Bone remodeling requires bone resorption by osteoclasts, bone formation by osteoblasts, and a poorly investigated reversal phase coupling resorption to formation. Likely players of the reversal phase are the cells recruited into the lacunae vacated by the osteoclasts and presumably preparing these lacunae for bone formation. These cells, called herein reversal cells, cover >80% of the eroded surfaces, but their nature is not identified, and it is not known whether malfunction of these cells may contribute to bone loss in diseases such as postmenopausal osteoporosis. Herein, we combined histomorphometry and IHC on human iliac biopsy specimens, and showed that reversal cells are immunoreactive for factors typically expressed by osteoblasts, but not for monocytic markers. Furthermore, a subpopulation of reversal cells showed several distinctive characteristics suggestive of an arrested physiological status. Their prevalence correlated with decreased trabecular bone volume and osteoid and osteoblast surfaces in postmenopausal osteoporosis. They were, however, virtually absent in primary hyperparathyroidism, in which the transition between bone resorption and formation occurs optimally. Collectively, our observations suggest that arrested reversal cells reflect aborted remodeling cycles that did not progress to the bone formation step. We, therefore, propose that bone loss in postmenopausal osteoporosis does not only result from a failure of the bone formation step, as commonly believed, but also from a failure at the reversal step.