Research report
A research report is the sourced material a domain
dossier is synthesized from — generated on a plan and a cadence, one topic per file.
A report carries no confidence tags. Its bracketed markers say who might have
an incentive to shade a line; none of them says anyone checked it. To reach a score, a line has
to be drafted onto a candidate as unverified and pass the Verifier or the
Corroborator, like everything else.
Section technology · Version 2026-09-01 · Cadence annual · Evidence class mixed · Sources 8 · Supersedes none
Domain: veterinary-diabetes-glucose-monitoring · Scope: public primary sources reachable by this repo's connectors, WebSearch and WebFetch. Excludes CPT descriptors (AMA-licensed), veterinary datasets (no US federal dataset exists), fda.gov guidance documents (intermittent bot mitigation), and any jurisdiction other than US.
Sourcing: two FDA 510(k) clearance records and one linked recall establish the timeline and defect history of the human CGM generation this domain's off-label incumbent draws on; five peer-reviewed studies (two human, three veterinary/production-animal) describe that generation's accuracy and its spread into new clinical windows and species — every quantitative line rests on one source
The domain dossier already covers the 1962/1967 enzyme-electrode history and the 2020-2023 clearance cadence for product code QLG. What this report adds is what happened after that window closes and where the same sensing technology is being pushed next. FDA cleared a 15-day, vitamin-C-interference-resistant successor sensor (FreeStyle Libre 3 Plus, K233537) in 2024, two years after the original Libre 3 (K212132, 2022) — and by late 2025 both generations were caught in the same carbon-buildup recall for false-low readings. No veterinary study of the newer 15-day generation was found; every veterinary accuracy figure on record still describes the older 14-day-labeled sensor. Separately, the same off-label device is being tested in a new clinical window — perioperative glucose monitoring in dogs — and the broader class of continuous interstitial-glucose sensors is being piloted in a different species and indication entirely (dairy calves, metabolic transition monitoring, not diabetes), with anatomical placement site itself now a studied variable.
_Baseline (v1). No prior version; this establishes the starting point for future diffs._
FDA cleared the original FreeStyle Libre 3 Continuous Glucose Monitoring System under K212132 on 2022-05-26 [1]. Two years later, on 2024-04-23, FDA cleared K233537, adding the FreeStyle Libre 3 Plus Sensor alongside the existing Libre 3 and Libre 2 Plus lines [2]. A peer-reviewed multicenter accuracy study of that newer sensor describes what changed in the hardware [3]:
"This study evaluates the performance of a 15-day factory-calibrated continuous glucose monitoring sensor used in FreeStyle Libre 2 Plus (Libre 2 Plus) and FreeStyle Libre 3 Plus (Libre 3 Plus) Systems, featuring an improved sensor design to reduce vitamin C interference." — PMID 40183340 (332 enrolled participants, 149 adults and 124 pediatric with paired YSI-reference data, seven US sites, 2026)
Reported accuracy in that human study was 94.0-94.2% of values within ±20 mg/dL/20% of the YSI reference and MARD 8.1-8.2% (adults and 6-17-year-olds) [3] [single-source]. A separate multicenter study of the standard 14-day Libre 3 (100 participants, four US sites) reported MARD 7.8% and 93.4% within ±20 mg/dL/20% against the same YSI reference [4] [single-source]. For comparison, the veterinary evaluation of the same third-generation sensor already on record in this domain's dossier — 20 client-owned diabetic cats, 288 paired readings against AlphaTrak2 — reports MARD 13.6% [5] [single-source]; median sensor lifespan in that feline cohort was 11 days against a 14-day label, and no veterinary study of the newer 15-day Plus generation named in [2] was found in a targeted search ("Libre 3 Plus veterinary dog cat," "Libre 2 Plus dog cat sensor," zero PubMed hits both).
