Chapter 6 - Hyperparathyroidism in chronic kidney disease

Updated 1 October 2009

Abbreviations

[Ca2+e] : extracellular Ca2+

CaR : Ca2+-sensing receptor

CKD : chronic kidney disease

EGF-R : epidermal growth factor receptor

1° hyperparathyroidism : primary hyperparathyroidism

2° hyperparathyroidism : secondary hyperparathyroidism

Ki67 : cell-cycle linked antigen

MEN-1 : multiple endocrine neoplasia type-1

PCNA : cell-cycle linked antigen (“ proliferating cell nuclear antigen ”)

PTH : parathyroid hormone

iPTH : intact PTH

PTHrp : parathyroid hormone related peptide

PTX : parathyroidectomy

TGF- : transforming growth factor-

VDR : vitamin D receptor

Chronic kidney disease (CKD) is almost constantly associated with a systemic disorder of mineral and bone metabolism, which has recently been named CKD-MBD (1). It is manifested by either one or a combination of biochemical abnormalities (abnormal calcium, phosphorus, PTH, or vitamin D metabolism), bone abnormalities (abnormal bone turnover, mineralization, volume, linear growth, or strength) and vascular or other soft tissue calcification. This disorder generally becomes apparent in CKD stage 3, i.e. a glomerular filtration rate between 60 and 30 ml/min x 1.73 m2. Initially, it is characterized by a tendency towards hypocalcemia, fasting normo or hypophosphatemia, and diminished plasma calcitriol concentration, together with a progressive increase in plasma intact parathyroid hormone (iPTH) (2) and the development of osteitis fibrosa (3). The latter is the consequence of a longstanding stimulation of bone turnover by excessive PTH secretion. In the recent past, however, the increasing recognition of this complication led to frequent oversuppression of PTH by the administration of excessive calcium and/or vitamin D supplements, with its skeletal consequence of iatrogenic low-turnover bone disease, also called adynamic bone disease. Nephrologists became progressively aware of the fact that the abnormally high serum phosphorus levels associated with either hyper or hypoparathyroidism might be detrimental to CKD patients, not only in terms of abnormal bone structure and strength, but also in terms of the relative risk of soft-tissue calcification and cardiovascular as well as all-cause mortality (4-6). As to serum PTH, observational studies consistently reported an increased relative risk of death in CKD stage 5D patients who have PTH values at the extremes, that is less than two or greater than nine times the upper normal limit of the assay. For PTH values within this range, reports of associations with relative risk of cardiovascular events or death were inconsistent. Of note, however, a recent report in elderly men of the community identified a strong association between plasma iPTH in the normal range and cardiovascular mortality (7).

Sequence of events in early kidney failure

Phosphate retention. The precise sequence of metabolic anomalies in incipient CKD leading to secondary (2°) hyperparathyroidism remains a matter of debate. Many years ago, it was postulated that a retention of phosphate in the extracellular space due to the decrease in glomerular filtration rate and the accompanying reduction in plasma ionized calcium concentration was the primary event in the pathogenesis of 2° hyperparathyroidism. These anomalies would only be transient and a new steady state would rapidly occur, with normalized plasma calcium and phosphorus in response to excessive PTH secretion and the well-known effect of this hormone on tubular reabsorption of phosphate (“trade-off hypothesis” of Bricker and Slatopolsky) (8). However, this hypothesis has become less attractive since it was demonstrated that plasma phosphorus is not often elevated in early CKD. It is generally normal until CKD stage 4-5 (2) and may be even moderately diminished in some cases (9), and the urinary elimination of phosphate after an oral overload is actually accelerated (9). Nonetheless, one could argue that in early kidney failure normal or even subnormal concentrations of plasma phosphorus might be observed subsequent to a slight, initial increase, causing an enhanced release of PTH which in turn corrects plasma phosphorus immediately, due to a permanent inhibition of tubular phosphate reabsorption. Of note another, more recent study identified slight increases of plasma phosphate in a large US population sample (NHANES III) with CKD stage 3, that is a creatinine clearance of 50-60 ml/min, as compared to a healthy control population without evidence of renal disease (10). It is possible that subtle changes in circulating and local factors involved in the control of phosphate balance determine the actual level of plasma phosphorus in CKD patients.

Fibroblast growth factor-23 (FGF-23), a recently identified phosphatonin, is one of these additional factors, if not a major new player in the control of phosphate metabolism. It is mainly produced by osteocytes and osteoblasts. It decreases plasma phosphorus by reducing tubular phosphate reabsorption similar to, but independent of, PTH. Moreover, in contrast to PTH, it decreases the renal synthesis of calcitriol. To activate its receptor on proximal tubular epithelial cells it requires the presence of klotho, which confers receptor specificity for FGF23 (11). The regulation of FGF23 production and its interrelations with PTH, calcitriol, and phosphorus are complex and not yet completely understood. Serum FGF23 increased when serum PTH levels were drastically reduced after parathyroidectomy in uremic patients (12) but FGF23 was also increased in PTH-cyclin D1 transgenic mice, a model of primary hyperparathyroidism with high PTH levels, compared to wild-type mice (13). FGF23 decreased PTH gene expression and secretion in vitro and in rats in vivo via a direct effect on the parathyroid cell (14, 15). Calcitriol administration has been shown to increase serum FGF23 in uremic rats (16). Acute phosphate loading or deprivation does not change serum FGF23 in healthy volunteers (17) or in patients with CKD (18), whereas chronic phosphate loading increased it in normal and uremic mice, respectively (16, 19). Circulating FGF23 levels increase with the progression of chronic renal failure, are independently associated with serum phosphorus, calcium, PTH, and calcitriol (17, 20), and may thereby contribute to the development of 2° hyperparathyroidism. Figure 1 shows the complex interrelations between serum FGF23, calcium, calcitriol, and parathyroid function.

Figure 1. Complex interrelations between serum FGF23, calcium, calcitriol and regulation of parathyroid function.

Calcium deficiency. In early CKD stages, disturbances of calcium metabolism may already be present. They include a calcium deficiency state due to a decrease in oral calcium intake and an impairment of active intestinal calcium absorption, a tendency towards hypocalcemia due to skeletal resistance to the action of PTH, and a reduced expression of the calcium-sensing receptor (CaR) in the parathyroid cell. All these factors contribute to the development of parathyroid overfunction (21). Their relative importance increases with the progression of CKD. It also depends on individual patient characteristics such as the underlying type of nephropathy, comorbidities, dietary habits, and amount of food intake.

Inhibition of calcitriol synthesis. A primary role of the disturbed renal synthesis of calcitriol in incipient CKD has become the preferred hypothesis in the last decade. A relative or absolute impairment of renal calcitriol production could be an important player in the initial sequence of events leading to secondary 2° hyperparathyroidism. The major underlying cause would be a reduced tubular transformation of 25OH vitamin D (calcidiol) to calcitriol, due to an intracellular accumulation of phosphate and/or the action of FGF23 whose serum levels increase with the degree of CKD and are a predictor of CKD progression (22). In addition, the progressive loss of nephron mass and the well-known tendency towards metabolic acidosis could also play a role.

Another hypothesis has been proposed, based on the observation that calcidiol does not penetrate into proximal tubular epithelium from the basolateral side, but only from the luminal side. Thus it has been shown that the complex formed by calcidiol and its binding protein (DBP) are ultrafiltered by the glomerulus, subsequently enters this epithelium from the apical side via the multifunctional brush border membrane receptor megalin, and then serves as substrate for the renal enzyme, 1-OH vitamin D hydroxylase for calcitriol synthesis (23) (Figure 2). In case of a reduction of glomerular filtration rate a decreased transfer of the calcidiol-DBP complex into the proximal tubular fluid would occur and hence a reduced availability of calcidiol substrate for luminal reabsorption and ultimately calcitriol formation. However, the validity for the human situation, of this mechanism established in the mouse, has subsequently been questioned since 1-OH vitamin D hydroxylase expression was found not only in proximal, but also in distal tubular epithelium of human kidney, that is in tubular areas in which megalin apparently is not expressed (24).

Figure 2. Schematic representation of the role of megalin, a multifunctional brush border membrane receptor, in the renal tubular reabsorption of 25 OH vitamin D (calcidiol), allowing thereby the synthesis of 1,25 diOH vitamin D by 1-alpha 25 OH vitamin D hydroxylase.

In addition, resistance to the stimulatory action of PTH on renal tubular 1-OH vitamin D hydroxylase activity probably occurs as well. A disturbance of the calcitriol synthesis pathway could explain the long reported direct relation in CKD patients between plasma calcidiol and calcitriol, and between plasma calcitriol and glomerular filtration rate (25). Such relations are not observed in subjects with normal renal function.

Finally, the concentration of plasma calcidiol is diminished in the majority of patients with CKD (26, 27). The reasons for this deficiency state include insufficient hours of sunshine or sun exposure, skin pigmentation, ageing, intake of antiepileptic drugs, enhanced urinary excretion of calcidiol complexed to vitamin D binding protein (DBP) in presence of heavy proteinuria, or loss into the peritoneal cavity during peritoneal dialysis treatment. All these factors may contribute to a reduction in calcitriol synthesis (28). However, low plasma calcidiol has also been postulated to be a risk factor per se for hyperparathyroidism, based on an observational study in Algerian hemodialysis patients with insufficient exposure to sunshine (29).

Disturbances during advanced kidney failure

The above mentioned roles of relative or absolute deficiency states of calcium and vitamin D are steadily gaining importance with the progression of CRF, and phosphate becomes a major player.

