The aim of this study was to estimate the additional cost of medical care (the incremental cost) caused by incident hip and vertebral fractures, using a matched case cohort design within a longitudinal follow-up study. Incident hip fractures were recorded using the regular follow-up system of the Rotterdam Study. Incident vertebral fractures were recorded by morphometric comparison of spinal radiographs taken at intervals of 2.2 years on average. The matched control group was randomly selected from other participants of the Rotterdam Study in whom no fracture occurred during follow-up, but who were otherwise comparable at baseline. Cases were matched for age, gender, self-perceived health, ability to perform activities of daily life, living situation and general practitioner. Medical expenditure was assessed by retrieval of the general practice medical records and by recording all hospital and nursing home admissions, and all general practice and outpatient visits. Pharmaceutical consumption was recorded through the computerized records of the central pharmacy. Valid results were obtained for 44 pairs (91%) in the hip fracture and for 42 pairs (93%) in the vertebral fracture group. Cost of medical consumption in the year before the hip fracture was similar in patients and control subjects, but the incremental cost in the first year after the hip fracture was almost US$10 000. In the second year after hip fracture the incremental cost was still about $1000. Accounting for the excess mortality in hip fracture patients had little effect on cost in the first year, but cost in the second year was doubled to almost $2000. For vertebral fractures, we did not detect important acute care costs, but these fractures were associated with a yearly recurrent incremental cost of over $1000. However, almost half this difference was already present before the occurrence of the fracture, and was attributable to hospital admissions. The remainder of the incremental cost was mainly due to pharmaceutical consumption and to a lesser extent to admissions to orthopedic surgery wards. We conclude that hip fractures cause excess mortality and an important incremental cost especially during the first year, and that these could probably be avoided by prevention of hip fractures. For vertebral fractures we found no evidence of important acute care costs but we observed a yearly returning incremental cost. Part of this incremental cost, however, was pre-existing and might therefore by caused by co-morbidity.