Glucose monitoring in insulin-treated diabetic dogs and cats

Insulin dosing in diabetic dogs and cats has always depended on a glucose curve, and the question this domain holds context on is what constructing that curve actually costs — in stress, in staff time, in sensor failures — and what practices and owners have already done about it, since no FDA-cleared veterinary CGM predicate exists to structure the question around a device instead. Nearly all of the useful evidence here is about the two things people already use: in-hospital serial sampling, and off-label human CGMs (chiefly Abbott FreeStyle Libre) applied to dogs and cats. Read the thinness before the content: sections 2 and 3 are essentially unsearched, epidemiology's best US-specific figures are 25+ years old (though a second research pass added two recent, large-scale non-US comparators — §1), and no connector in this repository reaches veterinary population, pricing, or payer data at all — every quantity below that isn't a PubMed record or an openFDA record was retrieved by following a web search to the primary document it pointed at, per the source hierarchy this repo requires, and several attempts to do that (VIN, AAHA, Today's Veterinary Practice, NAPHIA) returned 403s or bot-verification walls and are recorded honestly as not retrieved rather than papered over. Section 6 (reimbursement) was searched hard in the second pass and came back genuinely negative on the one open question (pet-insurance device coverage) rather than unsearched — that distinction is recorded in §6 itself.

This file carries no confidence tags. Every line below is context with a source attached. To affect a score, a line must be copied into a candidate's ## Claims as unverified via engine/scripts/append_research_claims.py and pass the Verifier or Corroborator.

1. Condition & epidemiology

as_of: 2026-08-26 (refreshed same-day, second research pass) · sources searched: PubMed via literature/literature-detail, this pass targeting recent (post-2010) and larger-scale/insurance-claims-based prevalence estimates specifically · not searched: AVMA member surveys, Banfield/BluePearl "State of Pet Health" reports (no connector reaches them; a targeted web search for them, both in the first pass and again this pass, returned only unrelated PubMed results — human-diabetes systematic reviews that happen to share the word "Banfield" or "prevalence" — and market-research-firm summaries, never the reports themselves)

Verbatim, for anything load-bearing:

"The prevalence of DM in dogs presented to veterinary teaching hospitals increased from 19 cases per 10,000 admissions per year in 1970 to 64 cases per 10,000 in 1999, while the case-fatality rate decreased from 37% to 5%." — PMID 12672370

"Hospital prevalence increased from eight cases per 10,000 in 1970 to 124 per 10,000 in 1999 (P < 0.001). Case fatality percent at first visit decreased from 40% to 10% (P < 0.001)." — PMID 17449313

"Annual prevalence was 0.39% (95% confidence interval [CI]: 0.37-0.42). Incidence risk (2019) was 0.14% (95% CI, 0.13-0.16)... Almost 1/250 cats in the UK live with DM annually." — PMID 40525629 (n=1,225,130 cats under UK primary veterinary care, VetCompass, 2019)

"The observed prevalence rose from approximately 0.07% in 2015 to 0.30% in 2023, highlighting a steady upward trend... an odds ratio of 1.128 (95% confidence interval: 1.116-1.142), indicating an annual 12.8% increase in the odds of diabetes." — PMID 41443822 (Japanese pet-insurance dataset, dogs, 2015-2023)

Gaps, updated this pass: the two US figures remain hospital-presented prevalence at US veterinary teaching hospitals (the "Veterinary Medical Data Base," VMDB) through 1999 — not a current population-level prevalence estimate, and not general-practice prevalence. A second, targeted search this pass (queries in §10) specifically for a post-2010, larger-scale, or insurance-claims-based US prevalence estimate again found none — the two new prevalence studies located (PMID 40525629, PMID 41443822) are recent, large-scale, and methodologically stronger than the 1999-cutoff VMDB studies, but both are non-US (UK primary-care EHR data, Japanese pet-insurance claims data respectively) and are recorded here as comparators, the same way the existing Swedish incidence studies are, not as substitutes for a US figure. This remains the single largest gap standing behind candidate claims 7 and 8(a): what exists for the US specifically is a dated, hospital-only proxy; what exists more recently and at scale exists only for the UK and Japan.

2. Care pathway & clinical setting

as_of: 2026-08-26

Mostly not searched. One finding bears directly on site-of-service: in-hospital glucose curves in cats are confounded by stress-induced hyperglycemia, and cats need an acclimatisation period before results are reliable.

The one finding this section can state with a source: stress hyperglycemia in a hospital setting is a documented, measured confound on the in-clinic glucose curve, not merely an assumed one, which is directly relevant to why a device readable at home has appeal independent of any comfort claim.

