When only the tip of the iceberg was visible, a diagnosis of primary hyperparathyroidism was a sufficient indication for surgical intervention [1]. Many physicians are still most comfortable with this policy, but it has a serious flaw. There are not now, and probably never will be, enough experienced and competent parathyroid surgeons to ensure that operating on all affected persons would do more good than harm [4]. Several authors [1, 2] have suggested that controlled clinical trials be done to determine how persons with fortuitously diagnosed hyperparathyroidism should be treated, but no such trials have yet been completed. Even if a net benefit of surgery in the most expert hands could be shown, recommending surgery would be a reasonable policy only for the few physicians who have ready access to such a surgeon.

Most persons with fortuitously discovered primary hyperparathyroidism accept the role of patient despite uncertainty about how the condition will be managed. Some reasonable indications for surgery in asymptomatic patients include young age, severe hypercalcemia, and individual preference, but something more tangible is needed. Bone densitometry is very attractive as a decision-making tool: It addresses a previously unsuspected possible benefit of surgical treatment [5], it is consistent with the current interest in detecting and preventing age-related bone loss, and it permits the formulation of numerically precise guidelines [6]. Furthermore, as Grey and colleagues discuss in this issue [7], it should help to define the role of hormone replacement therapy in the management of these patients. However, the proper use of bone densitometry in this context requires some understanding of bone remodeling, how it causes bone loss, and how it responds to estrogen deficiency and chronic hypersecretion of parathyroid hormone.

Bone remodeling is a replacement mechanism. For reasons that are poorly understood, bone that has reached a certain age becomes less able to carry out its mechanical or metabolic functions and must be renewed [8]. The instrument of bone remodeling is the basic multicellular unit, a unique temporary anatomical structure that consists of a team of osteoclasts in front; a team of osteoblasts behind; and associated blood vessels, connective tissue, and nerves [9]. Each basic multicellular unit originates in a particular place at a particular time, travels through or across the surface of bone toward its target, and continues beyond this target for a variable distance [8, 10]. Bone replacement by each basic multicellular unit is complete for only a few years before and after the attainment of peak adult bone mass. As a result of menopausal estrogen deficiency, more osteoclasts are recruited for bone remodeling than are needed; with increasing age, too few osteoblasts are recruited to refill each cavity [11].

Excess parathyroid hormone and estrogen deficiency increase bone turnover and accelerate bone loss. However, only the former causes hypercalcemia, because calcium homeostasis and bone remodeling are independent. This explains why hormone replacement therapy can reduce bone turnover in primary hyperparathyroidism without decreasing plasma calcium levels [7]. The hypercalcemia of primary hyperparathyroidism is caused by the combined effects of upward resetting of the blood-bone equilibrium at quiescent bone surfaces and increased tubular reabsorption of calcium [12]. Estrogen has no direct effect on these processes. Bone turnover increases because new basic multicellular units are created with increased frequency or because existing basic multicellular units continue to progress beyond their target [8]. When an increase in turnover is hormonally mediated, the latter is most likely the result of an increased supply of mononuclear osteoclast precursor cells [10]. These cells arise from the hematopoietic stem cell in the bone marrow at the behest of a complex network of cytokines and their receptors [11]. Estrogen deficiency and an excess of parathyroid hormone increase the local production of interleukin-6, and probably other cytokines, by stromal osteoblast precursor cells [13].

Estrogen deficiency and an excess of parathyroid hormone have both similar and different effects on bone remodeling. In cancellous bone, estrogen deficiency increases the depth of resorption from the surface, probably by delaying osteoclast apoptosis [10]; this leads to perforation and disconnection of trabeculae. With an excess of parathyroid hormone, however, these structural changes are less severe [14]. Estrogen deficiency can lead to thinning of cortical bone; this effect is also the result of increased depth of resorption. However, on the endocortical surface, an excess of parathyroid hormone potentiates this effect rather than protecting against it [15]. Why parathyroid hormone has such different effects on cortical and cancellous bone is a mystery, but these effects do explain why the bone deficits in primary hyperparathyroidism, particularly when expressed in SDs, are greater in cortical bone than in cancellous bone and greater in the appendicular skeleton than in the axial skeleton [2, 7, 14-16].