The role of hyperphosphatemia. In advanced stages of CRF hyperphosphatemia is a nearly constant feature, due to phosphate retention caused by the progressive loss of functioning nephrons characterized by an increasing difficulty to augment the filtered load of phosphate and to further reduce its tubular reabsorption when it is already maximally inhibited. A high extracellular phosphate concentration stimulates FGF23 production and PTH secretion, the latter both directly and indirectly (see below).

The uremic syndrome itself could also play a role. Thus several uremic toxins, that is substances which accumulate in the uremic state, have been shown to interfere with vitamin D metabolism and action (30, 31).

Mechanisms involved in the induction of 2° hyperparathyroidism

Generally speaking, there are at least two different mechanisms which determine the magnitude of 2° hyperparathyroidism. The first results in an increase in PTH synthesis and secretion, and the second in an increase in parathyroid gland mass, mostly due to enhanced cell proliferation (hyperplasia) and to a lesser degree due to an increase in cell size (hypertrophy) (see schematic representation in Figure 3). Whereas acute stimulation of PTH synthesis and/or release generally occurs in the absence of cell growth stimulation, these two processes appear to be tightly linked whenever there is longstanding stimulation. The main factors involved in the control of the two processes are again calcitriol, calcium, phosphate, whereas the effects of FGF23 appear to be limited to the control of PH synthesis and secretion. In the following, the disturbances of the mechanisms controlling parathyroid function will be discussed subsequently for each of these three factors, although there are numerous interactions between them. Thereafter, the influence of other factors and comorbid conditions related to renal failure will be discussed.

Figure 3. Schematic representation of parathyroid hormone (PTH) synthesis and secretion (upper part) and parathyroid cell proliferation and apoptosis (lower part), as regulated by a number of hormones and growth factors.

Calcitriol. The above mentioned decrease in plasma calcitriol aggravates hyperparathyroidism via several mechanisms. The first is direct and results from an insufficient inhibition of PTH synthesis due to low circulating calcitriol levels and a disturbed action of calcitriol at the level of the preproPTH gene. It is well established that calcitriol, after forming a complex with its receptor, vitamin D receptor (VDR) and heterodimerizing with the retinoic acid receptor (RXR), directly inhibits preproPTH gene transcription by binding to a specific DNA response element (VDRE) located in the 5’-flanking region of the gene. In CKD, in addition to low extracellular concentrations of calcitriol, at least two other factors interfere with calcitriol’s action on the preproPTH gene (32). The first factor is a reduced expression of the VDR gene product in hyperplastic parathyroid tissue of CKD patients (33). This reduction is particularly marked in nodular parathyroid tissue, as compared to diffusely hyperplastic tissue. The second factor is reduced binding of calcitriol to VDR, a slowed nuclear migration of the calcitriol–VDR complex and a less efficient action on the preproPTH gene, in association with the uremic state (31, 34). Of note, extracellular calcium appears to play a role in the regulation of VDR expression. Thus, Garfia et al. showed that low extracellular Ca reduced VDR expression in rat parathyroid glands independently of calcitriol, whereas high extracellular Ca reduced VDR expression (35). Hypocalcemia may prevent by this mechanism the feedback of increased plasma calcitriol concentrations on the parathyroids.

The second level at which calcitriol regulates PTH gene expression involves calreticulin. Calreticulin is a calcium binding protein which is present in the endoplasmic reticulum of the cell, and also may have a nuclear function. It regulates gene transcription via its ability to bind a protein motif in the DNA-binding domain of nuclear hormone receptors of sterol hormones. Sela-Brown et al. proposed that calreticulin might inhibit vitamin D's action on the PTH gene, based on in vitro and in vivo experiments (36). They fed rats either a control diet or a low calcium diet, which led to increased PTH mRNA levels despite high serum calcitriol levels that would be expected to inhibit PTH gene transcription. Their postulate that high calreticulin levels in the nuclear fraction might prevent the effect of calcitriol on the PTH gene was strongly supported by the observation that hypocalcemic rats had increased levels of calreticulin protein in their parathyroid nuclear fraction. This could explain why hypocalcemia leads to increased PTH gene expression despite high serum calcitriol levels, and might also be relevant for the refractoriness of 2° hyperparathyroidism to calcitriol treatment observed in many CRF patients.

The third mechanism of calcitriol action could be indirect, via a stimulatory effect on parathyroid CaR expression, as shown by Brown et al (37) and subsequently confirmed by Mendoza et al (38).

The fourth mechanism by which calcitriol acts on parathyroid tissue is again a direct one. It concerns the well-known inhibitory effect of vitamin D on cell proliferation and the induction of differentiation towards mature, slowly growing cells. A decrease in plasma calcitriol and a perturbed action at molecular targets favors abnormal cell growth. This is the case with parathyroid tissue as well, and secondary parathyroid hyperplasia ensues (39). The importance of vitamin D in the parathyroid hyperplasia of experimental uremia has first been shown by Szabo et al (40). These authors administered increasing doses of calcitriol to rats either at the time of inducing chronic renal failure or at a later time point, when uremia was already well established. They were able to prevent parathyroid cell proliferation entirely when calcitriol was given at start of uremia, but not when given later on. Fukagawa et al showed that pharmacologic doses of calcitriol repressed c-myc expression in the parathyroid tissue of uremic rats and suggested that the hormone might suppress parathyroid hyperplasia by this pathway (41). In contrast, Naveh-Many et al. (42) failed to observe such an antiproliferative effect of calcitriol in parathyroid cells of uremic rats but they administered the hormone for only three days. Such short-term administration may not have been sufficient for an efficacious suppression of cell turnover.

To answer the question of a possible direct calcitriol action on parathyroid cells, several studies were performed in experimental models in vitro. Nygren et al. (43) showed in primary cultures of bovine parathyroid cells maintained in short-term culture, that these cells underwent significant increases both in number and size in response to fetal calf serum, and that the addition of 10-100 ng/ml calcitriol almost completely inhibited cell proliferation whereas hypertrophy was unaffected. Kremer et al (44) subsequently confirmed in same parathyroid cell model that calcitriol exerted an anti-proliferative action. They further suggested that this inhibition occurred via a reduction of c-myc mRNA expression. One report showed an inhibitory action under long-term culture conditions (up to 5 passages) of the effect of calcitriol on bovine parathyroid cell proliferation (45). Our group subsequently confirmed such a direct antiproliferative effect of calcitriol in a human parathyroid cell culture system derived from hyperplastic parathyroid tissue of patients with severe 2° uremic hyperparathyroidism (46) (Figure 4).

Figure 4. Antiproliferative effect of increasing medium calcitriol concentrations in the incubation milieu of a human parathyroid cell culture system derived from hyperplastic parathyroid tissue of patients with severe 2° uremic hyperparathyroidism.

A final mechanism is the potential association between parathyroid function and vitamin D receptor (VDR) polymorphism. Fernandez et al (47) separated hemodialysis patients with same serum calcium and time on dialysis treatment into two groups, according to their serum iPTH levels, namely low PTH (<12 pmol/L) or high PTH (>60 pmol/L). They found that the BB genotype and the B allele were significantly more frequent in the low PTH than in the high PTH group (32.3 % vs 12.5 %, and 58.8% vs 39.1%, respectively). This information suggests that VDR gene polymorphism influences parathyroid function in chronic renal failure. Similar results have been reported by an Italian group (48) and in a large sample of Japanese hemodialysis patients (49). In this latter study, after excluding diabetics and patients with less than ten years on dialysis treatment, the authors observed lower plasma iPTH levels in patients with BB than with Bb or bb alleles. A relationship between Apa I polymorphism (A/a alleles) and the severity of hyperparathyroidism has also been sought in Japanese hemodialysis patients (50). Plasma PTH levels in AA and Aa groups were approximately half that of the aa group. However, other groups found no difference in PTH levels for various VDR polymorphisms (51). Moreover, although in some clinical conditions VDR polymorphism may be associated with variations of the half life of the VDR gene transcript (52) or of VDR function (53), there has been no report showing that in uremic patients with 2° hyperparathyroidism the density of parathyroid cell VDR varies with different VDR genotypes. In addition, although VDR genotypes may have some influence on the degree of 2° parathyroid hyperplasia, the mechanism by which this may occur remains unknown at present.

Calcium. It has long been known that extracellular Ca2+ is the major regulator of PTH secretion. Small changes in Ca2+ concentration result in immediate changes of PTH release which are short-lived or long-lived, depending on the velocity of the restoration of Ca2+ towards its normal level. A recent study showed that postprandial urinary calcium excretion was increased in patients with CKD as in healthy volunteers, but this was accompanied by significantly reduced serum calcium and increased PTH levels in the CKD patients only (18). The inverse relation between Ca2+ and PTH in the circulation obeys a sigmoidal curve (54). While the majority of in vitro studies have reported a decreased responsiveness of hyperplastic parathyroid cells to calcium, in vivo studies have not always led to the same conclusion. This is likely due to different methods used to assess the dynamics of PTH secretion (55).

Several in vitro studies have shown that the set point of calcium for PTH secretion (that is the calcium concentration required to produce half maximal PTH secretion) is greater in parathyroid cells from 1° (adenomas) and 2° (uremic) hyperplastic parathyroid glands than in normal parathyroid cells (56). Such a relatively poor response to calcium should contribute to the increased PTH levels observed in uremic patients with 2° hyperparathyroidism.

We and others have demonstrated that both 1° parathyroid adenoma and 2° uremic, hyperplastic parathyroid gland tissue exhibit a decrease in the expression of CaR protein (57, 58). In 2° uremic hyperparathyroidism, there is a significant CaR decrease in diffusely growing hyperplastic tissue, although the decrease is even more marked in nodular areas (characteristic of advanced hyperplasia) (57) (Figure 5). Since changes in intracellular calcium elicited by hyper or hypocalcemia depend on the CaR, its decreased expression explains, at least in part, the impaired intracellular calcium response to extracellular calcium and hence a reduced inhibitory effect of calcium on PTH secretion.