The newer generation did not escape the sensor-manufacturing problem the dossier already records for the older one. Recall Z-1021-2026, initiated 2025-11-24 and tied specifically to K233537 (the Libre 3 Plus clearance), gives the identical reason already on record for the older Libre 3 sensor:
"Due to a manufacturing process issue that may cause carbon build up in the sensor that may result in incorrect low glucose results." — Recall Z-1021-2026 [6]
Beyond home glycemic-curve monitoring, a 2025 feasibility study applied the same off-label Libre device to perioperative glucose monitoring in 32 dogs undergoing general surgery — a use this domain's dossier does not yet describe — comparing 224 paired readings against a portable blood glucose meter across the pre-, intra- and post-operative phases [7] [single-source]. The device tracked the comparator closely in all three phases (Spearman's rho 0.89, 0.82 and 0.80 respectively) but the direction of its bias flipped specifically during surgery:
"FGMS provides continuous and noninvasive glucose monitoring in dogs with clinical feasibility throughout the perioperative period, although its accuracy decreases during the intraoperative phase." — PMID 40461420 (32 dogs, 224 paired FGMS/portable-meter readings, 2025)
Bland-Altman bias was -16.0 mg/dL preoperatively and -4.6 mg/dL postoperatively, but +23.3 mg/dL intraoperatively — all data pairs still fell within Parkes zones A+B throughout [7]. This is a different physiological window than any of the dossier's existing accuracy record and a different, phase-specific failure mode than the ones already catalogued there (stress hyperglycemia, sensor detachment, the measurement-range ceiling).
Outside the domain's stated companion-animal, diabetic population, a 2026 study evaluated continuous interstitial glucose sensors (brand not named in the retrieved abstract) in six male dairy calves aged 15-38 days, for a non-diabetic indication — feeding-practice and metabolic transition monitoring, not insulin titration — with sensors placed simultaneously at five different anatomical sites (ear base, neck, thorax, rump, inner thigh) [8] [single-source]:
"Sensors placed on the rump and thorax demonstrated the longest functional lifetimes, along with a constant negative bias (intercept) and a positive proportional bias (slope) compared with the reference standard." — PMID 42508481 (6 calves, jugular blood glucose by hexokinase method 8x/day, Passing-Bablok regression, 2026)
Read narrowly: this is a different species, a different clinical indication, and a device not identified as FreeStyle Libre or any product named elsewhere in this domain. It is included here only as evidence that the underlying sensing technology class — and anatomical placement site as a design variable — is actively being studied outside this domain's core population, not as a finding about dogs, cats, or diabetes.
[1] 510(k) K212132 — FreeStyle Libre 3 Continuous Glucose Monitoring System, Abbott Diabetes Care, Inc. — FDA openFDA device/510k database (decision 2022-05-26; accessed 2026-09-01). K212132 — https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpmn/pmn.cfm?ID=K212132 [federal-registry] [2] 510(k) K233537 — FreeStyle Libre 3 / FreeStyle Libre 3 Plus Continuous Glucose Monitoring System, Abbott Diabetes Care, Inc. — FDA openFDA device/510k database (decision 2024-04-23; accessed 2026-09-01). K233537 — https://www.accessdata.fda.gov/cdrh_docs/pdf23/K233537.pdf [federal-registry] [3] Accuracy of a 15-day Factory-Calibrated Continuous Glucose Monitoring System With Improved Sensor Design — Journal of Diabetes Science and Technology (published 2026-07; accessed 2026-09-01). PMID 40183340 — https://pubmed.ncbi.nlm.nih.gov/40183340/ [peer-reviewed] [4] Accuracy of the Third Generation of a 14-Day Continuous Glucose Monitoring System — Diabetes Therapy (published 2023, month not given in the PubMed record; accessed 2026-09-01). PMID 36877403 — https://pubmed.ncbi.nlm.nih.gov/36877403/ [peer-reviewed] [5] Evaluation of the FreeStyle Libre 3 in client-owned diabetic cats — Journal of Veterinary Internal Medicine (published 2026, month not given in the PubMed record; accessed 2026-09-01). PMID 41742494 — https://pubmed.ncbi.nlm.nih.gov/41742494/ [peer-reviewed] [6] Recall Z-1021-2026 — FreeStyle Libre 3 Plus Sensor, Abbott Diabetes Care, Inc. — openFDA device/recall database (initiated 2025-11-24, status Open, Classified; accessed 2026-09-01). Z-1021-2026 — https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfRES/res.cfm [federal-registry] [7] Perioperative use of flash glucose monitoring system in dogs undergoing general surgery: a clinical feasibility study — Journal of Veterinary Science (published 2025, month not given in the PubMed record; accessed 2026-09-01). PMID 40461420 — https://pubmed.ncbi.nlm.nih.gov/40461420/ [peer-reviewed] [8] Evaluation of continuous interstitial glucose monitoring in healthy calves — Journal of Dairy Science (published 2026, month not given in the PubMed record; accessed 2026-09-01). PMID 42508481 — https://pubmed.ncbi.nlm.nih.gov/42508481/ [peer-reviewed]
Well established: nothing in this report rests on two independent, non-affiliated sources. The two 510(k) records [1][2] and the recall record [6] are registry facts stated plainly, per this fleet's convention that a single named federal record does not need a [single-source] marker; every peer-reviewed accuracy or feasibility finding is a single study and is marked accordingly.