Two related concepts are of cardinal importance for understanding the effects of treatment on bone density in primary hyperparathyroidism, whether by surgical excision or hormone replacement therapy: the remodeling transient and the distinction between reversible and irreversible bone loss [17-19]. In each basic multicellular unit, several weeks elapse between the removal of bone by osteoclasts and the onset of bone replacement by osteoblasts, and several months elapse before the replacement process is finished. Consequently, each basic multicellular unit creates a temporary deficit of bone, and the aggregate of such deficits throughout the skeleton constitutes the remodeling space. The volume of remodeling space is proportional to the rate of bone turnover, so that the volume expands when turnover increases and contracts when turnover decreases. These changes in remodeling space volume are large enough to be readily detectable by bone densitometry and invariably account for almost all changes in bone density during the first 6 months of any treatment that reduces bone turnover (for example, the changes Grey and colleagues [7] saw in patients with hyperparathyroidism who were given hormone replacement therapy).

Reversible bone loss caused by expansion of the remodeling space underlies most of the effects of hyperthyroidism on bone density and much of the initial response to estrogen deficiency [19]. Of greater importance in the long-term, however, are the mechanisms of irreversible bone loss [18]. As each basic multicellular unit advances through or across the bone surface, it creates and leaves behind successive cycles of resorption followed by formation; each cycle constitutes a separate remodeling transaction. Each transaction represents the balance between the thickness of the old bone that has been resorbed and the thickness of new bone formed in its place. Once the transaction has been completed, its outcome is irrevocable. In the long term, bone loss as measured by densitometry is the cumulative aggregate effect of completed remodeling transactions; in each transaction, only about 90% to 95% of the resorbed bone is replaced. The normal rate of bone turnover is about 10% per year, so that the usual rate of age-related bone loss is about 0.5% to 1% per year. If the rate of bone turnover increases, the number of remodeling transactions completed in the same period increases in the same proportion with a corresponding increase in the rate of irreversible bone loss. This explains why primary hyperparathyroidism modestly accelerates this process, which in turn leads to the consequences discussed by Grey and colleagues [7].

Most agents used to treat osteoporosis reduce bone turnover by directly or indirectly limiting the supply of mononuclear osteoclast precursor cells, thereby curtailing the progression of basic multicellular units beyond their target [8]. Reducing turnover has two major effects on bone density. First, as mentioned earlier, contraction of the remodeling space to a new steady-state level causes a short-term increase in bone density that reaches a plateau after about 6 to 12 months. The magnitude of this effect depends on the local rate of turnover and thus is higher in the spine than in the extremities. Second, completion of fewer remodeling transactions in the same time leads to a sustained reduction in the rate of bone loss [17, 18]. Other salutary effects of reducing turnover are less easily captured by noninvasive methods. The mechanical threat to vertical trabeculae in the spine that have lost their horizontal supports is lessened, and the adverse effects of impaired osteoblast recruitment are mitigated. Reduced bone turnover, with the consequences just described, is the major effect that therapy with estrogens, bisphosphonates, calcitonins, calcium salts, or vitamin D congeners has on bone. However, only estrogens have other health benefits unrelated to the prevention of bone loss, the most important of which is protection against ischemic heart disease and consequent reduction in all-cause mortality [20].

Grey and colleagues [7] convincingly show that the presence of primary hyperparathyroidism is no reason to deprive women of the benefits of hormone replacement therapy. In their study, this therapy reduced biochemical indices of bone turnover and produced early increases in bone density at every site measured that were similar in magnitude to those produced by parathyroid surgery. The decrease in urine calcium excretion was large enough to account for the 3% advantage in total body calcium that patients who received hormone replacement therapy had over the controls after 1 year. Evidence suggested that bone loss was retarded at the spine and femoral neck, but a longer observation period would have been needed to establish this. Nevertheless, there is every reason to believe that the long-term benefits of hormone replacement therapy would be the same in all postmenopausal women, whether or not they have primary hyperparathyroidism. Consequently, hormone replacement therapy should be strongly considered in all such women. In persons who are not candidates for hormone replacement therapy, for whatever reason, the need for alternative therapy to prevent bone loss should be determined by bone densitometry and, if the results are equivocal, by biochemical evaluation of bone turnover. In osteopenic women with primary hyperparathyroidism who are not candidates for hormone replacement therapy, the most logical alternative therapy for preventing bone loss is parathyroid surgery. The question therefore is not, When is hormone replacement therapy a reasonable alternative to parathyroid surgery? but rather, When is parathyroid surgery a reasonable alternative to hormone replacement therapy?