass-screening-the-worried-well market still doesn't exist. Symptomatic use is spreading; screening healthy people is not.
  • Everything else stayed dark. No Kisunla or Leqembi commercial data, nothing from Roche, AbbVie, or the tau programs, no muscarinic/psychiatry follow-up, no PET-imaging or specific diagnostics-company news. A genuinely thin week with one rich signal.

What's new

Only one relevant episode surfaced across the whole podcast universe this week, and it is not a company or analyst show, it's a certified teaching program. But it's a good one, and it's an operator's-eye view of exactly the bottleneck this newsletter keeps circling.

A front-line family doctor says the blood test now works in primary care, and puts a number on it. On PeerView Neuroscience & Psychiatry CME/CNE/CPE Audio Podcast ("Minding Cognitive Health Across the Continuum," published August 14, also cross-posted to PeerView's Internal Medicine feed), Dr. Chuck Vega (OPERATOR/INSIDER: a practicing clinical professor of family medicine at UC Irvine who runs a safety-net clinic in Orange County) laid out how blood biomarkers have gone from research curiosity to something he uses in a regular visit. His headline claim: "Adding pTau-217 has increased the accuracy in identifying cognitive impairment in primary care from 40 percent to 90 percent." He frames this as the tool the field has wanted "for a long time," and stresses it's already reachable, not aspirational: "if it's available in my clinic in a safety net setting, then I bet you can get it in your clinics, too. So definitely start ordering biomarkers." Why it moves numbers: every prior week this newsletter has flagged the "demand-side wall," doctors who won't order the test. This is the first operator voice actively pushing the other way, and doing it in the least-resourced setting imaginable. If the ordering habit is diffusing into general practice, the diagnostics volume story gets a lot more real.

He described which tests he uses, and it's the pTau-217 ratio that leads. Vega said his preferred test is "the P-tau-217 to amyloid beta 42 ratio," because "it has a very strong sensitivity and specificity... you can be pretty sure that this is Alzheimer's disease," especially in patients with a higher pre-test probability. He contrasted it with a second test, pTau-181, which he described as a rule-out tool, "most valuable when it's negative... when it is negative, the chance that they have Alzheimer's disease is very, very low" (approved for ages 55 and up). And he flagged that the market is moving fast: "it's not something that's been changing even on a monthly basis. It seems to be changing on a weekly basis... this graph may change, and in fact it will change in the coming months as more testing becomes available." Why it matters: the pTau-217/amyloid-beta-42 ratio is the exact format C2N's PrecivityAD2 reports (my read-through, not his claim; he named no companies), so a primary-care doctor defaulting to the ratio as his go-to test is a quiet tailwind for the ratio-based players over single-analyte tests.

The reimbursement gate and the "symptomatic only" rule are still firmly in place. Vega was blunt about who qualifies and who pays: "these blood biomarker tests are not meant for folks who are asymptomatic. We don't use those," and "you probably won't get paid for [it] unless you have symptoms... so that's something you want to make clear on your order form, that they have cognitive impairment." He also walked through the referral funnel: a positive blood test sends the patient to a neurologist, and "if you're thinking about amyloid-targeted therapies, that's not going to be something that we offer in the primary care suite. That's going to be infusions that really require a specialist." Why it matters: this is the read-through map in one sentence: GP orders the blood test → positive result → neurologist referral → infusion. It's a volume funnel for the whole chain, but it's gated on symptoms and on documentation, which caps the near-term addressable population at the sick, not the worried.

The debate

Does under-the-skin dosing + cheap blood tests + broader coverage finally turn anti-amyloid (and eventually anti-tau) into a multi-billion-dollar franchise, or do modest benefit, brain-bleed risk, and diagnostic bottlenecks keep uptake weak while the tau bets stay unproven?

Bull: This is the best diagnostics data point the bulls have had in weeks, and it comes from the least glamorous, most telling place: a working family doctor in a safety-net clinic. If pTau-217 really moves primary-care diagnostic accuracy from 40% to 90%, and if ordering the test is spreading down-market to general practice rather than staying locked inside memory clinics, then the funnel that feeds the whole space (test, refer, infuse) is widening at the front end. Vega even predicted the last holdout guideline body (the US Preventive Services Task Force, which still doesn't back cognitive screening) "may change their approach" now that disease-modifying drugs exist. And he made the case for early diagnosis as a mission, not a chore, calling therapeutic nihilism something with "no place" in this disease.

