The RVC's Life-Expectancy Figures, Breed by Breed

A database built from the consulting room, not the claim form

VetCompass is the Royal Veterinary College's primary-care surveillance programme, drawing anonymised clinical records directly from participating veterinary practices across the United Kingdom. Unlike insurance-based datasets — the kind assembled by Swedish insurance company Agria or the Finnish pet-insurance pool — VetCompass captures animals whether or not they are insured, and records the full arc of a patient's life, from first vaccination to the date and cause of death. That methodological difference matters considerably when median-age-at-death figures are being read across breeds.

The landmark life-expectancy paper produced by the VetCompass programme and published in Scientific Reports in February 2024 drew on records for more than 580,000 dogs whose deaths were documented in the programme's primary-care database between 2016 and 2020. Dan O'Neill of the Royal Veterinary College, one of VetCompass's most frequently cited authors, was among the lead researchers. The study's headline finding — a median lifespan of 12.5 years across all breeds and sexes combined — was widely reported, but the breed-level data beneath it carries substantially more diagnostic weight.

What the breed-level table actually shows

The study ranked breeds by median age at death, with sample sizes that vary sharply from one row to the next. Flat-coated retrievers recorded a median of just 9.4 years, a figure consistent with their breed's well-documented susceptibility to histiocytic sarcoma. Jack Russell terriers, among the most numerous dogs in the VetCompass database, recorded a median of 12.7 years — a figure drawn from a sample large enough to be treated as statistically stable. Lancashire heelers came in at 13.3 years, the longest median recorded in the published table, though their absolute sample size is far smaller than the terrier entries.

The French bulldog, a breed whose registration trajectory reflects its surge in commercial popularity, recorded a median of 4.5 years in the same dataset — the lowest of any breed studied and less than half the all-breed median. The authors attributed this to the compound welfare burden of brachycephalic obstructive airway syndrome (BOAS) — a set of upper-airway anatomical defects that restrict breathing — alongside spinal, skin and reproductive disorders concentrated in the breed. English bulldogs recorded a median of 7.4 years; American bulldogs, 7.8 years. The VetCompass figures gave the short-faced breeds' poor longevity a primary-care evidential basis that earlier insurance-only studies had approximated but could not as precisely anchor.

Among the larger traditional breeds, golden retrievers showed a median of 12.0 years, Labrador retrievers 12.0 years also, and border collies 13.1 years. German shepherd dogs recorded 10.3 years, with the VetCompass authors noting that disorders of the musculoskeletal and neurological systems feature prominently among recorded causes of death. Cavalier King Charles spaniels, whose syringomyelia and cardiac vulnerabilities have been extensively documented in clinical literature, recorded 11.0 years — a figure that sits measurably below the all-breed median, consistent with findings from earlier VetCompass papers examining the breed's disorder burden.

Why primary-care records produce different numbers

An insurance dataset captures dogs whose owners chose to insure them, and the decision to insure correlates with ownership type, income level, breed popularity, and perceived value — all factors that introduce selection bias. An insured Labrador retriever and an uninsured crossbreed in the same practice are both VetCompass subjects; only the first appears in an insurance company's actuarial record.

Insurance datasets also tend to under-represent deaths: a dog that dies at home without a final veterinary consultation may never appear in a claim record, while a VetCompass practice that has seen the same dog throughout its life will often log the death, the cause and the age. This means VetCompass median figures are not simply higher or lower than insurance equivalents — they are constructed from a differently shaped population, and direct numerical comparison between the two sources requires caution.

The programme's method does carry its own limitations. Participating practices are not a random sample of all UK veterinary premises; urban and suburban practices are over-represented, and within those, certain client demographics skew the breed population. Crossbreeds — which the VetCompass data shows living to a median of 11.8 years in the 2024 paper — appear in large numbers precisely because the database does not restrict its intake to registered pedigree animals. That inclusion is methodologically valuable: it allows the study to compare pedigree and non-pedigree longevity within the same primary-care framework rather than across different recording systems.

Rasowy covers the pedigree dog breeds as they are registered, regulated and measured

The sample-size variation across breeds also warrants reader attention. A median derived from several hundred deaths is more reliable than one drawn from fewer than fifty, and the 2024 paper's supplementary data makes these denominators visible. Breeds with small national populations — the otterhound, the Irish red and white setter, several of the Kennel Club's Vulnerable Native Breeds — either do not appear in the published table or appear with sample sizes too small to support confident interpretation. Their absence from a lifespan ranking is not evidence of long or short life; it reflects the programme's dependence on clinical encounter volume.

Reading the numbers against the regulatory record

The VetCompass figures have entered the policy literature. The Norwegian Society for the Protection of Animals cited shortened lifespan data — including findings consistent with VetCompass outputs — in the litigation that reached the Borgarting Court of Appeal in 2023, where the court found that breeding English bulldogs and cavalier King Charles spaniels in their current form violated the Norwegian Animal Welfare Act. The RVC data was not the sole evidence before the court, but primary-care longevity figures of the kind VetCompass supplies are increasingly treated by regulators and courts as harder evidence than show-ring assessment or breed-club testimony.

The British Veterinary Association has referenced VetCompass research in its public positions on brachycephalic breeds, and the Kennel Club's own health monitoring work cites the programme. Whether the figures translate directly into breed-standard revision, registration restriction or veterinary guidance depends on political and institutional processes that the data itself does not resolve. What the 2024 paper does resolve is that the question of how long a breed typically lives can now be answered with primary-care evidence at scale — and that for several breeds, the answer disclosed is considerably shorter than their registries' promotional histories have suggested.