Living with a canine companion is postulated to increase physical activity. We test the hypotheses that adults living with a canine companion have a higher level of physical activity and reduced mortality risk compared to those not living with a companion animal. A U.S. national health survey with longitudinal mortality follow-up studied 11,394 American men and women aged 40 years and over examined in 1988–1994 followed an average 8.5 years. Measurements at baseline included self-reported companion animals in the household, socio-demographics, health status, physical and biochemical measurements. Outcome measures were leisure-time physical activity (LTPA), and death from all causes. Death during follow-up occurred in 3,187 persons. In bivariate cross-sectional analyses living with a dog was associated with more frequent LTPA and higher survival. In proportional hazards regression analysis, no significant interaction of age, gender or ethnicity with animals was found. After adjusting for confounding by baseline socio-demographics and health status at ages 40+, the hazards ratio (95% confidence limits) for living with a canine companion compared to no animals was 1.21(1.04–1.41, p < 0.001). After also controlling for health behaviors, blood pressure and body mass, C-reactive protein and HDL-cholesterol, the HR was 1.19 (0.97–1.47, NS). In a nationwide cohort of American adults, analyses demonstrated no lower risk of death independent of confounders among those living with canine or feline companions, despite positive association of canine companions with LTPA.