Overweight and obesity have been adversely associated with longevity in dogs but there is scarce knowledge on the relation between body composition and lifespan. We aimed to investigate the effects of body composition, and within-dog changes over time, on survival in adult Labradors using a prospective cohort study design. The dogs had a median age of 6.5 years at study start and were kept in similar housing and management conditions throughout. The effects of the various predictors, including the effect of individual monthly-recorded change in body weight as a time varying covariate, were evaluated using survival analysis.
All dogs were followed to end-of-life; median age at end-of-life was 14.0 years. Body composition was measured annually with dual-energy x-ray absorptiometer (DEXA) scans between 6.2 and 17.0  years. All 39 dogs had DEXA recorded at 8, 9 and 10 years of age. During the study the mean (± SD) percent of fat (PF) and lean mass (PL) was 32.8 (± 5.6) and 64.2 (± 5.5) %, respectively, with a mean lean:fat ratio (LFR) of 2.1 (± 0.6); body weight (BW) varied from 17.5 to 44.0 kg with a mean BW change of 9.9 kg (± 3.0). There was increased hazard of dying for every kg increase in BW at 10 years of age; for each additional kg of BW at 10 years, dogs had a 19% higher hazard (HR = 1.19, P = 0.004). For the change in both lean mass (LM) and LFR variables, it was protective to have a higher lean and/or lower fat mass (FM) at 10 years of age compared to 8 years of age, although the HR for change in LM was very close to 1.0. For age at study start, older dogs had an increased hazard. There was no observed effect for the potential confounders sex, coat colour and height at shoulders, or of the time-varying covariate.
These results suggest that even rather late-life control efforts on body weight and the relationship between lean and fat mass may influence survival in dogs. Such "windows of opportunity" can be used to develop healthcare strategies that would help promote an increased healthspan in dogs.