Figure 5. Calcium-sensing receptor expression in normal parathyroid glands, 1° hyperparathyroidism (adenomas), and 2° hyperparathyroidism (glands exhibiting diffuse or nodular hyperplasia, from dialysis patients). Decreased expression of both CaR protein and mRNA in the majority of hyperplastic glands, with a particularly marked decrease in nodular type 2° uremic hyperparathyroidism.

Almaden et al studied the calcium-regulated PTH response in vitro, using respectively primary parathyroid adenoma and uremic hyperplastic tissue, the latter both of the nodular and the diffuse type (59). They found that in parathyroid adenomas PTH secretion was less responsive to an increase in extracellular calcium than in 2° hyperplasia; among the latter, nodular tissue was less responsive than diffusely hyperplastic tissue. The decreased secretory response to calcium observed in nodular hyperplasia may be explained by the markedly reduced CaR expression in CKD, as demonstrated by Gogusev et al (57). This decrease can be overcome, at least partially, by PTHrp, as shown by Lewin et al (60). These authors observed that the the administration of PTHrp significantly stimulated the impaired secretory capacity of the parathyroid glands of uremic rats in response to hypocalcemia. Of note, this observation also implies that the PTH/PTHrp receptor is expressed on the parathyroid cell.

The shift of the calcium set point to the right in dialysis patients in vivo has been a much less constant finding than the right shift observed in the above mentioned studies in uremic parathyroid tissue in vitro. While in CKD patients with a mild to moderate degree of hyperparathyroidism the set point was most often found to be normal, an altered set point was observed in presence of severe parathyroid overfunction with hypercalcemia (61). This anomaly could at least in part be due to CaR down-regulation. In CKD patients with less severe parathyroid overfunction, there was actually a considerable controversy regarding the results of in vivo assessments of parathyroid gland function (62, 63). In part, disparities among reports reflected technical differences in experimental methods and/or variations in the mathematical modeling of PTH secretion in vivo (64). Another difficulty in interpreting the results of in vivo dynamic tests of parathyroid gland function relates to the issue of parathyroid gland size. Because there is a basal, or non-suppressible, component of PTH release from the parathyroid cell even at high extracellular calcium concentrations, excess PTH secretion may result solely from increases in parathyroid gland mass (61). This can occur in the absence of a defect in calcium-sensing at the level of the parathyroid cell. Since parathyroid gland hyperplasia is present to some extent in nearly all patients with chronic renal failure, alterations in PTH secretion due to increases in parathyroid gland mass cannot readily be distinguished from those attributable to changes in calcium-sensing by the parathyroid cell using the four parameter model for in vivo studies (63).

The role of calcium in parathyroid cell proliferation is less clear than is generally assumed. Calcium deficiency, in the presence or absence of hypocalcemia, together with vitamin D deficiency or reduced generation of calcitriol, probably is a major stimulus of parathyroid hyperplasia. Thus Naveh-Many et al showed that calcium deprivation, together with vitamin D deficiency, greatly enhanced the rate of parathyroid cell proliferation in normal rats and also in rats with chronic renal failure, using the cell cycle-linked antigen, PCNA (42). The concomitant decrease in CaR expression in CRF, as observed in parathyroid glands of both dialysis patients and uremic rats (57, 65), should theoretically enhance parathyroid tissue hyperplasia further. Indirect support for this contention came from the observation that the administration of the calcimimetic compound NPS R-568, a calcium-sensing receptor agonist, led to the suppression of parathyroid cell proliferation in rats with renal insufficiency (66). However, in the study by Naveh-Many et al the dietary regimen was poor in both calcium and vitamin D. In contrast, when feeding normal rats on a calcium-deficient diet alone, in the absence of concomitant vitamin D deficiency, Wernerson et al observed parathyroid cell hypertrophy, not hyperplasia (67).

The question whether the effect of calcium is direct or indirect remains therefore unsolved at present. It can only be answered by in vitro studies. Unfortunately, until recently none of the available culture systems using normal parathyroid cells allowed the maintainance of functionally active cells for prolonged time periods. They were all characterized by a rapid and significant loss of PTH secretion, within 3 to 4 days (68-70). One culture model has been described, using bovine parathyroid cell organoids, which maintained the ability to modulate PTH secretion in response to extracellular Ca2+ [Ca2+e] and tissue-like morphology for 2 weeks in culture (71). However, only one long-term study of bovine parathyroid cells demonstrated a release of bioactive bovine PTH but with reduced sensitivity to calcium (72). Other reports showed that the rapid decrease in PTH responsiveness of cultured bovine parathyroid cells to changes in [Ca2+e] was associated with a marked reduction in CaR expression (73, 74). Yet other parathyroid cell-derived culture models proposed in the literature were in fact devoid of any PTH secretory capacity (75, 76).

To study direct effects of [Ca2+e] on the parathyroid cell in vitro, we developed a functional human parathyroid cell culture system capable of maintaining regulation of its secretory activity and the expression of extracellular CaR mRNA and protein for several weeks. For this purpose, we used parathyroid cells derived from hyperplastic parathyroid tissue of dialysis patients with severe 2° hyperparathyroidism (77). In a subsequent study, we obtained evidence with this experimental model that parathyroid cell proliferation index, as estimated by [3H]-thymidine incorporation into an acid-precipitable fraction as a measure of DNA synthesis, could be directly stimulated by high [Ca2+e] in the incubation medium, compared with low [Ca2+e] (46) (Figure 6). We confirmed this finding in independent experiments using the cell cycle-linked antigen Ki-67 to determine parathyroid cell proliferation. However, the addition of the calcimimetic NPS R-467 to the incubation medium led to a decrease in proliferation (Figure 7). Of interest, calcimimetics have subsequently been shown to upregulate the expression of the CaR (38, 78) and the VDR (38) in parathyroid glands of uremic rats. In an attempt to unify our apparently contradictory in-vitro observations with respect to findings made in vivo, we propose the following hypothesis. The effect of calcium on parathyroid cell proliferation could occur along two different pathways, via two distinct mechanisms. Inhibition of proliferation would occur via the well-known parathyroid CaR-dependent pathway, whereas stimulation of proliferation would occur via a second pathway (Figure 8). It must be noted that all parathyroid tissue samples used in our study stemmed from uremic patients with long-term renal failure and severe secondary hyperparathyroidism. Since such parathyroid tissues generally exhibit decreased CaR expression, it is possible that the number of CaR expressed in the parathyroid cell membranes of our culture model was insufficient to inhibit cell proliferation. Of note, the human CaR gene has two promoters and two 5’ untranslated exons and that the alternative usage of these exons leads to production of multiple CaR mRNAs in parathyroid (79). The expression of CaR mRNA produced by one of the two promoters of CaR gene is specifically reduced in PT adenomas, suggesting a role in PTH hypersecretion and proliferation. Moreover, the membrane-bound 550-kD Ca2+-binding glycoprotein megalin, belonging to the low-density lipoprotein receptor superfamily, has been identified in parathyroid chief cells as another putative calcium-sensing molecule which could be involved in calcium-regulated cellular signalling processes as well (80). Based on these observations, one can postulate that parathyroid cells express multiple CaR-like molecules. Consequently, if the well-known parathyroid CaR is down-regulated, parathyroid cell proliferation by calcium may occur via a different CaR isoform. Another possibility is an alteration in post-receptor signal transduction that could occur in hyperparathyroid states or under cell culture conditions. Our observations are in line with findings by Ishimi et al. which were incompatible with a direct effect of low [Ca2+e] in the pathogenesis of parathyroid hyperplasia (45). However, the extrapolation from such in vitro observations to the in vivo setting should be done with caution, and further work is needed to define the precise pathway(s) by which calcium regulates parathyroid tissue growth.

Phosphate. Hyperphosphatemia is associated with increased PTH secretion. The stimulation of PTH release occurs via direct and indirect mechanisms. The initially proposed indirect mechanism, which remains true according to present knowledge, is via a decrease in plasma Ca2+ concentration (see above). Hyperphosphatemia also leads to an inhibition of the renal synthesis of calcitriol, probably mostly via stimulation of FGF23 production.

A direct action of phosphate on PTH secretion by the parathyroid cell has long been suspected. However, it has been formally demonstrated in vitro only in 1996 (81). This demonstration required the use of either intact parathyroid glands (from rats) (Figure 9) or parathyroid tissue slices (from cows) whereas it had not been possible to obtain such a direct stimulation using the classic model of isolated bovine parathyroid cells. The elevation of plasma phosphate concentration in the incubation milieu of the experimental models using intact (or partially intact) parathyroid tissue leads to a stimulation of PTH secretion within some hours, in the absence of any change in extracellular ionized calcium concentration. It can however be abrogated by an increase in cytosolic calcium (82).

Figure 6. Stimulatory effect on parathyroid cell proliferation (measured by KI-67 staining method) of high medium calcium concentrations in the incubation milieu of a human parathyroid cell culture system derived from hyperplastic parathyroid tissue of patients with severe 2° uremic hyperparathyroidism.

Figure 7. Inhibitory effect of the calcimimetic NPS R-467 on parathyroid cell proliferation (measured by [3H]-thymidine method) of high medium calcium concentrations in the incubation milieu of a human parathyroid cell culture system derived from hyperplastic parathyroid tissue of patients with severe 2° uremic hyperparathyroidism.

