I am an independent researcher and data analyst with experience parsing FDA/CVM public adverse-event data files and building reproducible data summaries and charts.

For the Librela chart work, I parsed FDA/CVM public quarterly JSON adverse-event data files, filtered the reports involving Librela, applied deduplication and counting rules, and prepared the charts included as Exhibit 13(b).

I am not a veterinarian, physician, lawyer, or regulator. My role is limited to the data analysis: where the numbers came from, how the reports were counted, what the charts show, and what the charts do not prove.

“All Cases” means reports where Librela was one of the listed drugs.

This includes both monotherapy reports, where Librela was the only listed drug, and polytherapy reports, where other medications were also listed.

Adverse-event reports do not prove causation by themselves. The charts show reported death/euthanasia outcomes over time where Librela was one of the listed drugs.

The charts are intended to show reported patterns in the public data, not to determine the cause of individual outcomes.

“Monotherapy” means Librela/bedinvetmab was the only identifiable named active ingredient in the adverse-event report.

These reports remove cases where other medications were also listed, but they still do not prove causation by themselves.

The charts do not show the death rate or incidence rate for all dogs treated with Librela.

The charts show reported death/euthanasia outcomes in FDA/CVM public adverse-event data. Because the public data do not include the total number of dogs treated with Librela, the charts cannot be used to calculate an incidence rate.

The charts include both death and euthanasia because both were reported as fatal outcomes in FDA/CVM public adverse-event data.

Some reports mentioned both death and euthanasia. In those reports, the dog was counted only once as euthanasia to prevent double-counting.

Yes. Where age was reported, most dogs in the all-cases time-to-onset analysis were older dogs.

Age data were available for 3,005 dogs, or 86.3% of the dogs with known time-to-onset. The average age was 12.0 years, and the median age was also 12.0 years.

Among dogs with age data:

<3 years: 6 dogs
3–5 years: 47 dogs
6–8 years: 232 dogs
9–12 years: 1,359 dogs

12 years: 1,361 dogs

Age is important context, but age alone does not determine whether a reported outcome was expected, unrelated, or potentially associated with the drug. The charts show reported outcomes in FDA/CVM public adverse-event data; they do not determine causation for individual dogs.

The numbers came from FDA/CVM public quarterly adverse-event JSON data files made available through openFDA.

Data download page: https://open.fda.gov/data/downloads/

For the petition chart package, I used FDA/CVM public quarterly JSON files from 2023 Q2 through 2025 Q4. I parsed the JSON files, filtered for dog reports involving Librela/bedinvetmab, applied deduplication and counting rules, and summarized reported death/euthanasia outcomes and time-to-reported-onset information.

The charts are my analysis of FDA/CVM public data. They are not ready-made FDA charts or FDA summaries.

One important limitation is that public quarterly files are database snapshots. FDA/CVM may later add, update, recode, remove duplicate entries, or otherwise process reports after a quarterly file first appears. For that reason, counts can change if the public JSON files are refreshed later.

Yes. The formal methodology explanation is filed as Exhibit 13(a) in the FDA citizen-petition docket.

Exhibit 13(a) explains the source data, cutoff date, counting units, all-cases and monotherapy methods, fatal-outcome classification, time-to-reported-onset calculation, traceability checks, validation outputs, relationship to the FDA FOIA clinical-detail listing, and limitations.

You can download Exhibit 13(a) here:

https://www.regulations.gov/document/FDA-2026-P-6249-0021

Because the FOIA document and Jeffrey’s charts are not counting the same thing.

The FDA ADE Report Bedinvetmab (Librela): 5/5/2023–12/31/2025 is a CVM ADE Comprehensive Clinical Detail Report Listing. It lists reported clinical signs or reaction terms, grouped by species and route of administration. FDA states that more than one clinical sign may be reported per ADE case report, so the “Number of times reported” column is not additive and does not necessarily represent the total number of reports received. FDA also states that this clinical-detail listing does not include the final outcome of the reaction.

Jeffrey’s charts are different. They analyze the underlying FDA/CVM public quarterly JSON records to identify death/euthanasia outcome information.

A report may contain death/euthanasia information in structured outcome data, fatal reaction terms, or both. The FOIA listing is not designed to function as a complete fatal-outcome count.

Some reports also involve more than one dog. Jeffrey’s charts count individual dogs, while the FOIA clinical-sign listing may reflect term frequencies rather than a full count of affected dogs.

Public JSON records can also change over time as reports are corrected, supplemented, or updated. That can affect totals later.

Bottom line: Jeffrey’s charts are an outcome-based analysis of FDA/CVM public JSON records. The FOIA document is a clinical-sign frequency listing, not a final-outcome total. Different totals do not automatically mean either source is wrong.

These are reported adverse-event data, not incidence or causation findings.

Yes. Adverse-event reporting systems are generally understood to involve underreporting.

FDA’s own Librela ADE report states that underreporting occurs with most adverse-event reporting systems and that reporting frequency can vary over time, including when a drug is newly marketed or when media publicity occurs.

A 2024 study published in the Journal of Small Animal Practice reviewed 10,565 veterinary records and identified 827 suspected adverse drug reactions. Approximately 90% of those suspected adverse drug reactions were not recorded as reported.

Study link: https://onlinelibrary.wiley.com/doi/full/10.1111/jsap.13721

Because of underreporting, the charts should be viewed as reported adverse-event data, not as a complete count of every event that may have occurred.