Objective. -To compare two series of patients admitted to a rheumatology department for bone metastases 30 years apart and to obtain a pragmatic estimation of the percentage of bone metastases revealing cancer.

Methods. - The recent series was composed of 132 patients seen between 1989 and 1996 and the earlier series of 50 patients admitted between 1958 and 1967. All patients were admitted to the same rheumatology department for bone metastases from a diagnosed or undiagnosed primary. Both series were studied retrospectively. The rate of occurrence of each cancer diagnosis was determined in the patients with and without known primaries.

Results. -Among the patients with metastatic disease as the first manifestation of cancer, the percentage with lung cancer increased from the early to the recent series, particularly among women, whereas the percentages with prostate and breast cancer decreased. The primary remained unknown in 27% and 38% of patients in the early and recent series, respectively.

Conclusion. -Despite the introduction of new investigations, our ability to identify primaries responsible for bone metastases does not seem to have improved. However, our data should be interpreted with caution since recruitment probably differed between the two series.