Suppression of ovulation by transdermal oestradiol patches.

implied that the indications for surgery might be more important than surgery itself. There are several potential biases. Early symptoms in women with osteoarthritis may have been misinterpreted and resulted in hysterectomy. This is unlikely given the long median delay (six years) between operation and diagnosis and that those with spinal osteoarthritis did not have the greatest rate of hysterectomy. The difference in rate of hysterectomy between the women with osteoarthritis and those with rheumatoid arthritis may have been due to fewer gynaecological abnormalities occurring in those with rheumatoid arthritis. Previous studies found no difference between women with rheumatoid arthritis and controls,2 and the rates ofhysterectomy in women with rheumatoid arthritis were similar to those in our population controls. By contrast, a link between hysterectomy and predisposition to osteoarthritis is biologically plausible. Dysfunctional uterine bleeding and fibroids, the main reasons for hysterectomy in our study, have both been attributed to an excess of oestrogens. Oestrogens have also been found to affect adversely several different animal models of osteoarthritis,4 possibly by affecting turnover of proteoglycan.