BackgroundMean body weight gradually increases with age. Yet, little data exists on the prevalence of excess weight in populations aged 80 years or older. Moreover, little is known about predictors of overweight and obesity in old age. Thus, the purpose of this study was: To present data on the prevalence of excess weight in old age in Germany, to investigate predictors of excess weight in a cross-sectional approach and to examine factors affecting excess weight in a longitudinal approach.MethodsSubjects consisted of 1,882 individuals aged 79 years or older. The course of excess weight was observed over 3 years. Excess weight was defined as follows: Overweight (25 kg/m2 ≤ BMI < 30 kg/m2) and obesity (BMI ≥ 30 kg/m2). We used fixed effects regressions to estimate effects of time dependent variables on BMI, and overweight or obesity, respectively.ResultsThe majority was overweight (40.0 %) or obese (13.7 %). Cross-sectional regressions revealed that BMI was positively associated with younger age, severe walking impairments and negatively associated with cognitive impairments. Excess weight was positively associated with younger age, elementary education, walking impairments and physical inactivity, while excess weight was negatively associated with cognitive impairment. Longitudinal regressions showed that age and severely impaired walking disabilities reduced BMI. The probability of transitions to excess weight decreased considerably with older age and occurrence of severe walking impairments (overweight).ConclusionsMarked differences between predictors in cross- and longitudinal setting exist, underlining the complex nature of excess weight in old age.Electronic supplementary materialThe online version of this article (doi:10.1186/s12877-015-0081-5) contains supplementary material, which is available to authorized users.