Figure 8. Schematic representation of the regulation of parathyroid cell proliferation by extracellular calcium concentration, involving an inhibitory pathway via the calcium-sensing receptor, and a stimulatory pathway via an unknown transmembrane transduction mechanism. Normally, pathway 1 predominates over pathway 2 in parathyroid tissue. In presence of parathyroid hyperplasia with calcium-sensing receptor down-regulation pathway 2 could become dominant and favor parathyroid cell proliferation over suppression.

Figure 9. Direct inhibition of parathyroid hormone secretion by increases in phosphate concentration in the incubation medium bathing intact parathyroid glands from normal rats.

Silver’s group reported subsequently that phosphate, like calcium, regulates pre-pro-PTH gene expression post-transcriptionally by changes in protein-PTH mRNA interactions at the 3'-UTR which determine PTH mRNA stability. They identified the minimal sequence for protein binding in the PTH mRNA 3'-UTR and determined its functionality. They found that the conserved PTH RNA protein-binding region conferred responsiveness to calcium and phosphate and determined PTH mRNA stability and levels (83). Thus a low calcium diet increased stability, whereas a low phosphate diet decreased stability of PTH mRNA (Figure 10). The PTH mRNA 3’-untranslated region-binding protein was subsequently identified by this research group as AUF1 (84).

Figure 10. Post-transcriptional regulation of PTH mRNA stability by calcium, phosphorus, and chronic renal failure. Pre-pro-PTH gene expression is modulated via changes in protein-PTH mRNA interactions at the 3'-UTR region which determine PTH mRNA stability. Low calcium diet increases stability, whereas low phosphate diet decreases stability of PTH mRNA. The PTH mRNA 3’-untranslated region-binding protein was subsequently identified as AUF1.

In addition to its stimulatory effect on PTH secretion a high phosphate diet also rapidly induces parathyroid hyperplasia. It has long been shown in experimental animal models that a phosphate-rich diet induced an increase in parathyroid gland function and volume (85). Subsequently, studies showed that phosphate-rich diets, when fed to animals with chronic renal failure leading to high plasma phosphate levels, induced parathyroid hyperplasia even when changes in plasma calcium and calcitriol concentration were carefully avoided, pointing to a direct effect of phosphate on cell proliferation (42, 86). Conversely, early dietary phosphate restriction was capable of preventing both PTH oversecretion and parathyroid hyperplasia (42, 86, 87). Interestingly, dietary phosphate restriction following phosphate overload in rats also led to an immediate decrease in PTH secretion, in the absence of a regression of parathyroid gland size (88).

Our group wished to know whether the stimulatory effect of phosphate on parathyroid cell proliferation was direct or indirect. To answer this question, we used the above described in vitro model of human parathyroid cell maintained in long-term culture (77). We could show that the cell proliferation index was directly stimulated by high phosphate concentrations in the incubation medium, compared with low phosphate concentration (46) (Figure 11). These experiments demonstrate that phosphate is capable to stimulate not only PTH secretion, but also to induce parathyroid tissue hyperplasia by a direct mode of action.

Figure 11. Stimulatory effect on parathyroid cell proliferation (measured by KI-67 staining method) of high medium phosphate concentrations in the incubation milieu of a human parathyroid cell culture system derived from hyperplastic parathyroid tissue of patients with severe 2° uremic hyperparathyroidism.

FGF23 plays an important role in the control of plasma phosphorus. Its plasma concentration is elevated in chronic renal failure. This allows a more efficient inhibition of proximal tubular phosphate reabsorption and the maintenance of plasma phosphorus in the normal range. However, since serum phosphorus directly stimulates PTH secretion, its decrease by FGF23 indirectly leads to a reduction of PTH release, in addition to the direct inhibitory action of this growth factor on the parathyroid secretory activity (see above).

Other factors and conditions. As already pointed out above the uremic state is another long suspected, albeit still ill defined factor in the pathogenesis of 2° hyperparathyroidism. Recently, several pieces of evidence have been provided in favor of a role of uremic toxins which interfere with the binding of calcitriol to VDR (31) and with the nuclear uptake of the hormone-receptor complex (34) . This should have consequences not only for PTH synthesis and secretion, but also for parathyroid cell proliferation.

Diabetic patients on dialysis have relatively low plasma PTH levels, compared with non diabetic dialysis patients. The high incidence of low bone turnover in uremic diabetic patients (89-92) has been attributed to low PTH levels, possibly via an inhibition of PTH secretion or a modification of the PTH peptide by the accumulation of advanced glycation end-products such as pentosidine (93). However, experimental studies have demonstrated that the metabolic abnormalities associated with diabetes can also directly decrease bone turnover, independent of PTH (94). In general, patients with low bone turnover tend to develop hypercalcemia when on a normal or high calcium intake, probably due to a diminished skeletal capacity of calcium uptake. This in turn tends to reduce plasma PTH. Thus low bone turnover favors the occurrence of hypoparathyroidism. Another question is whether in diabetic patients abnormalities such as hyperglycemia and insulin deficiency or resistance may directly affect parathyroid function. In an in vitro study using dispersed bovine parathyroid cells, high glucose and low insulin concentrations suppressed the PTH response to low calcium (95). These results are compatible with the view that diabetes directly inhibits parathyroid function. However, in experiments in renal failure rats fed a high phosphorus diet to induce 2° hyperparathyroidism, diabetes did not affect the development of parathyroid overfunction (94).

Aluminum bone disease is generally associated with low serum PTH levels (96, 97) and a decreased PTH response to stimulation with hypocalcemia (98, 99). In such patients, high amounts of aluminum are also found in parathyroid glands (100). The relatively low PTH levels may reflect either an inhibition by the hypercalcaemia commonly observed in this condition (101) or a direct inhibitory effect of aluminum on parathyroid cell function (102). Direct toxic effects of the trace element have also been demonstrated in studies in vitro (103, 104).

Data obtained in experimental animals and results of clinical studies have been more controversial. Whereas some experiments indicated that aluminum did not decrease plasma PTH levels in vivo (103, 104), other experiments were affirmative (105, 106). Whatever the mechanisms involved, clinical data clearly showed that the introduction of aluminum-free dialysis fluid and the discontinuation of aluminum contamination of the dialysate or aluminum removal with deferoxamine resulted in an increase in plasma PTH levels and in PTH response to hypocalcemia (107). Thus, although there appears to be an association between aluminum toxicity and parathyroid gland function, the interaction is complex.

Post-receptor mechanisms involved in polyclonal parathyroid tissue growth

As pointed out above, calcitriol reduces parathyroid cell proliferation by decreasing the expression of the early gene, c-myc. This gene modulates the progression from G1 to S phase in cell cycle. A decrease in plasma calcitriol and/or a disturbance of its action at the level of the parathyroid cell, which are both frequently observed in uremic patients, may cause progression into the cell cycle and disinhibition of c-myc expression. Another mode of action involves the cyclin kinase inhibitor p21WAF1. Calcitriol has long been shown to induce the differential expression of p21WAF1 in the myelo-monocytic cell line U937 and to activate the p21 gene transcriptionally in a VDR-dependent, but p53-independent, manner (108). Slatopolsky’s group further showed that the administration of calcitriol to moderately uremic rats enhanced parathyroid p21 expression and prevented high phosphate-induced increase in parathyroid TGF- content (108). In addition, they found that calcitriol altered membrane trafficking of the epithelial growth factor receptor (EGFR, which binds both EGF and TGF-) and down-regulated EGFR growth signaling (109). Induction of p21 and reduction of TGF-content in the parathyroid glands also occurred when uremia-induced parathyroid hyperplasia was suppressed by high dietary Ca. The mechanisms by which a phosphate-rich diet and hyperphosphatemia induce parathyroid hyperplasia, and conversely a phosphate-poor diet and hypophosphatemia inhibit parathyroid tissue growth, have also been explored by this group in a detailed fashion. Thus, Dusso et al showed that feeding a low phosphate diet to uremic rats increased parathyroid p21 gene expression through a vitamin D-independent mechanism (110). When administering a high phosphate diet there was however no reduction in p21 expression. In this condition, they observed an increase in parathyroid tissue TGF- expression and a direct correlation between this expression and parathyroid cell proliferation rate. This finding is in line with the previous observation by our group of de novo TGF- expression in severely hyperplastic parathyroid tissue of uremic patients (111). Although these findings provide more insight into the pathways by which changes in phosphate intake, and ultimately variations in extracellular phosphate concentration, control parathyroid tissue growth, the exciting question of the transmembrane signal transduction mechanism and subsequent nuclear events triggered by phosphate remains yet to be answered.

In addition to p21 and TGF- , a number of other growth factors and inhibitors may be involved in polyclonal parathyroid hyperplasia. Thus, PTHrp has also been proposed as a posssible growth suppressor in the human parathyroid (112). PTHrp, and probably PTH itself, also exert an inhibitory effect on PTH secretion by acting via a negative feedback loop on the PTH/PTHrp receptor (PTH-R1) which appears to be expressed in the parathyroid cell membrane as well (60). Table 1 summarizes various changes in gene and growth factor expression which are potentially involved in the parathyroid tissue hyperplasia of 2° uremic hyperparathyroidism. Gcm2 has been identified as a master regulatory gene of parathyroid gland development, since Gcm2 knockout mice lack parathyroid glands (113). Correa et al. found high Gcm2 mRNA expression in human parathyroid glands in comparison with other non-neural tissues and underexpression in parathyroid adenomas but not in lesions of HPT secondary to uremia (114). It is unknown whether Gcm2 itself is submitted to regulation by other factors.