Thin: all five PubMed sources [3][4][5][7][8] are [single-source] — each is one study, none independently corroborated by a second paper in this pass. The calf study [8] additionally names no device brand in the abstract retrieved, only "the sensor"; that identification gap is stated in §3 rather than resolved by inference.
Rescoped from class 3: none in this report. Every proposition here is a registry fact (a 510(k) clearance, a recall) or a published measured finding, not a question about what a person would do.
Out of scope: CPT descriptors, not reproduced anywhere in this repo; fda.gov guidance documents, excluded for intermittent bot mitigation; any non-US jurisdiction, including any EU/CE-mark equivalent clearance record for the newer Libre generation, since this domain's stated jurisdiction is US. The calf study [8] and dog-surgery study [7] both sit partly outside the domain's stated population (diabetic dogs and cats, at-home or in-clinic monitoring) — [8] is a different species and a non-diabetes indication, and neither the diabetic status of [7]'s surgical dogs nor a home/community setting is established. Both are included as technology-trajectory context, named as adjacent rather than folded into the domain's core accuracy or care-pathway record.
Not searched vs. not found: Not found — a veterinary (dog or cat) accuracy study of the 15-day Libre 2 Plus / Libre 3 Plus generation cleared under K233537; a PubMed record for "GluCurve" under any phrasing tried. Not resolved this pass — the patent landscape: running python3 -m connectors patent returned unconfirmed, reason "EPO_OPS_KEY / EPO_OPS_SECRET not set" — a different named credential than the PATENTSVIEW_API_KEY the domain dossier's open-questions table names for the same blocked item. Both are unset; this report cannot say which the connector actually requires without an environment change, so the item stays blocked rather than searched. A web search (not a connector, and not treated as a source of record) for a GluCurve-specific hardware patent found only a 2023 company announcement of a "notice of allowance" for an unrelated human-diabetes-product algorithm, with no patent number disclosed and no stated connection to the pet CGM hardware — seen and not used, not recorded as a finding. Full-text review of either 510(k) summary PDF beyond the Indications for Use statement already quoted was also not performed.
[inference] None of the numeric comparisons in §3 — human MARD (7.8-8.2%) against the feline MARD (13.6%) reported for the same third-generation sensor, or the pre/intra/postoperative bias figures in the dog-surgery study — are combined into a causal or predictive claim by this report; they are placed adjacent to let a reader compare them, which is a presentation choice worth flagging rather than a finding in itself.
Propositions a checker could attempt. This report does not append them — /research drafts them onto a candidate with engine/scripts/append_research_claims.py, which constructs the unverified tag itself.
| Proposition | Evidence class | Resolvable identifier | Dossier section |
|---|---|---|---|
| FDA cleared the original FreeStyle Libre 3 Continuous Glucose Monitoring System under K212132 on 2022-05-26 | 1 registry | K212132 | 7 |
| FDA cleared the FreeStyle Libre 3 Plus Sensor (a 15-day, factory-calibrated successor sensor with improved sensor design to reduce vitamin C interference) under K233537 on 2024-04-23 | 1 registry | K233537 | 7 |
| A multicenter human accuracy study of the Libre 2 Plus/3 Plus 15-day sensor (332 enrolled, 149 adults + 124 pediatric with paired YSI data) reported MARD 8.1-8.2% and 94.0-94.2% of values within ±20 mg/dL/20% of the YSI reference | 2 published | PMID 40183340 | 7 |
| Recall Z-1021-2026, tied to 510(k) K233537, cites a manufacturing process issue causing carbon buildup that may produce incorrect low glucose readings in the FreeStyle Libre 3 Plus Sensor | 1 registry | Z-1021-2026 | 5 |
| A 2025 feasibility study in 32 dogs undergoing general surgery found flash glucose monitoring system bias of -16.0 mg/dL preoperatively, +23.3 mg/dL intraoperatively, and -4.6 mg/dL postoperatively against a portable blood glucose meter, with 100% of paired readings in Parkes zones A+B throughout | 2 published | PMID 40461420 | 7 |
| A 2026 study of continuous interstitial glucose sensors placed at five anatomical sites in 6 dairy calves found the rump and thorax sites gave the longest functional sensor lifetimes | 2 published | PMID 42508481 | 7 |