Bear: Read the fine print and the wall is still there, it just moved. Vega is adamant the test is for symptomatic patients only, that you won't be reimbursed without documented impairment, and that mass screening of healthy people is off the table. That is the same limit the neurologist drew two weeks ago and the guidelines drew before that. So the "everyone over 60 gets a blood test" fantasy remains a fantasy. Worse for the near-term numbers: this was a teaching program, not a report of commercial traction: no volumes, no company names, no sales. And the drugs at the end of the funnel are still modestly effective infusions with a real brain-bleed risk that Vega himself flagged. Widening front end, unchanged back end.

Net: the bull case gained its first genuine operator-side evidence that the blood test is being adopted (a real crack in the demand wall) while the bear case kept its strongest brick: adoption is confined to symptomatic patients and gated by reimbursement. The re-rating catalysts are unchanged and still ahead: hard Kisunla and Leqembi sales prints, a guideline that green-lights broader testing, clean Medicare coverage, and the 2027 muscarinic Alzheimer's-psychosis data.

Stocks in play

No company was discussed by name or ticker this week; the one relevant episode was a clinician-education talk, not a company or analyst show. So everything below is read-through, not fresh company news. Flagged as such, not dressed up.

  • LLY (Eli Lilly) and BIIB (Biogen) / Eisai, read-through only. Vega referred to "two amyloid-targeted therapies... available" (that's Lilly's Kisunla and Eisai/Biogen's Leqembi), calling them "disease modifying" drugs that can give patients "potentially years back in terms of your cognitive function," while noting "there's definitely some risks associated" (ARIA). Bull: a primary-care doctor teaching peers to diagnose earlier and refer for these drugs is exactly the top-of-funnel behavior both franchises need. Bear: zero commercial data: no starts, no ramp, no infusion-capacity color; it's a call option in the P&L, same as the last two weeks. Next catalyst: actual Kisunla/Leqembi launch metrics (none this week). (PeerView, Aug 14)
  • Blood-diagnostics complex (C2N, Roche, QTRX / Quanterix, Fujirebio, LH / LabCorp, DGX / Quest), read-through only. Bull: a safety-net family doctor says the blood test lifts accuracy from ~40% to 90%, that he orders it routinely, and that his default is the pTau-217/amyloid-beta-42 ratio, the format C2N's PrecivityAD2 uses. Adoption spreading into general practice is the volume story these names need. Bear: symptomatic-only, reimbursement-gated, and no company-specific traction or numbers; the screening-the-healthy market still doesn't exist. Next catalyst: Medicare coverage and/or a guideline (watch USPSTF) that green-lights broader testing. (PeerView, Aug 14)
  • OPCH (Option Care Health) and infusion/specialty pharmacy, read-through only. Bull: Vega's funnel ends in specialist-administered infusions: every early diagnosis he makes is a potential infusion referral downstream. Bear: the volume is gated all the way up the chain (symptoms → referral → eligibility), so the flow-through is slow and indirect. Next catalyst: infusion-capacity or utilization data (none this week). (PeerView, Aug 14)
  • PET imaging (LNTH / Lantheus, GEHC / GE HealthCare), read-through only. Vega mentioned PET scanning as an alternative confirmatory test to blood biomarkers, particularly for tricky lower-probability cases. Bull: still the objective gold-standard confirmer. Bear: he clearly frames blood tests as the more efficient, patient-friendly first line; structurally, cheap blood work eats into the case for scanning everyone. Next catalyst: none named this week. (PeerView, Aug 14)
  • No coverage this week: Roche (trontinemab, gantenerumab successors, bepranemab/UCB tau); ABBV (AbbVie) neuroscience (emraclidine, Vyalev, Cerevel); the tau programs (E2814, remternetug, Ionis/Biogen's diranersin/BIIB080, ACI-35, TREM2, complement); BMY (Bristol Myers Squibb) Cobenfy and the muscarinic/Alzheimer's-psychosis story that led last week; MapLight and Neumora; and any specific CMS/coverage decision. None appeared in this week's podcasts; flagged, not fabricated.

