Digest — 2026-09-27, The Clinical
One thing crosses the escalation bar, and it closes out the item the last two digests on this device left open: what Abbott's Amulet IFU change actually says.
1. Abbott's pulmonary-artery warning on the Amulet, in full — with numbers you can use
The 2026-09-13 digest told you Abbott had revised the Amulet IFU to address a pulmonary-artery-injury risk, but that the actual wording hadn't been read — only the FDA's one-line paraphrase. It has now been read in full, via Abbott's own field safety notice (reference FA-Q326-SH-1), corroborated independently by the UK's MHRA and by Abbott's own US advisories page. All three agree.
This is about the Amulet you're implanting today — not Amulet 360. The affected model numbers (9-ACP2 series, the full 16–34 mm size range) are the current, commercially available device. Amulet 360 isn't mentioned anywhere in the notice and shouldn't be assumed to inherit this until it has its own published IFU.
What's changing: the contraindication about obstructing intracardiac/intravascular structures now specifically names the pulmonary artery, and there's a new pre-deployment step — measure the distance from the landing zone to the pulmonary artery. Abbott's own numbers for that assessment:
- ≤ 1.5 mm — significant risk
- 1.5–3.0 mm — theoretical risk
- > 3.0 mm — lower risk
If it's close, consider a different landing zone or orientation; if the risk can't be brought down enough, Abbott's own language says to weigh whether to proceed at all. This is a labeling correction, not a recall — nothing needs to come back.
One retrieval gap worth knowing about: a copy of the same notice was not found on HRS's site (their safety-alerts page blocked the fetch). That's a gap in this corpus's coverage, not a sign HRS doesn't have one — the IGJ (Dutch regulator) and MHRA copies were enough to corroborate this fully on their own.
→ substrate/clinical--amulet-360
Also written this run, below the bar
A LINQ II lead from last week's cardiac-AI sweep turned out to be nothing. Medtronic's LINQ II ICM got a new FDA clearance in late August, and because LINQ II's AccuRhythm AI is a named comparator in your ICM false-positive notes, it was worth a look. Read the actual FDA paperwork: it's a manufacturing change (a new material used inside the device), not a software or algorithm update. Doesn't affect anything you already know about LINQ II's data-burden claims.
→ substrate/clinical--icm-false-positive-evidence
A cardiac MRI measurement tool (Heartvue.Proton) got FDA clearance in August — not your area of day-to-day practice, but worth flagging once: it's the third cardiac-adjacent AI tool this corpus has now seen get classified under a general-purpose FDA category (Radiology, in this case) rather than a cardiovascular-specific one. If you ever need to argue that "not on the cardiovascular AI list" doesn't mean "not cardiac AI," this is a clean example.
→ substrate/clinical--siu-cardiovascular-outcome-prediction
A remote ECG patch from a small company (Peerbridge, "Cor MDx") was checked and set aside. Thin information, no confirmed AI angle, no obvious relevance to your work — recorded as checked rather than silently dropped, but not written up as its own item.
Two HeartMate/CentriMag corrections showed up on Abbott's own advisories page while looking for the Amulet notice (a May 2026 battery-warning issue, an October 2025 correction). Mechanical circulatory support isn't in this corpus's charter scope, so nothing was written up — flagging only so it isn't a silent gap if it ever becomes relevant to you.
Instruments
The "download the PDF, then read the saved file directly" workaround worked twice more this run — once for the LINQ II clearance letter, once for the Amulet notice itself. Both times, the automated web-fetch tool could download the file but couldn't read the text out of it; reading the same saved file with a PDF-capable reader worked cleanly both times. This is now a reliable fallback whenever an FDA or manufacturer PDF won't extract normally.