"Stress-induced hyperglycaemia was observed in the hospital. The acclimatisation cut-off time for the cats' return to euglycaemia was 6 h... Elevated HR, RR, CSS, glucose and cortisol concentrations were observed in cats at the hospital compared with those at home (P <0.001)." — PMID 40230085 (10 healthy cats, blood glucose measured hourly over 10h, at home and during hospitalisation)

Gaps: no source establishes where the majority of US diabetic-pet monitoring actually happens today (general practice vs. specialty/internal-medicine referral), typical visit frequency, or staffing/throughput implications of either sampling method. Not searched.

3. User groups & incentives

as_of: 2026-08-26

Mostly not searched, but the owner-facing row has real sourced content — see §5, which is where this domain's demand-adjacent evidence actually lives.

Gaps: no source separates veterinary-technician incentives from veterinarian incentives; no source addresses purchaser-side clinic economics (does the clinic mark up the sensor, sell it near cost, or absorb it).

4. Standard of care & clinical evidence

as_of: 2026-08-26 · sources searched: PubMed via literature/literature-detail, 13 PMIDs read at abstract depth · not searched: AAHA's 2018 Diabetes Management Guideline full text (aaha.org returned HTTP 403 to an automated fetch — recorded as not retrieved, not as "guideline is silent")

Read the fifth row carefully — it is exactly the title-vs-conclusion trap this repo's own backlog names: a title reading "limited analytical accuracy" sits over a conclusion that qualifies it heavily ("rarely produced ... detrimental clinical decisions"). Both halves are recorded above rather than only the title.

Verbatim:

"The FSL3 underestimates BG across most of the hypo-euglycemic range but overestimates BG in marked hypoglycemia (< 55 mg/dL). Recognizing the proportional, glycemic-dependent bias of FSL3 improves the safety of its clinical application in cats." — PMID 40105400

"In this model of rapidly induced hypoglycemia, the FGMS demonstrated limited analytical accuracy compared to the reference standard. However, it rarely produced glucose readings likely to lead to detrimental clinical decisions." — PMID 40280161

Gaps: no guideline position independently retrieved. No study found comparing owner-home FGMS-based titration outcomes (glycemic control, hospitalization rate) against clinic-curve-based titration head-to-head with a clinical (not just analytical) endpoint. No canine analogue of the feline hyperinsulinemic-clamp accuracy studies was found — the clamp-based hypoglycemia-accuracy work above is cats-only; dog accuracy evidence is field-cohort (33562672, 40280161), not clamp-controlled.

5. Reference products — and why people stop using them

as_of: 2026-08-26 (refreshed same-day, second research pass — added a per-unit-price second-source search) · sources searched: the PMIDs in §1 and §4, plus openFDA recalls on product code QLG, plus company/clinic web sources followed to the primary page · not searched: MAUDE (no connector in this repo, docs/backlog.md 1.9); commercial-distribution status of any specific human CGM SKU in the veterinary channel

This domain has two incumbents worth separate abandonment records: the practice being displaced (serial blood-glucose-curve sampling) and the off-label product already displacing it in some practices (human CGMs, chiefly Abbott FreeStyle Libre).

Serial blood-glucose-curve sampling (venipuncture / lancet, in-clinic or at-home)

• What it is: repeated blood glucose measurements over 12-24h to construct a glycemic curve for insulin-dose titration; historically performed in-hospital.
• Price / what it realizes per unit: not established here (no fee-schedule or billing study found; CPT is licensed and out of scope regardless — see §6).
• Evidence behind it: §4 (comparator arm in every accuracy study above; also PMID 40230085 on the confound it is prone to).
• Adoption: the historical standard of care; not quantified as an installed base because it is a practice, not a product.
• Abandonment record — why people stop:

• Verbatim:

"More than 80% of DPOs considered FGMS easier to use and less stressful and painful for the animal compared to blood glucose curves (BGCs). Overall, 92% of DPOs reported that their pet had better diabetes control since using FGMS." — PMID 36977242

Read this the same way the potassium dossier reads its own capillary-sampling section: an owner's stated preference for the alternative they already switched to is real, retrieved, revealed-behaviour evidence about the incumbent's burden. It is not proof that a new device (this candidate's) would be adopted — it is evidence about what people already did with the alternative that already exists (FreeStyle Libre), which is the closest available twin, not a stand-in for validating a not-yet-built product.