Read-throughs

  • Diagnostics (C2N, Roche, QTRX, LH, DGX): the demand-side wall finally has a crack in it, but only on the symptomatic side. For two weeks this newsletter reported doctors refusing to order the blood test. This week an operator in the toughest possible setting said he orders it routinely and that it lifts his diagnostic accuracy dramatically. That's the first real-world adoption signal, and it matters because it came from primary care, not a memory clinic. But the boundary is unchanged: symptomatic patients only, with documentation, or you don't get paid. The investable read is that near-term volume comes from specialist and now GP work-ups of symptomatic patients, not from screening the well, and the pTau-217 ratio format (C2N-style) is the doctor's default, a small edge for ratio tests over single-analyte ones. (PeerView, Aug 14)
  • Prevention is quietly a bigger story than any single drug, and it's not a drug. Vega spent real time on the FINGER and POINTER lifestyle trials (POINTER completed in the US) where multidomain lifestyle programs (blood pressure, lipids, diet, exercise, cognitive and social activity) improved cognition, and where even the control group improved just from light coaching. He noted 45% of dementia risk is modifiable, and that US dementia incidence per 100,000 is actually falling thanks to better education and cardiovascular control. The read-through isn't a ticker; it's a reminder that the biggest lever on this disease may sit outside the pipeline entirely, which is a long-run governor on how large the drug-and-diagnostic TAM can ever get.
  • A useful caveat on the whole diagnostic funnel: 68/30. Vega noted ~68% of US dementia is Alzheimer's and ~30% is something else (vascular/multi-infarct, Lewy body, frontotemporal), for which these amyloid/tau tests come back negative. That's a natural ceiling on how much of the dementia population the blood tests and amyloid drugs can ever address, worth keeping in the back of any TAM model.
  • Muscarinics / psychiatry (BMY, MapLight, Neumora) and Lilly's GLP-1 engine: last week's two big threads got no follow-up at all this week, no read to give.

What changed vs last week

Last week the loudest neuro story was psychiatry (a muscarinic drug's schizophrenia miss and the setup for a 2027 Alzheimer's-psychosis fight) plus Lilly's monster GLP-1 quarter in which Alzheimer's was a footnote. This week, none of that recurred; those threads went silent.

What actually moved: the diagnostics debate flipped sides. Two weeks ago a practicing neurologist said he won't order pTau-217: patients panic, no guidelines on what to do with a positive. This week a practicing family doctor said he does order it, routinely, in a safety-net clinic, and put a hard number on the benefit (accuracy ~40% → 90%). On the surface that's a contradiction; underneath, the two reconcile into a sharper rule: both agree you don't screen asymptomatic, worried-well people, but they diverge on symptomatic patients: the neurologist hesitates, the GP orders. The net is that real-world adoption for symptomatic diagnosis is spreading further down the care chain than the last two weeks implied. That's the most citable change of the week.

What we did not get this week (the honest gap list): any Kisunla launch metrics or infusion-capacity color: nothing (just the generic "two ATTs are available" framing); any Leqembi/IQLIK sub-Q ramp or international numbers: nothing; any Roche pipeline update (trontinemab, gantenerumab successors, bepranemab/UCB): nothing, again; anything on AbbVie's neuroscience assets (emraclidine, Vyalev, Cerevel): nothing; any tau-program news (diranersin/BIIB080, E2814, remternetug, ACI-35, TREM2): nothing; any specific diagnostics-company update (Quanterix, C2N, Roche Elecsys, LabCorp, Quest): only the generic blood-biomarker discussion above; PET imaging (Lantheus, GE HealthCare): only a passing mention; a hard CMS/coverage decision: nothing; the muscarinic/Alzheimer's-psychosis story (BMY, MapLight, Neumora) that led last week: nothing; Parkinson's and Huntington's: nothing; infusion/specialty pharmacy operator data (OPCH): nothing. It was a diagnostics-adoption week and nothing else. The referees that matter are all still ahead: a Medicare coverage decision, a guideline expansion (watch the USPSTF), hard Kisunla and Leqembi sales prints, and the 2027 muscarinic Alzheimer's-psychosis data.