Healthy premenopausal women with regular cycles are believed to be increasing or maintaining bone density. However, few studies have prospectively documented spinal cancellous bone, the bone that changes rapidly in response to reproductive hormones, in this population. Furthermore, our previous one-year study documented that 24% of the one-year bone change by quantitative computed tomography (QCT) was related to subclinical ovulatory disturbances (short luteal phase and non-ovulation) in the presence of regular menstrual cycles. The purpose of this study was to document the cancellous bone change over five years in this initially ovulatory, premenopausal cohort of 66 healthy women. Thirty-seven women, who continued to be premenopausal and have regular cycles, completed this five-year study. Those enrolled differed only by being older and weighing less than those who could not be contacted (n = 19) or who declined to participate (n = 10). Documentation of current ovulatory characteristics was obtained for at least three cycles in 27 women. At the five-year assessment, the volunteers were 40.6 (range 26-47) years old, weighed 58.5 (41-77) kg, and were 160.9 (149-174) cm in height. All were premenopausal, healthy, nonsmokers with regular menstrual cycles (mean 27.7, range 24-33 days). Six women with intervening events (such as pregnancy or use of oral contraceptives) had interval (12 to 60 months) QCT changes similar to the remaining 31 (-7.98 vs. -4.92 mg/cm, p = 0.1, respectively). Mean five-year QCT was 143.0 +/- 20.2 mg/cm, whereas the initial mean value was 151.9 +/- 20.1 mg/cm. Significant QCT loss over five years (-8.9 +/- 6.2 mg/cm) (95% Cl -6.9 to -11.0) correlated with QCT change in the first year (r = 0.629, p < 0.001). First-year change was not related to the subsequent four-year interval change (r = -0.056, p = 0.74), however. Five-year QCT change was not related to age, weight, osteoporosis family history, estimated calcium intake, or exercise, but did correlate with year one luteal index (luteal/cycle length) (r = 0.339, p = 0.043). Significant cancellous spinal bone loss occurs in healthy, ovulatory premenopausal women, and is influenced by subclinical disturbances of ovulation.