Table 1. Changes in gene and growth factor expression potentially involved in parathyroid tissue hyperplasia of 2° uremic hyperparathyroidism

Early immediate genes and receptor genes * enhanced c-myc gene expression (Fukagawa et al, Kidney Int 1991; 39: 874-81) * decreased calcium-sensing receptor (CaR) gene expression (Kifor et al, J Clin Endocrinol Metab 1996; 81: 1598-1606. Gogusev et al, Kidney Int 1997; 51: 328-36) * decreased vitamin D receptor (VDR) gene expression (Fukuda et al, J Clin Invest 1993; 92: 1436-42) Gene polymorphisms * vitamin-D receptor (VDR) gene polymorphism (Olmos et al, Methods Find Exp Clin Pharmacol 1998; 20: 699-707. Fernandez et al, J Am Soc Nephrol 1997 ; 8: 1546-52. Tagliabue et al, Am J Clin Pathol 112: 366-70, 1999) Growth factors and cell cycle inhibitors * increased acidic growth factor (aFGF) gene expression (Sakaguchi, J Biol Chem 1992; 267: 24554-62) * decreased parathyroid hormone-related peptide (PTHrp) gene expression (Matsushita et al, Kidney Int 1999; 55: 130-8) * de novo transforming growth factor-alpha (TGF-alpha) gene expression (Gogusev et al, Nephrol Dial Transplant 1996; 11: 2155-62. ) * induction of TGF-alpha by high phosphate diet (Dusso et al, Kidney Int 2001; 59: 855- 865) * insufficient inhibition of cyclin kinase inhibitor p21WAF1 (Dusso et al, Kidney Int 2001; 59: 855-65) ; p21WAF1can be induced by calcitriol (Cozzolino et al, Kidney Int 60(6): 2109-2117) Gene mutations : association with monoclonal or multiclonal growth * mutation of menin gene (Falchetti et al, J Clin Endocrinol Metab 1993; 76: 139-44. Tahara et al, J Clin Endocrinol Metab 2000; 85: 4113-7. Imanishi et al, J Am Soc Nephrol 2002 ;13 :1490-8) * mutation of Ha-ras gene (Inagaki et al, Nephrol Dial Transplant 1998; 13: 350-7) * no involvement of VDR or CaR gene mutations (Degenhardt et al, Kidney Int 1998; 53: 556-61. Brown et al, J Clin Endocrinol Metab 2000; 85: 868-72)

Mechanisms involved in the transformation of polyclonal to monoclonal parathyroid growth

In severe forms of 2° hyperparathyroidism nodular formations within diffusely hyperplastic tissue are a frequent finding (115). This observation probably corresponds to the occurrence of a monoclonal type of cell proliferation within a tissue which initially exhibits polyclonal growth. Such clonal, benign tumoral growth was initially shown by Arnold et al using chromosome X-inactivation analysis method (116), and subsequently confirmed by other groups (117, 118). In fact, one could also speak of multiclonal proliferation since several different clones may coexist in the same patient, and sometimes even in a single parathyroid gland (Figure 12). Acquired mutations of tumor enhancer or tumor suppressor genes are almost certainly involved in the development of such cell clones but precise knowledge about acquired genetic abnormalities remains limited (117). To identify new locations of parathyroid oncogenes or tumor suppressor genes important in this disease, Imanishi et al performed both comparative genomic hybridization (CGH) and genome-wide molecular allelotyping on a large group of uremia-associated parathyroid tumors (119). One or more chromosomal changes were present in 24% of tumors, markedly different from the values in common sporadic adenomas (28% and 72%, respectively), whereas no gains or losses were found in 76% of tumors. Two recurrent abnormalities were found, namely gain of chromosome 7 (9% of tumors) and gain of chromosome 12 (11% of tumors). Losses on chromosome 11, the location of the MEN1 tumor suppressor gene, occurred in only one uremia-associated tumor (2%), as compared to 34% in adenomas. The additional search for allelic losses with polymorphic microsatellite markers led to the observation of recurrent allelic loss on 18q (13% of informative tumors). Lower frequency loss was detected on 7p, 21q, and 22q. Interestingly, the cyclin D1 oncogene, activated and overexpressed by clonal gene rearrangement or other mechanisms in 20-40% of parathyroid adenomas (120, 121), has not been found to be overexpressed in uremia-associated tumors (121).

Figure 12. Schematic representation of the different types of parathyroid hyperplasia encountered in CKD patients with 2° uremic hyperparathyroidism.

Another interesting question was that of a possible involvement of somatic genes playing a major role in the normal reguation of parathyroid function, such as the CaR and VDR genes. The expression of these two genes was found to be decreased in the hyperplastic parathyroid tissue of uremic patients (33, 57). The decrease was particularly marked in nodular areas, as compared to diffuse areas of parathyroid gland hyperplasia (Figure 13). Moreover, in uremic rats the decrease in CaR expression was inversely related to the degree of parathyroid cell proliferation (122). However, the search for mutations or deletions of the VDR gene or the CaR gene in uremic hyperparathyroidism has remained unsuccessful (117, 123, 124). The question remains unsolved at present whether the downregulation of CaR and VDR expression is a primary event or whether it is secondary to hyperplasia.

Figure 13. Increase in apoptotic parathyroid cells in glands removed from patients with 1° or 2° uremic hyperparathyroidism, as compared to normal parathyroid tissue.

Whether benign parathyroid tumors may evolve towards malignant forms is still subject to debate. Since in dialysis patients parathyroid carcinoma is a rare event (125), malignant transformation of clonal parathyroid neoplasms is probably exceptional.

Thus genome-wide allelotyping and CGH have directly confirmed the presence of monoclonal parathyroid neoplasms in uremic patients with refractory secondary hyperparathyroidism whereas the candidate gene approach has led to only modest results. Somatic inactivation of the MEN1 gene does contribute to the pathogenesis of uremia-associated parathyroid tumors, but its role in this disease appears to be very limited, and there is probably no role for DNA changes of the CaR and VDR genes. Recurrent DNA abnormalities suggest the existence of new oncogenes on chromosomes 7 and 12, and tumor suppressor genes on 18q and 21q, involved in uremic hyperparathyroidism. Finally, patterns of somatic DNA alterations indicate that markedly different molecular pathogenetic pathways exist for clonal outgrowth in severe uremic hyperparathyroidism, as compared to common sporadic parathyroid adenomas. It must be pointed out that our group did not find a correlation between the presence of microscopically evident nodules and the clonal character of resected parathyroid tissue; appearances of several glands with histologic patterns of diffuse hyperplasia also were unequivocally monoclonal in the absence of detectable nodular formations, suggesting that the current criteria for pathological diagnosis do not reflect the genetic differences among these two pathohistological types (116).

Parathyroid cell apoptosis

It is uncertain at present whether a change in the rate of apoptosis can also contribute to parathyroid tissue hyperplasia (39, 126, 127). One research group examined this issue in rats with short-term renal failure (5 days) and failed to detect apoptosis in hyperplastic parathyroid glands (128). However, this failure could be due to lack of sensitivity of the employed methods.

Negative findings in rats, with no identifiable apoptotic figures at all in parathyroid glands (39, 126, 128), contrast with subsequent positive observations in rats by others (129, 130) and with personal observations of significant apoptotic figures in hyperplastic parathyroid glands removed from uremic, severely hyperparathyroid patients during surgery (127). In our study in human glands, we found approximately ten times higher apoptotic cell numbers in uremic patients than in normal parathyroid tissue, using the Tunel method (Figure 13) (127). Of note, the uremic state appears to stimulate apoptosis in other cell types as well such as circulating monocytes (131), possibly via the well-known increase of cytosolic Ca2+ which has been observed in a variety of cell types in renal failure (132), and also possibly via the noxious effect of bioincompatible dialysis membranes used for renal replacement therapy (133). If confirmed by others, the observed enhancement of parathyroid tissue apoptosis could compensate, at least in part, the increase in parathyroid cell proliferation observed in 2° uremic hyperparathyroidism.

Regression of parathyroid hyperplasia?

Whether regression of parathyroid hyperplasia occurs in animals or patients with chronic renal failure remains a matter of debate. According to some authors regression must be an extremely slow process, if it occurs at all (42, 126). This is in sharp contrast to the rapid reversibility of excessive parathyroid function in uremic rats after normalization of renal function by kidney transplantion (134), although parathyroid mass probably remained markedly increased in this experimental model.

The issue of regression is of clinical importance. If for example a chronic dialysis patient with a dramatic increase in total parathyroid mass has practically no chance to experience regression of hyperplastic glands after uremia correction by a succesful kidney graft it would seem appropriate to perform a surgical parathyroidectomy prior to renal transplantation. If however significant regression of hyperplasia occurs as an active or passive process, namely by enhanced apoptosis or reduced proliferation, prophylactic surgery should be avoided. That regression of parathyroid hyperplasia secondary to vitamin D deficiency can occur has been convincingly demonstrated many years ago in experiments done in chicks (135). Thus the administration of cholecaciferol to birds that had developed an increase in parathyroid gland mass when fed a rachitogenic, vitamin D-free diet for 8-10 weeks led to a significant (50%) reduction in gland weight. Calcitriol failed to achieve same effect at low, albeit hypercalcemic, dose but was capable of reducing gland mass at higher dose. However, in an experimental dog model no regression was found. Thus in dogs first submitted to a low-calcium, low-sodium and vitamin D deficient diet for two years and subsequently to a normal diet for another 17 months, no involution of hyperplastic parathyroid glands was observed (136). In uremic animals, published evidence for or against regression of increased parathyroid mass is sparse. The calcimimetic drug, NPS R-568 which has been shown to decrease parathyroid cell proliferation and to prevent parathyroid hyperplasia in 5/6th nephrectomized rats, was unable to reverse it (128, 137).

