Parathyroid Hormone and Periosteal Bone Expansion

Excess parathyroid hormone (PTH) has been linked to the periosteum for many years. Subperiosteal resorption of the phalanges is the most sensitive radiographic sign of osteitis fibrosa1, 2 and periosteal neosteosis used to be a frequent component of renal osteodystrophy.2, 3 With earlier diagnosis of both primary4 and secondary5 hyperparathyroidism, these observations became little more than clinical curiosities, but the link between PTH and the periosteum recently has become of more than academic importance.6 When used in the treatment of osteoporosis, PTH increases bone formation rate within a few weeks on the periosteal as well as the endocortical and cancellous surfaces of iliac bone biopsy specimens.7 After several years, there is increased iliac cortical thickness,8 but whether this resulted from increased net bone formation on the endocortical or periosteal surfaces or both could not be determined. Some old observations, presented at meetings but never published in full,9, 10 may shed some light on this question. Between 1962 and 1971, Martyn Lloyd and I saw almost every patient with primary hyperparathyroidism at the Royal Brisbane Hospital in Queensland, Australia; an X-ray of the hands was part of our standard workup. During the same period, I was in charge of a general medical service and obtained hand X-rays on every patient with a stable plasma creatinine of at least 1.5 mg/dl. I took copies of all these films to Detroit, MI, in 1971; there were 41 patients with primary and 86 with presumed secondary hyperparathyroidism,11 none of whom had any symptoms of bone disease or renal disease or hand X-rays that showed subperiosteal erosion with meticulous expert examination.2 Using Helios calipers the length (L) and outer (D) and inner (d) diameters at the midpoint of the second left metacarpal were measured. In both groups there had been excess cortical bone loss,9 indicated by increased d and decreased cortical thickness (D − d) and percent cortical area ((D2 − d2)/D2 ∗︁ 100). Of greater importance in the this context, both groups showed periosteal bone growth indicated by increased D and unchanged L (Table 1). Soon after I moved to Detroit, MI, we acquired the capability for single photon absorptiometry (SPA) of the radius using the Norland-Cameron instrument, which measured bone mineral (BM; g/cm), bone width (BW; cm), and BM/BW (g/cm2) at standard proximal and distal sites.14 By 1983 we had accumulated data on 115 white women with primary hyperparathyroidism and 284 skeletally healthy, white women who were attending the same general medical clinic.10, 14 The groups did not differ in age, height, or weight.10 As for metacarpal morphometry, there had been significant cortical bone loss in the patients with hyperparathyroidism, indicated by reduced values for BM and BM/BW.10 Also, as in the metacarpal there had been periosteal bone growth, indicated by a significant increase in BW at the proximal but not at the distal site (Table 2). Because of its flared shape, the measurement of width is less precise in the distal than in the proximal radius. Collectively, the data indicate that a chronic mild increase in PTH secretion, not severe enough to cause osteitis fibrosa, is accompanied by increased net bone formation beneath the periosteum in the peripheral skeleton. Assuming a cylindrical shape, the absolute increase in circumferential periosteal deposition in primary hyperparathyroidism was 0.38 mm in the metacarpal and 0.19 mm in the radius, a difference that could reflect greater mechanical stimulation at the former site.15 If PTH secretion had been increased for 20 years before diagnosis, the rate of periosteal expansion would have been ∼10–20 μm/year, which is considerably faster than the maximum rates calculated for the normal ilium of 4 μm/year,16 and observed in the metacarpal of 5 μm/year.17 Because bone further from the neutral axis contributes more to bone strength,18 periosteal growth may explain why the increase in fracture risk in hyperparathyroidism is of smaller magnitude and is more difficult to show than would be expected for the degree of cortical thinning,19 but I am more concerned here with the mechanism. Very slow net periosteal growth after cessation of longitudinal growth is a general feature of the normal adult human skeleton16, 17, 20-23 but its cellular basis remains obscure, because of continued uncertainty about whether adult periosteal bone cell activity should be classified as “modeling” or “remodeling.”16, 18 In the growing skeleton there is more or less continuous expansion of all bones by periosteal apposition; bone formation on the same surface for prolonged periods without interruption is one component of modeling.18 Continuation of the same process at a much-reduced rate18, 20-22 could account for slow net periosteal growth in the adult skeleton. Stimulation of this process by PTH could reflect increased recruitment of osteoblasts from local periosteal connective tissue precursors.24 An alternative explanation would be inhibition of periosteal osteoblast apoptosis, but this effect usually is restricted to intermittent rather than continuous PTH excess.25 The only systematic tetracycline-based study of the human periosteum was more consistent with remodeling than with continued modeling.16 Bone formation rate and the extents of osteoclast- and osteoblast-covered surfaces were all much lower than on the endocortical surface but in similar proportion, and cement line configuration was more often scalloped than smooth.16 The association of remodeling with net bone gain at the periosteal surface and net bone loss at the endocortical surface can be explained by differences in cell recruitment and lifespan during each remodeling cycle, reflecting the operation of local factors.26 A PTH-induced increase in activation frequency would be expected to amplify the existing bone balance on each surface without changing its direction, accounting for increases in periosteal gain and endocortical loss by a single mechanism. Although many observers continue to believe that modeling is the dominant mode of periosteal bone cell activity,18 the evidence from the ilium has not been challenged. At first sight, it is paradoxical that intermittent and continued PTH excess appear to have opposite effects on the endocortical surface but the same effects on the periosteal surface. Nevertheless, on both surfaces the cellular responses to intermittent and continuous PTH are different. The increase in periosteal bone formation rate with intermittent PTH administration6, 7 happens too quickly to be the result of increased remodeling with positive balance, and could not be solely caused by abrogation of osteoblast apoptosis because the normal periosteal bone formation rate is so low.16 By exclusion there must be direct transformation of quiescent surface to bone formation, a process that has been termed renewed modeling.27 Intermittent PTH renews modeling and postpones osteoblast apoptosis28 on both periosteal and endocortical surfaces, but renewed modeling is more important on the periosteal surface and postponement of apoptosis is more important on the endocortical surface (Table 3). By contrast, continuous PTH excess increases remodeling on both periosteal and endocortical surfaces. The existing balance (gain or loss) is maintained, but its magnitude is increased by the increase in activation frequency26 (Table 3). Thus, the effects of intermittent and continuous PTH differ consistently on each surface, and each mode of PTH increase has the same effects on both surfaces (Table 3).