Off-label human CGMs in veterinary use (chiefly Abbott FreeStyle Libre 1/2/3)

• What it is: a human-labeled, human-510(k)-cleared flash/continuous glucose monitor (product code QLG, regulation 862.1355) applied off-label to a dog or cat, typically over the dorsolateral thorax.
• Price / what it realizes per unit: one veterinary clinic's own published pricing states "$75.08" per 14-day sensor for pet use — a single-clinic data point, not a systematic survey (craigrd.com, Craig Road Animal Hospital blog, retrieved directly, no date stamp on the page itself). Separately, Abbott's own consumer cost page states insured human patients typically pay $0-$20/sensor and cites a copay-card threshold of "over $75 for two sensors" for uninsured human patients — a **human retail/insurance figure, not a veterinary one**, included here only as the source the vet-channel figure should be read against, per the instruction not to port human pricing over without basis. Second-source attempt, this pass: searched specifically for a corroborating or contrasting veterinary per-unit price. Found none citable. Two US clinic pages (accessveterinarycare.com/freestyle-libre, retrieved directly) state that pet owners purchase a reader "from a pharmacy" but name no dollar figure; a Canadian clinic page (chestermerevet.com) returned HTTP 403 (not retrieved, wrong jurisdiction regardless). A web search surfaced a "$40-$60 per sensor" range attributed to consumer health-content sites (singlecare.com, goodrx.com) rather than to any single named clinic, insurer, or manufacturer page with a dated, attributable price — exactly the "search result, not a source of record" case this repo's sourcing rule excludes, so it is named here and not used. GluCurve Pet CGM (the one purpose-built veterinary product, §8) still discloses no per-unit price on its own site (glucurve.com, alrt.com) as of this pass — confirmed unchanged from the first pass, not merely un-re-checked. **Net: the $75.08 single-clinic figure remains the only per-unit veterinary CGM price found in either pass.**
• Evidence behind it: §4 in full.
• Adoption: not established at a population level — every study above is a single-institution cohort (n=7 to n=34 animals); no aggregate US adoption-rate figure was found.
• Abandonment record — why people stop:

• Verbatim:

"Sensors, which stopped working in situ (70% [46/66]), had a median functional life of 8.3 (1.6-14) days." — PMID 33229057

"The most challenging aspects of using the FGMS were ensuring proper sensor fixation during the wearing period (47%), preventing premature detachment (40%), and purchasing the sensor (34%). Moreover, 36% of DPOs reported that the device cost was difficult to afford in the long term." — PMID 36977242

"Due to a manufacturing process issue that may cause carbon build up in the sensor that may result in incorrect low glucose results." — openFDA recall Z-1020-2026

Read this carefully, in both directions, the same way the potassium dossier's capillary section does. The complication and failure-mode record above is evidence that the incumbent off-label product has a real, measured abandonment/complication profile — not merely an assumed inconvenience. It is not evidence that owners or clinicians would adopt this candidate's purpose-built veterinary product instead; a device engineered around haired-skin adhesion and species-specific calibration might address some of these exact failure modes (sensor detachment, skin reaction, ceiling effects) or might not, and nothing here tests that. The FDA recalls attach to the human-labeled product and its own 510(k)s — they are included because the physical sensor a veterinarian applies off-label is the identical recalled unit, not because any recall names a veterinary use.

Gaps: no veterinary-labeled competing product's abandonment record exists yet (GluCurve Pet CGM, §8, has no complication/accuracy literature found); MAUDE narratives are unreachable (no connector); no data on what fraction of US veterinary practices offer FGMS at all versus curve-only.

6. Reimbursement landscape

as_of: 2026-08-26 (refreshed same-day, second research pass — attempted a direct search for a citable published finding on pet-insurance coverage of veterinary diagnostic/monitoring devices)

Settled negative, not unsearched. Per candidate claim 1 (21 CFR 807.65(b), 807.20(a), 807.81(a), 814.1(a), all independently verified via regulation), veterinary-only devices sit entirely outside FDA premarket review, and Medicare/Medicaid/CLIA/HCPCS are human-payer constructs with no veterinary counterpart. hcpcs, payment, coverage, and procedures in this repo query exclusively human federal-payer systems, and there is no veterinary analogue to point them at — this is a structural fact about US health-payer architecture, not a gap this dossier can close with more searching. What remains genuinely open is pet insurance policy language (Trupanion, Nationwide, ASPCA, etc.), which is licensed/commercial and not reachable by any connector here; that is the one live open question in this section (see §9).