In patients, a rapid remission of parathyroid overfunction may occur either due to parathyroid “ apoplexy ”, that is necrosis, as has been shown in rare cases of 1° hyperparathyroidism (138), or due to enhanced apoptosis. The diagnosis of necrosis is more difficult in 2° than in 1° forms of hyperparathyroidism because the hyperplasia of the former is not limited to one gland.

Regression of parathyroid hyperplasia in hemodialysis patients in response to calcitriol pulse therapy for 12 weeks has been reported by Fukagawa et al using ultrasonography (139). These authors observed a significant decrease in mean gland volume from 0.87 to 0.51 cm3 of this time period, concomitantly with a reduction in serum iPTH of more than 50%. In contrast, Quarles et al who also examined parathyroid gland anatomy in hemodialysis patients in vivo in response to intermittent i.v. or oral calcitriol treatment for 36 weeks failed to observe a decrease in parathyroid gland size as assessed by high resolution ultrasound and/or magnetic resonance imaging (140). Mean gland size was 1.9 and 2.1 cm3 before and 3.3 and 2.3 cm3 after oral and i.v. calcitriol therapy, respectively. The authors achieved an overall maximum average serum PTH reduction of 43% over this time period. There were marked differences between these two studies. Hyperparathyroidism probably was more severe in the latter than in the former. Although initial mean serum iPTH levels were similar, serum phosphorus was higher and the decrease in serum PTH achieved in response to calcitriol was less marked in the latter. Moreover, parathyroid mass was more than double.

In another study, Fukagawa et al examined the possible relationship between parathyroid size and the long-term outcome after calcitriol pulse therapy, by subdividing patients into different groups according to estimation of initial parathyroid gland volume (141). In two hemodialysis patients with detectable gland(s), in whom the size of all parathyroid glands as well as PTH hypersecretion regressed to normal, 2° hyperparathyroidism remained controllable for at least 12 months after switching to conventional oral active vitamin D therapy. In contrast, in seven hemodialysis patients, in whom the size of all parathyroid glands did not regress to normal by calcitriol pulse therapy, 2° hyperparathyroidism relapsed after switching to conventional therapy although PTH hypersecretion could be controlled temporarily. Similarly, in a recent study Okuno et al. showed that in hemodialysis patients plasma PTH levels and the number of detectable parathyroid glands decreased in response to maxacalcitol (22-oxacalcitriol) given for 24 weeks only when the maximum diameter of one of the parathyroid glands was less than 11.0 mm, but not when it was above that value (142).

Taken together, these findings suggest that the degree of parathyroid hyperplasia, as detected by ultrasonography, is an important determinant for regression in response to calcitriol therapy. It is probable, although not proven, however, that the type of hyperplasia, namely monoclonal, multiclonal or polyclonal growth, is even more important with respect to regression potential than the actual size of each gland.

Figure 3 (see above) summarizes in a schematic view the main mechanisms involved in the abnormal synthesis and secretion of PTH and in the hyperplasia of parathyroid tissue. It further points to the possible counterregulatory role of apoptosis.

Altered PTH metabolism and resistance to PTH action

PTH metabolism is greatly disturbed in CRF. Normally, most of full-length PTH1-84 is transformed in the liver into the biologically active N-terminal PTH1-34 fragment and several other, inactive C-terminal fragments. The latter are mainly catabolized in the kidney and the degradation process involves solely glomerular filtration and tubular reabsorption, whereas the N-terminal PTH1-34 fragment undergoes both tubular reabsorption and peritubular uptake, as does the full-length PTH1-84 molecule (143). Tubular reabsorption involves the multifunctional receptor megalin (144).

In CRF, both pathways of renal PTH degradation are progressively impaired. This leads to a marked prolongation of the half-life of C-terminal PTH fragments in the circulation (145-147) and their accumulation in the extracellular space. Moreover, there is no peritubular metabolism of PTH1-84 in uremic non-filtering kidneys, in contrast to peritubular uptake by normal, filtering kidneys (148). Hepatic PTH catabolism appears however to be unchanged in CRF. Thus uremic livers released equal amounts of immunoreactive C-terminal PTH fragments as control livers (148).

A decreased response to the action of PTH may be another factor involved in the stimulation of the parathyroid glands in CRF. A diminished calcemic response to the infusion of PTH has long been reported, suggesting that PTH oversecretion was necessary to maintain eucalcemia. The skeletal resistance to PTH is probably due to several different factors, including impaired vitamin D action in association with hyperphosphatemia, overestimation of true PTH(1-84) by assays measuring iPTH (see below), accumulation of inhibitory PTH fragments, increase in circulating osteoprotegerin levels, and changes in PTH-R1 expression (149, 150). Concerning the latter mechanism, studies suggest the presence of PTH receptor isoforms in various organs of normal rats. Moreover, downregulation of PTH-R1 mRNA was observed in various tissues in uremic rats (151-153) and also in osteoblasts of bones from patients with end-stage renal disease (154). However, the issue of PTH-R1 expression in bone tissue remains a matter of controversy since another group actually found it to be upregulated in patients with moderate to severe renal hyperparathyroid bone disease (155).

Clinical features

In most patients with ESRD, even advanced 2° hyperparathyroidism remains a clinically silent disease. Clinical manifestations are generally related to severe osteitis fibrosa and to the consequences of hypercalcemia and/or hyperphosphatemia.

Osteo-articular pain may be present. When patients become symptomatic, they usually complain of pain on exertion in skeletal sites that are subjected to biomechanical stress. Pain at rest and localized pain are rather unusual and suggest other underlying causes. Severe proximal myopathy is seen in some patients, even in the absence of vitamin D deficiency. These symptoms and signs are more frequent in patients who suffer from mixed renal osteodystrophy, resulting from a combination of parathyroid overfunction and vitamin D deficiency. Skeletal fractures may occur after only minor injury. They may also develop on the ground of cystic bone lesions, the so-called “ brown tumors ”, which occur for still unknown reasons in a small number of uremic patients with 2° hyperparathyroidism. Rupture of the patella or avulsion of tendons may be seen in advanced cases.

Uremic pruritus is most often associated with an elevated Ca x P product although other factors may also be involved. Related symptoms and signs are the red eye syndrome due to the deposition of calcium in the conjunctiva, cutaneous calcification, and pseudogout. The latter is a form of painful arthralgia of acute or subacute onset caused by intra-articular deposition of radio-opaque crystals of calcium pyrophosphate dehydrate.

The syndrome of “ calciphylaxis ” is an infrequent manifestation of cutaneous and vascular calcification in uremic patients which may occur in association with 2° hyperparathyroidism, although this association is by no means constant. It is characterized by a rapidly progressive skin necrosis involving buttocks and the legs, particularly the thighs. It can produce gangrene and may be fatal. It occurs as the result of arteriolar calcification and has also been termed “ calcific uremic arteriolopathy ” to reflect more accurately the nature of the lesion (156).

Diagnosis

The biochemical diagnosis of 2° hyperparathyroidism relies since approximately 20 years on the determination of plasma iPTH. This is true for primary and secondary forms of hyperparathyroidism. In patients with CRF, it has however become apparent in recent years that there are several limitations to the measurement of iPTH, in addition to the usual day-to-day variations in healthy people (157). Normal iPTH plasma values are not normal for uremic patients since values in the normal range are often associated with low bone turnover (adynamic bone disease) whereas normal bone turnover may be observed in presence of elevated plasma intact PTH levels (158-161). It is currently unclear to what extent this is due to imperfections in the PTH assays used (see below), PTH receptor state, post-receptor events, non-PTH-mediated changes in bone metabolism (e.g. supply of vitamin D or its metabolites, supply of estrogens or androgens), or a combination of these factors.

The accumulation of a large non (1-84) molecular form of PTH, which is detected by iPTH (so-called "intact" PTH) assays, has been described in patients with CRF (162). The large PTH fragment was tentatively identified as hPTH(7-84) (163). This finding is of importance in the interpretation of PTH values, since true hPTH(1-84) represents only about 50-60% of the levels detected by the currently used intact PTH assays, and since PTH(7-84) antagonizes PTH(1-84) effects on serum calcium and on osteoblasts (164). Moreover, the secretory responses of hPTH(1-84) and non-hPTH(1-84) to changes in extracellular calcium concentration are not proportional for these two PTH moieties (54). Moreover, a large variability has been found between different assay methods used for plasma PTH measurement in patients with CKD, recognizing PTH7-84 with various cross-reactivities (165). Varying plasma sampling and storage conditions may further complicate the interpretation of PTH results provided by clinical laboratories (166). The development of assays which detect full-length (whole) human PTH, but not amino-terminally truncated fragments (167), was initially considered as a major progress in this field.. Monier-Faugere et al proposed to further improve the assessment of uremic hyperparathyroidism and the associated increase in bone turnover by calculating the ratio of PTH-(1-84) to large C-PTH fragments (168). The usefulness in the clinical setting of the whole PTH assay and of the ratio of whole PTH to PTH fragments has however not been convincingly established at present for the diagnosis of parathyroid overfunction in adult (169, 170) or pediatric (171) dialysis patients. From a practical point of view, it must be pointed out that at present measurement of PTH with third-generation assays is not widely available. Moreover, from a theoretical standpoint, we are hoping that in the future we will be able to rely not only on serum PTH but also on circulating and other markers of bone structure and function for the assessment of renal osteodystrophy and on markers of cardiovascular disease related to secondary hyperparathyroidism (172).

