Protein and calorie restrictions extend median lifespan in many organisms. However, studies suggest that among-individual variation in the age at death is also affected. Ultimately, both of these outcomes must be caused by effects of nutrition on underlying patterns of age-specific mortality (ASM). Using model life tables, we tested for effects of dietary macronutrients on ASM in mice ( Mus musculus). High concentrations of protein and fat relative to carbohydrates were associated with low life expectancy and high variation in the age at death, a result caused predominantly by high mortality prior to middle age. A lifelong diet comprising the ratio of macronutrients self-selected by mouse (in early adulthood) was associated with low mortality up until middle age, but higher late-life mortality. This pattern results in reasonably high life expectancy, but very low variation in the age at death. Our analyses also indicate that it may be possible to minimize ASM across life by altering the ratio of dietary protein to carbohydrate in the approach to old age. Mortality in early and middle life was minimized at around one-part protein to two-parts carbohydrate, whereas in later life slightly greater than equal parts protein to carbohydrate reduced mortality.