This pass's specific attempt, and its result: no citable, resolvable-identifier source was found establishing pet-insurance coverage of veterinary diagnostic or monitoring devices, and that absence is recorded here as a finding, not a shortfall of searching. What was tried:

• NAPHIA (North American Pet Health Insurance Association) State of the Industry Report — the aggregate-industry source this pass was specifically pointed at. Its own site (naphia.org/industry-data/, naphia.org/industry-data/about-this-data/, naphia.org/news/naphia-news/soi-report-2025/) returned only a bot-verification interstitial page to a direct fetch on every URL tried ("One moment, please... Please wait while your request is being verified...") — the same class of failure this dossier already names for fda.gov and aaha.org: a gate page parsed as content would read as "nothing found," which is why it is recorded as not retrieved, not as "NAPHIA reports no device coverage." Search summaries (not the report itself) describe NAPHIA's tracked figures as aggregate written premium ($5.2B, 2024) and enrollment counts (7.6 million pets), not device- or procedure-level coverage categories — but this is a description of a search result, not a primary-document finding, and is not recorded as dossier content for that reason.
• A citable published finding on pet-insurance uptake was located and read at full text: Springer, Lund, Sandøe & Corr, "Investigating dog and cat owners' uptake of pet health insurance and spending on veterinary treatment," Veterinary Record, 2026, DOI 10.1002/vetr.70621. Read directly (not from a search summary). It reports insurance uptake and owner spending intentions across Austria, Denmark and the UK (e.g. "41.7% of owners had insurance" across the three countries, with country-level rates from 21.3% in Austria to 56.3% in Denmark) but **is non-US (jurisdiction mismatch) and does not mention diagnostic devices, monitoring devices, or CGMs anywhere in the text retrieved** — confirmed by direct read, not inferred from the title. Named here so a later reader knows it was found and read, and specifically why it settles nothing for this candidate: wrong jurisdiction and wrong subject matter, not merely a weak match.

"41.7% of owners had insurance" — Springer et al. 2026, DOI 10.1002/vetr.70621

• A CAS (Casualty Actuarial Society) seminar document on pet insurance (casact.org/sites/default/files/2022-02/CS28-Pet_Insurance.pdf, dated by its own file path to Feb 2022) was fetched but its content could not be extracted in readable form by the fetch tool used (an encoded/slide-format PDF); no finding is recorded from it because none was actually read — this is a "not retrieved in readable form," distinct from "found and read but silent on devices."
  • Non-primary sources seen and deliberately not used: consumer-health content sites (singlecare.com, goodrx.com) describe FreeStyle Libre as sometimes "covered by pet insurance... as a reimbursement... after purchase," but name no specific insurer, policy document, or claims figure — exactly the "search result, not a source of record" case this repo's sourcing rule excludes. Not recorded as a finding.

Net for this section: the reimbursement-pathway open question is unchanged in substance from the first pass — no citable source establishes pet-insurance coverage terms for this category of device — but the search this pass was real and specific, not a repetition of the first pass's general note, and every attempt and its outcome (retrieved-and-silent vs. not-retrieved-at-all vs. seen-and-rejected) is now distinguished rather than collapsed into one "inconsistent" line.

Gaps: commercial pet-insurance coverage-policy documents for CGM devices specifically — genuinely open, and the only item in this section that isn't closed by the underlying regulatory fact. NAPHIA's own report remains specifically unretrieved (bot-gated), not searched-and-silent — a materially different gap than "no report exists."

7. Technology trajectory

as_of: 2026-08-26

Verbatim:

"The concept of the biosensor for measuring glucose levels was first proposed in 1962 by Clark and Lyons from the Children's Hospital of Cincinnati." — PMID 22399892 (PMC3292132), quoting the article's own historical section, retrieved directly rather than taken from a search summary

"Updike and Hicks significantly simplified the electrochemical glucose assay by immobilizing and thereby stabilizing GOx. They immobilized GOx in a polyacrylamide gel on an oxygen electrode for the first time and measured glucose concentration in biological fluids." — PMID 22399892

Note per the template's own instruction: the 510(k) clearance count above is a counted series and cannot carry a confidence tag anywhere in this repo (docs/backlog.md 1.10) — it lives here as context only.

Gaps: no patent-landscape check was possible (candidate claim 6 remains blocked — PATENTSVIEW_API_KEY unset; see §9). No membrane-chemistry or manufacturing-process literature specific to the current generation of sensors was reviewed here; §4/§5's accuracy/complication literature is downstream-application evidence, not chemistry-trajectory evidence.