The radiological diagnosis is relatively easy in advanced stages of 2° hyperparathyroidism. Typical lesions include resorptive defects on the external and internal surfaces of cortical bone, particularly resorption on the subperiosteal surface. Resorption within cortical bone enlarges the Haversian channels, resulting in longitudinal striation; resorption at the endosteal surface causes cortical thinning. These lesions can be generally detected first in the hand skeleton, most characteristically at the periosteal surface of the middle phalanges (Figure 14). Accelerated bone deposition at this site (periosteal neostosis) can also be seen. Another characteristic feature is resorptive loss of acral bone (acro-osteolysis), in particular at the terminal phalanges, at the distal end of the clavicles, and in the skull (‘pepper-pot’ aspect) (Figure 15). Whereas cortical bone is progressively thinning, the mass of spongy bone tends to increase, particularly in the metaphyses. The latter phenomenon results in a characteristic sclerotic aspect of the upper and lower thirds of the vertebrae, contrasting with rarefaction of the center (‘rugger jersey spine’). Osteosclerosis is also commonly seen in radiographs of the metaphyses of the radius and tibia.

Figure 14. X-ray aspect of periosteal resorption within cortical bone of middle phalanges of the hand, indicative of osteitis fibrosa, and extensive finger artery calcification in a CKD stage 5 patient with severe secondary hyperparathyroidism.

Figure 15. Pepper-and-salt aspect of the skull in in a chronic hemodialysis patient with severe secondary hyperparathyroidism.

In addition to the skeletal lesions, radiographs often reveal various types of soft tissue calcification. These comprise vascular calcifications, i.e. calcification of intimal plaques (aorta, iliac arteries) (Figure 16a), as well as diffuse calcification (Mönckeberg type) of the media of peripheral muscular arteries (Figure 16b) (173). Of interest, media calcification of digital arteries can entirely regress after surgical parathyroidectomy (Figure 14). Calcium deposits may also be seen in periarticular tissue or bursas and may exhibit tumor-like features (Figure 17). Since the development of electron-beam computed tomography (EBCT) and multiple slice computed tomography (MSCT) more reliable means have become available to assess quantitatively vascular calcification and its progression in uremic patients (174). However, these techniques are not universally available and costly. Moreover, they do not allow a distinction between arterial intima and media calcifications. Such a distinction can be obtained by radiograms of the pelvis and the thigh, combined with ultrasonography of the common carotid artery. Using these simple methods, London et al could show that hemodialysis patients with arterial media calcification had a longer survival than hemodialysis patients with arterial initma calcification, but in turn their survival was significantly shorter than that of hemodialysis patients without calcifications (175). Both severe hyperparathyroidism and marked hypoparathyroidism favor the occurrence of the two types of calcification in ESRD patients (176-178). In contrast to permanent elevations in serum PTH, the intermittent administration of PTH1-34 has been shown to decrease arterial calcification in uremic rats (179) and in diabetic mice with LDL receptor deletion (180). This observation tends to demonstrate that normal parathyroid function is required not only for the maintenance of optimal bone structure and function, but also as an efficacious defense against soft tissue calcification, and that intermittent PTH administration may not only improve osteoporosis (181), but also reduce vascular calcification, at least in experimental animals.

Figure 16. Massive intima (a) and media (b) calcification of hypogastric artery from a chronic hemodialysis patient.

Figure 17. Tumorlike periarticular calcification in the shoulder of a chronic hemodialysis patient with adynamic bone disease due to aluminum intoxication.

Treatment

Medical management. Presently available options of medical treatment should take into account the levels of plasma biochemistry and x-ray findings, and as a more recently recognized parameter also the dimensions of the largest parathyroid glands, as assessed by ultrasonography. A gland diameter of 5-10 mm or more is considered by some groups as being indicative of autonomous growth which often is resistant to medical treatment (141).

Schematically, there are five major medical treatment options which can be combined in some cases, but not in others, namely the restriction of phosphate intake, the administration of calcium supplements, the administration of phosphate binders, and the prescription of vitamin D derivatives and calcimimetics (182, 183). In dialysis patients the weekly dose of renal replacement therapy is an additional important factor. An optimal dialysis technique allows the control of hyperphosphatemia, the adaptation of the dialysate calcium concentration to each patient’s needs to reduce the occurrence of hypercalcemia, the avoidance of metabolic acidosis and the removal of uremic toxins.

When trying to control hyperparathyroidism it is important to avoid hypocalcemia and hypercalcemia and to reduce or correct hyperphosphatemia as well. In patients with already controlled plasma phosphate, this can be achieved by giving either calcitriol or one of its synthetic analogs, or by administering oral calcium supplements. Until recently, most clinicians would have agreed that calcitriol or alfacalcidol should be the preferred therapy in patients with high to very high plasma intact PTH values and normal to moderatety elevated plasma calcium levels, if plasma phosphorus did not exceed recommended levels, namely 1.5 mmol/l for CKD stage 3-4 and 1.8 mmol/l for CKD stage 5, according to the K/DOQI guidelines of 2003 (184). However, the administration of active vitamin D derivatives often induces hypercalcemia and/or hyperphosphatemia. The recent KDIGO CKD-MBD guideline (185) suggests "maintaining iPTH levels in CKD stage 5D patients (i.e. patients on dialysis) in the range of approximately two to nine times the upper normal limit for the assay, to keep serum calcium normal, and to decrease serum phosphorus towards the normal range." Thus the recommended iPTH target range has become larger than with the K/DOQI guidelines. The KDIGO guideline suggests further that marked changes in iPTH levels in either direction within the newly defined, broadened range should "prompt initiation or change in therapy to avoid progression to levels outside of this range." It pursues : "In patients with CKD stages 3-5 not on dialysis, the optimal PTH level is not known. However, it is suggested that patients with levels of iPTH above the upper normal limit of the assay are first evaluated for hyperphosphatemia, hypocalcemia, and vitamin D deficiency. It is reasonable to correct these abnormalities with any or all of the following: reducing dietary phosphate intake and administering phosphate binders, calcium supplements, and/or native vitamin D."

Vitamin D and active vitamin D derivatives. A satisfactory degree of vitamin D repletion should probably be aimed at in case of vitamin D deficiency since most of them have at least some degree of vitamin D deficiency (26, 186). Relative vitamin D depletion has been shown to be an independent risk factor for 2° hyperparathyroidism in hemodialysis patients (29). Repletion with native vitamin D may lead to improved control of secondary hyperparathyroidism in patients with CKD not yet on dialysis (187) and in those treated by dialysis (188). In addition, it might allow optimal bone formation, help to avoid osteomalacia, and exert numerous other positive effects due to its pleiotropic actions. However, randomized controlled trials with native vitamin D have not yet been performed in CKD patients for the evaluation of hard patient outcomes. To correct secondary hyperparathyroidism of moderate to severe degree the oral administration of active vitamin D derivatives is generally more efficient. In hemodialysis patients, calcitriol or its analogs can be given either orally or intravenously. The oral administration can be on a daily basis (for instance 0.125 to 0.5 µg) or as intermittent bolus ingestions (for instance 0.5 to 2.0 µg or more for each dose) whereas the i.v. administration is always intermittent (also 0.5 to 2.0 µg or more per injection). The route and mode of administration of calcitriol or alfacalcidol probably play only a minor role. Since the highly active 1-hydroxylated vitamin D derivatives can easily induce hypercalcemia, intensive resarch has focused on the development of various non-hypercalcemic analogs, including the natural vitamin D compound 24,25(OH)2 vitamin D3, 22-oxa-calcitriol (maxacalcitol), 19-nor-1,25(OH)2 vitamin D3 (paricalcitol), and 1-(OH) vitamin D2 (hectorol). Despite numerous studies done in many patients, none of them has been shown to have entirely lost the capacity of inducing an increase in plasma calcium or phosphate, and none has been demonstrated thus far to be superior to calcitriol or alfacalcidol in the long run in controlling secondary hyperparathyroidism (189). A report by Teng et al. showed that paricalcitol administration to a large cohort of hemodialysis patients conferred a remarkable (16%) survival advantage over the administration of calcitriol (190). Numerous subsequent observational studies reported a survival benefit, either comparing treatment with active vitamin D derivatives to no treatment, or novel active vitamin D derivatives to calcitriol in either CKD patients not yet on dialysis (191) and in those receiving dialysis treatment (192-194). One recent observational study in hemodialysis patients, however, did not find a survival advantage with paricalcitol, as compared to calcitriol (195). In the absence of randomized controlled trials it appears to be premature to conclude that paricalcitol treatment is superior to calcitriol or alfacalcidol in terms of patient survival. Findings of observational studies can only be considered as hypothesis-generating and needs to be confirmed by a properly designed prospective investigation (196).

Calcimimetics. The introduction of the calcimimetic cinacalcet into clinical practice led to a change in the above treatment strategy since this enables parathyroid overfunction control without increasing plasma calcium or phosphorus. Calcimimetics modify the configuration of the CaR cloned by Brown et al (197) and thereby make it more sensitive to extracellular calcium whereas the so-called calcilytics decrease its sensitivity (Figure 18). Initial acute studies in chronic hemodialysis patients showed that the calcimimetic cinacalcet was capable of reducing plasma PTH within hours, immediately followed by a rapid decrease in plasma calcium and a minor decrease in plasma phosphorus (198-200). Moreover, in short-term and long-term studies performed in rats or mice with chronic renal failure the administration of the calcimimetic NPS R-568, starting at the time of uremia induction, allowed the prevention of parathyroid hyperplasia (128, 137, 201)..Perhaps more important from a clinical point of view, the administration of calcimimetics enabled an improvement of osteitis fibrosa (66), halted the progression of vascular calcification (201, 202), prevented vascular remodeling (203), improved cardiac structure and function (204), and prolonged survival (205) in uremic animals with 2° hyperparathyroidism.

Figure 18. Schematic representation of the modulation of the calcium-sensing receptor by calcimimetics and calcilytics. The former increase receptor sensitivity to calcium ions whereas the latter decreases it.

