Labrador retrievers (42 of original 48) were used to assess the effects of lifetime diet restriction on glucose tolerance at ages 9-12 y. Restricted-fed (RF) dogs were fed 75% of the same diet consumed by control-fed (CF) pair-mates. An intravenous glucose tolerance test was done annually (maximal stimulation, nonsteady-state). Diet treatment, age, and interactions were fixed effects. Statistical procedures used included mixed-model, repeated-measures ANOVA; least-squares means; Tukey's multiple comparison; paired t tests; and Spearman rank correlations. Glucose k-value and half-life, and insulin sensitivity (total, and 9, 10, 11 y, and per lean mass) were higher (P < 0.05) in RF than in CF dogs. Late-phase insulin release [area under the curve (AUC) 30-120 min] was less (P < 0.05) in RF than in CR dogs. Early-phase insulin release (AUC 0-5 min), y 12 insulin sensitivity and insulinogenic index did not differ between RF and CF dogs. Insulin peak, delta and total AUC increased (P < 0.05) with age, whereas the glucose k-value and glucose half-life were not affected by age. Insulin sensitivity was negatively, and insulin AUC 30-120 min, peak and delta glucose were positively correlated with body weight, body condition score, fat mass, percentage of fat and abdominal fat/total tissue. Higher insulinogenic indices tended (P = 0.053) to be associated with greater median survival and dogs with higher insulin sensitivity were at lower (P < 0.05) risk of dying or receiving chronic disease treatment. Time to first osteoarthritis treatment or death was greater with lower basal glucose and higher insulin sensitivity (P < 0.05), but diet restriction explained most of this relationship's variation. Glucose disposal efficiency and insulin response were associated with increased quality and length of life in diet-restricted dogs.