A key feature of postmenopausal osteoporosis is the loss of trabecular bone mass and connectivity. The current study focuses on these parameters in the assessment of long-term (12 and 18 months) parathyroid hormone (PTH) therapy and its withdrawal (6 months) in the ovariectomized cynomolgus monkey (Macaca fascicularis), a well-characterized model for bone changes associated with postmenopausal osteoporosis. We used static and dynamic histomorphometric parameters to assess the amount and architecture of cancellous bone in four clinically important sites for osteoporotic fractures, including the lumbar vertebra, femoral neck, distal radius, and iliac crest. Recombinant human PTH(1-34) was administered daily to two groups for 18 months at 1.0 microg/kg per day (n = 19) and 5.0 microg/kg per day (n = 21). To study the effects of PTH withdrawal, two groups were administered PTH(1-34) daily for 12 months at 1.0 microg/kg per day (n = 20) and 5.0 microg/kg per day (n = 20), followed by daily administration of vehicle for 6 months. Sham-ovariectomized and ovariectomized (ovx) groups each received daily injections of vehicle for 18 months. Treatment with PTH had minimal effects on bone formation rates at the timepoints studied, but markedly increased cancellous bone volume relative to ovx monkeys in iliac crest biopsies at 6 and 15 months, as well as in terminal specimens of lumbar vertebrae, femoral neck, and distal radius after 18 months. At all sites, PTH significantly improved trabecular architecture, as evidenced by increased trabecular number (Tb.N) and decreased trabecular separation (Tb.Sp), with no significant change in trabecular thickness (Tb.Th). The mechanism of these structural changes is suggested by qualitative observations of trabecular tunneling observed in the iliac crest and vertebra. Longitudinal tunneling of thickened individual trabeculae is hypothesized to convert them into multiple trabeculae, resulting in a normalization of Tb.Th, but an increase in Tb.N. A significant positive effect on cancellous bone volume was still apparent after a 3-6 month withdrawal period following 12 months of PTH treatment in the iliac crest, vertebra, and femoral neck. Corresponding increases in Tb.N and decreases in Tb.Sp also remained significant after PTH withdrawal at these three sites. The distal radius was relatively insensitive to PTH treatment or its withdrawal, compared with the other bones. In summary, PTH therapy dramatically improved cancellous bone mass and architecture in both axial and appendicular sites.