The long-term administration of cinacalcet to chronic hemodialysis patients proved to be superior to « optimal » standard therapy in controlling 2° uremic hyperparathyroidism, in that it was able to induce not only a decrease in plasma PTH but also in plasma calcium and phosphorus (206-208). The initial daily dose is 30 mg orally, which can be increased up to 180 mg if necessary. Cinacalcet is generally well tolerated, with the exception of gastrointestinal side effects, which however cease in the majority of patients with time. Since its administration generally leads to a decrease in serum calcium, a close follow-up is required, at least initially, to avoid hypocalcemia with possible adverse clinical consequences. Cinacalcet can be associated with calcium-containing and non-calcium containing phosphate binders and also with vitamin D derivatives. For PTH lowering, recent studies suggest that combination therapy may lead to a more complete correction than single drug treatment because of less side-effects and greater efficacy in the control of parathyroid overfunction (209, 210). Future randomized controlled trial must answer whether calcimimetics impact on cardiovascular events or survival and whether in this respect there are differences between vitamin D sterols and calcimimetics (182). The ongoing EVOLVE trial should provide an answer to this question in 2011 (211).

Phosphate binders, oral phosphate restriction, and phosphate removal by dialysis. Calcium-containing phosphate binders should be given, preferentially during or at the end of phosphate-rich meals, to patients with CKD and uncontrolled hyperphosphatemia who have no hypercalcemia or radiological evidence of soft tissue calcifications. In these latter cases non calcium-containing phosphate binders should be administered preferentially (see below). The administration of calcium salts alone such as calcium carbonate or calcium acetate may be sufficient for the control of hyperphosphatemia in many instances, particularly in patients with CKD stages 3-5 not yet on dialysis. It will at the same time prevent serum iPTH from rising in the majority of patients (212). It may however lead to PTH oversuppression and adynamic bone disease (213). In hemodialysis patients, the efficacy and tolerance of this treatment may be enhanced by the concomitant use of low-calcium dialysate, for instance 1.25 mmol/l, especially if plasma intact PTH levels are not very high. However, long-term studies have shown that the continuous use of a dialysate calcium of only 1.25 mmol/l requires close monitoring of plasma calcium and PTH because of the risk of inducing excessive PTH secretion (214). Moreover, the use of a low calcium dialysate may require higher doses of active vitamin D derivatives (215) or cinacalcet (216) for the control of 2° hyperparathyroidism. Finally, a low calcium bath favors hemodynamic instability during the hemodialyis session (217). In CAPD patients, the use of calcium carbonate, in the absence of vitamin D, together with a reduction of the dialysate calcium concentration from 1.75 to 1.45 mmol/l prevents the occurrence of hypercalcemia in most patients (218). However, the addition of daily low-dose alfacalcidol may lead to hypercalcemia, despite a further reduction of dialysate calcium to 1.0 mmol/l.

The development of calcium-free, aluminum-free oral phosphate binders such as sevelamer-HCl (219-221), sevelamer carbonate (222, 223), and lanthanum carbonate (224-226) allows control of hyperphosphatemia without the potential danger of calcium overload. Their phosphate binding capacity is roughly equivalent to that of Ca carbonate or calcium acetate. Sevelamer offers in addition the advantage to lower serum total cholesterol and LDL-cholesterol and to increase serum HDL-cholesterol, to slow the progression of arterial calcification in dialysis patients (220), and possibly to improve survival in such patients (227). The administration of sevelamer is probably more efficient in halting of the progression of vascular calcification than calcium carbonate or calcium acetate but this remains a matter of debate (228). No human outcome studies are as yet available for lanthanum carbonate. The effects of calcium-free, aluminum-free phosphate binders on serum iPTH are variable, depending on baseline iPTH and concomitant therapies. In general, iPTH levels are higher than with calcium-containing phosphate binders (229, 230).

The administration of aluminum-containing phosphate binders should be avoided because of their potential toxicity. They may be given in some treatment resistant cases, but only for short periods of time (185).

Dietary phosphate intake should be examined closely and diminished, if possible. The spontaneous reduction of protein intake with age probably explains the often better control of serum phosphate in elderly ESRD patients, compared with younger ESRD patients, and this might contribute to the relatively low PTH levels of the former and their propensity to develop adynamic bone disease (231). However, when reducing dietary phosphate intake and concomitantly protein intake, one has to take care to avoid the induction of a protein malnutrition state. The risk of controlling serum phosphorus by restricting dietary protein intake may outweigh the benefit of controlled phosphorus and may lead to greater mortality (232). In dialysis patients, an attempt should always be made as well to improve the efficiency of the dialysis procedure.

A better correction of metabolic acidosis by bicarbonate-buffered dialysate, as compared to acetate-buffered dialysate, probably helps to delay the progression of osteitis fibrosa in hemodialysis patients (233). A recently proposed mechanism for the beneficial role of acidosis correction is an increase in the sensitivity of the parathyroid gland to plasma ionized calcium (234).

Current recommendations for the medical treatment and prevention of patients with CKD-MBD, including 2° hyperparathyroidism, can be found in the 2009 KDIGO CKD-MBD guideline (185).

Local injection of alcohol and active vitamin D derivatives. Since in advanced forms of 2° hyperparathyroidism the hyperplasia of parathyroid glands is asymmetrical, with some glands being grossly enlarged and others remaining relatively small, local ethanol injection has been proposed as an alternative therapy in patients who become resistant to medical treatment (235, 236). The procedure is performed by fine needle injection of small amounts of ethanol under Doppler-ultrasonography guidance, targeted at the largest gland(s). In many instances a second and a third injection is required to decrease plasma PTH levels adequately. In the experience by Kakuta et al, parathyroid function could be maintained within target range in 38 of 46 patients (80.4%) at 1 year after percutaneous ethanol injection, followed by appropriate medical therapy (237). Surgical parathyroidectomy was not required in any patient. Conversely, in the eight remaining patients with recurrent hyperparathyroidism who required subtotal parathyroidectomy, plasma iPTH levels recovered in only 50% of them at 1 year after ethanol injection. Thus this technique can allow one to set a new stage for the successful treatment with active vitamin D derivatives, at least in highly specialized institutions. However, it has not reached at present a widespread use in clinical practice outside of Japan since other research groups were unable to obtain equally convincing results (238, 239).

Recently, the direct injection of maxacalcitol into parathyroid glands of dialysis patients (240) and uremic rats (241) was shown to lead to a rapid, significant decrease in circulating PTH. In the rats, the injection of vehicle solution alone did not induce a change in plasma PTH levels. The ultimate contribution of the administration of active vitamin D derivatives via this invasive route to the control of severe 2° uremic hyperparathyroidism remains to be seen.

Despite major advances in the medical treatment of mineral and bone metabolism disturbances in uremic patients the achievement of the targets for plasma calcium, phosphorus, Ca x P product, and PTH, as recommended by the K/DOQI guidelines (184) was far from being optimal in the DOPPS patient population for the years 2002-2004 (242). It was actually rare in the hemodialysis patients of this international cohort to fall within recommended ranges for all four indicators of mineral metabolism. However, consistent control of all three main CKD-MBD parameters, namely calcium, phosphorus, and PTH was found to be a strong predictor of survival in hemodialysis patients (243).

Surgical treatment. The surgical correction remains the final, symptomatic therapy of the most severe forms of 2° hyperparathyroidism which cannot be controlled by medical treatment. The most important issue is to prevent or correct the development of major clinical complications associated with this disease. The presence of a severe form of overfunction must be ascertained by clinical, biochemical and radiological evidence. Thus in general neck surgery should only be done in case plasma iPTH values are greatly elevated (> 600-800 pg/ml), together with an increase in plasma total alkaline phosphatases (or better bone-specific alkaline phosphatase), and only after one or several medical treatment attempts have been unsuccessful in decreasing plasma iPTH with cinacalcet and/or active vitamin D derivatives or if their use is absolutely contraindicated, namely in presence of persistent hypercalcemia or hyperphosphatemia. Bone histomorphometry examination is rarely needed. Clinical symptoms and signs such as pruritus and osteoarticular pain are non specific and therefore are not good criteria for operation by their own. Similarly, an isolated increase in plasma calcium and/or phosphorus, even in case of coexistent soft tissue calcifications, is not a sufficient criterion alone for surgical PTX. However, in the presence of a high plasma PTH the latter disturbances may contribute to favor the decision to proceed to the surgical correction of parathyroid overfunction. The results can be spectacular, including the complete disappearance of soft tissue calcifications (see Figure 14b). A concomitant aluminum overload should be excluded or treated, if present, before performing surgery.

Two main surgical procedures are generally used, namely either subtotal parathyroidectomy or total parathyroidectomy with immediate autotransplantation. There is no substantial difference of operative difficulties and treatment results between the two procedures. We found that the long term frequency of recurrence of hyperparathyroidism was similar (244). Some authors claim superiority of total parathyroidectomy without reimplantation of parathyroid tissue in terms of long-term control of parathyroid overfunction, tolerance, and safety (245), but this has been questioned by us and others (246-248). We do not recommend the performance of total parathyroidectomy without autotransplantation in uremic patients since subsequently permanent hypoparathyroidism and adynamic bone disease may develop, with possible harmful consequences especially for those patients who subsequently undergo kidney transplantation.

The frequency of parathyroidectomy has not changed significantly during the last decade according to a survey done in Northern Italy some years ago (249). In the USA, PTX was associated with higher short-term mortality, but lower long-term mortality among chronic dialysis patients (250). Measures to attenuate 2° hyperparathyroidism may play an important role in reducing mortality among patients with end-stage renal disease.

ACKNOWLEDGEMENTS

The author wishes to thank Ms Martine Netter for expert assistance in Figure design.