There is an increased risk of hip fracture and low bone mass in thin individuals. An association between weight loss and hip fracture has also been reported. In addition, it has been suggested that weight cycling might lead to bone loss. We studied weight variability and change in 19,938 women and 19,151 men who all attended three consecutive health examinations during an average period of 12 years, and assessed the effect of these on the incidence of hip fracture during a subsequent follow-up. Mean age at start of follow-up was 48.6 years in women and 48.5 years in men. For each subject weight variability and linear trend in weight change between the three examinations were assessed by linear regression of weight versus time. The cohort was followed on average 11.6 years from the third examination with respect to hip fracture. During follow-up, 148 hip fractures were identified in women and 59 in men. In both sexes, those with most weight variability had increased risk of fracture (relative risk (RR) = 2.07, 95% confidence interval (CI) 1.24-3.46 in women, and RR = 2.70, 95% CI 1.25-5.86 in men, high vs low quarter of weight variability). Overall, the effect of weight variability was not affected by adjustment for body mass index and linear trend in weight change. In men, there was also an association between weight loss and hip fracture. In summary, high weight variability defined a group with increased risk of hip fracture in this middle-aged cohort. The effect was independent of body mass index and linear trend in weight change. Whether weight variability leads to increased risk of fracture per se or whether it defines a group with otherwise increased risk of fracture is not known, and needs further investigation.