8. Market players & dynamics

as_of: 2026-08-26 · sources searched: company/trade press pages, followed from web search to the primary announcement, not cited from the search summary alone

What this section explicitly does not record as fact: several market-research-firm reports (grandviewresearch, gminsights, coherentmarketinsights, futuremarketinsights, precedenceresearch, indexbox) were returned by web search and quote a "pet diabetes care devices market" sized at $1.2-2.2 billion with various forecast CAGRs. None of these figures is recorded here. They are paywalled commercial market-research products with undisclosed methodology and sample, which is exactly what "a search result is never a source of record" rules out — they are named here only so a later reader knows they were seen and deliberately not used, not treated as an oversight.

Gaps: no stated reason for the ~2023-to-2026 gap in GluCurve's US availability was found in any source retrieved (not a withdrawal-reason finding — a genuine absence, recorded as such rather than guessed at). No funding data for ALR Technologies' animal-health division specifically was retrieved. No recall or complaint record for GluCurve was found (no veterinary-device recall channel exists to check, per §6's structural point).

9. Open questions — and what would settle them

10. Provenance & staleness

Queries run against connectors literature in this pass (for anyone re-running this): "FreeStyle Libre 3 cats diabetic accuracy," "continuous glucose monitoring diabetic dogs accuracy," "flash glucose monitor diabetic dog," "diabetes mellitus prevalence dogs cats United States," "canine diabetes mellitus incidence insurance claims," "continuous glucose monitoring history enzyme electrode review," "glucose curve variability diabetic dogs," "stress hyperglycemia cats blood glucose," "canine diabetes mellitus prevalence veterinary teaching hospital," "diabetes mellitus cats electronic health records prevalence." Several multi-clause queries (e.g., "blood glucose curve variability diabetic cats stress hyperglycemia hospital") returned zero hits and were narrowed rather than assumed to mean "no literature exists" — narrower re-runs are recorded in the section they informed.

Queries run in the second research pass (2026-08-26, same day): Reimbursement — "pet insurance coverage veterinary device reimbursement," "veterinary economics pet insurance claims cost." TAM/epidemiology — "diabetes mellitus dogs prevalence electronic health records United States," "diabetes mellitus cats prevalence electronic health records United States," "diabetes mellitus prevalence primary care veterinary hospitals dogs cats," "Banfield diabetes mellitus prevalence," "Nationwide pet insurance diabetes mellitus claims," "veterinary continuous glucose monitor cost price owner," "Banfield Applied Research Knowledge diabetes prevalence pet," "diabetes mellitus incidence dogs cats veterinary practice management software United States," "prevalence diabetes mellitus companion animals United States veterinary teaching hospitals 2000 2020," "canine diabetes mellitus prevalence," "feline diabetes mellitus prevalence" (these last two, unqualified by "United States," are what surfaced PMID 40525629 and PMID 41443822 — the earlier, more qualified queries returned only irrelevant human-diabetes and unrelated veterinary titles), "VetCompass diabetes mellitus dogs prevalence primary veterinary care," "diabetes mellitus dogs primary veterinary care United Kingdom prevalence," "diabetes mellitus prevalence dogs United States general practice electronic records," "Guptill diabetes mellitus dogs," "diabetes mellitus cats United States general practice prevalence Antech IDEXX" — this last block of US-qualified queries returned zero PubMed hits, confirming (not merely repeating) the first pass's finding that no post-1999, US-specific, general-population prevalence study surfaces in PubMed under any phrasing tried.

What this dossier has never looked at: MAUDE narratives (no connector exists in this repo, docs/backlog.md 1.9); the AAHA 2018 diabetes management guideline's full text (blocked by a 403 on two separate attempts, across both passes — not silent); canine-specific hyperinsulinemic-clamp accuracy data (only field-cohort dog data was found); GluCurve Pet CGM's own accuracy or complication literature (none exists in PubMed as of this pass — it may simply not have been published yet, or may not be published at all for a product without an FDA-review requirement to drive it); NAPHIA's State of the Industry Report content itself (bot-gated on three direct fetch attempts this pass, second pass — not silent, not retrieved); any AVMA member survey or Banfield/BluePearl aggregated-EHR report (still unreached by either PubMed queries or web search across both passes — this is the closest thing to a settled negative in this dossier: two passes, many phrasings, nothing found). Commercial pet-insurance policy text for Trupanion/Nationwide/ASPCA specifically remains unreached (no connector, and this pass's web search reached only NAPHIA's aggregate industry page, which is itself unretrieved, not individual insurers' policy documents). Absence of scanning is not absence of change, and a later reader should treat sections 2 and 3 as a starting point, not coverage — §6 (reimbursement) is now a searched-and-negative section rather than an unsearched one, which is itself a finding.