1.1 The vitamin D story

In 1920, Aldolf Windaus, a German physician and chemist, and his colleagues isolated a material from plant sterols that following irradiation with ultraviolet light was active in healing rickets - a disorder characterized by defect mineralization of the skeleton. The substance was termed "vitamin D". Dr. Windaus and his colleagues further isolated and identified other nutritional forms of vitamin D enabling them to cure diseases caused by lack of vitamin D. Thereby rickets was eliminated as a major medical problem. For this contribution Alfred Windaus received the Nobel Prize in chemistry in 1928 [8].

Besides the effects of vitamin D on bone mineralization, it was discovered in the 1920s and 1930s that an "endogenous factor" played an unequivocal role in calcium homeostasis by changing the intestinal absorption of calcium according to the skeletal need [9]. This "endogenous factor" was later shown to be the vitamin D analog, 1,25(OH) 2 D 3 . Inadequate plasma levels of calcium and phosphate were shown in the 1950s to be of importance for the defective mineralization of the skeleton in vitamin D deficiency. At that time it was discovered that another major function of vitamin D was to mobilize calcium from bone [10]. How vitamin D exerts this mobilization of calcium is still not known in details, but it became clear in the 1970s that both parathyroid hormone (PTH) and vitamin D, independently, were necessary for this process [11]. In the 1980s, the influence of vitamin D and PTH on renal calcium absorption was discovered [12] leading to vitamin D being accepted as a hormone and not just considered as a vitamin.

Already in 1960, Kodieck et al. observed that the inactive vitamin D compound was converted to an active form after digestion [13, 14]. In 1968, 25(OH)D 3 was purified and chemically identified as the first active metabolite of vitamin D [15]. Radio labelled 25(OH)D 3 was shown to be metabolized to more polar metabolites [16, 17] and in 1971 the structure of the active vitamin D metabolite was unequivocally demonstrated to be 1,25(OH) 2 D 3 [18]. The ultimate proof of 1,25(OH) 2 D 3 being the active metabolite was obtained in 1973 when Vitamin D-dependent rickets type 1 (an autosomal recessive disorder with a defect in the 1 α -hydroxylase enzyme) was successfully treated by what later was found to be physiological doses of synthetic 1,25(OH) 2 D 3 , whereas much larger amounts of vitamin D 3 or 25(OH)D 3 were needed to heal the disorder [19]. A few years later, two chemical different structures 1 α ,25(OH) 2 D 3 [20] and 1 β ,25(OH)2D3 [21] were synthesized, with 1 α ,25(OH) 2 D 3 being the active metabolite.

The stringent, physiological control of plasma 1,25(OH) 2 D 3 levels is exerted through the coordinated action of classic mineral regulating organs: the kidney, intestine, bone and parathyroid glands. In chronic renal failure, formation of 1,25(OH) 2 D 3 is impaired [22, 23].

1.2 Intracellular actions of 1,25(OH) 2 D 3

In the beginning of the 1960s, Zull et al. showed that cellular nuclear activity was required for vitamin D to carry out its function [24]. The first clear demonstration of the existence of a nuclear vitamin D receptor took place in 1973 [25] and in 1988 the Vitamin D receptor was purified and the full coding sequence for the Vitamin D receptor (VDR) described [26]. VDRs are not only present in the classical target tissues (intestine, kidney, bone) regulating calcium homeostasis, but also in a wide variety of non-classical tissues including keratinocytes, malignant cells and cells belonging to the immune system [27]. Numerous genes induced or suppressed by the 1,25(OH) 2 D 3 / VDR complex are relevant for the efficacy of 1,25(OH) 2 D 3 therapy. The biological actions of 1,25(OH) 2 D 3 on genes include the classical calcium homeostasis in bone, intestine and kidney; the regulation of the rates of synthesis and catabolism of 1,25(OH) 2 D 3 , the suppression of PTH synthesis, modulation of immune responses, and suppression of cell proliferation [27]. The simultaneous presence of 25(OH)D 3 -1 α -hydroxylase and VDRs in several tissues suggests a paracrine role for 1,25(OH) 2 D 3 locally modulating cell proliferation and differentiation [28]. The current model of 1,25(OH) 2 D 3 -VDR action is that 1,25(OH) 2 D 3 after entering the cell can either be inactivated by mitochondrial 24-hydroxylase to 24,25(OH) 2 D 3 or bind to VDR. Ligand binding activates the VDR to translocate from the cytosol to the nucleus where it heterodimerizes with its partner - the retinoid X receptor RXR. The VDR/RXR complex binds to specific sequences in the promoter region, the vitamin D response element (VDRE), of the target genes - and recruits basal transcription factors and co-regulator molecules to either increase or suppress the rate of gene transcription by RNA-polymerase II [27].

Results of several animal studies have suggested that 1,25(OH) 2 D 3 also can exert direct action on its target cells via rapid effects on the cell membrane [29, 30]. In 1998, the first paper was published identifying a membrane receptor for 1,25(OH) 2 D 3 mediating a rapid activation of protein kinase C in rat chondrocytes. Until now, a similar membrane receptor has not been demonstrated in human tissues.

In this thesis first a short introduction to the calcium and phosphate homeostasis in the normal man will be given. Then an overview of the causes, development and consequences of secondary hyperparathyroidism in chronic uremic patients is presented. Finally, follows a short overview of conventional treatment modalities of secondary hyperparathyroidism in dialysis patients at the time when the first of the presented studies was initiated. The clinical data presented are from 3 different experimental scenarios:

1. Short and long-term experiences with intermittent intravenous 1 α (OH)D 3 treatment to patients on hemodialysis, and a short-term cross-over study between intermittent oral and intermittent intravenous administration.

2. Experience with intermittent oral and intravenous 1 α (OH)D 3 treatment, combined with reduced calcium concentration in the dialysis fluid to patients on CAPD (Continuous Ambulatory Peritoneal Dialysis) and on hemodialysis and at the same time a simultaneous change from aluminum- to calciumcontaining phosphatebinders.

3. Studies on the pharmacokinetic properties of 1 α (OH)D 3 as compared to active vitamin D, 1,25(OH) 2 D 3 , in normal subjects and uremic patients.

The results obtained will be discussed from diagnostic and treatment perspectives with respect to new knowledge obtained from studies during the recent years. At the end, some aspects of treatment in the future will be discussed.

2.1 Calcium homeostasis in normal man

Regulation of the calcium homeostasis involves several organs: Intestine, kidneys, skeleton, and the parathyroids, and different hormones: Parathyroid hormone, vitamin D and calcitonin [31]. The concentration of ionized calcium (Ca 2+ ) in the extracellular fluid (ECF) is 104 times higher than the concentration in the intracellular fluid (ICF) [32, 33]. Most of the calcium in the body (99%) is stored in the skeleton, which serves as a relatively inexhaustible reservoir. Approximately 47% of the extracellular calcium is ionized, 46% is proteinbound and the remainder is complexed to small ions. Appoximately 75% of the proteinbound fraction is bound to albumin [32, 33]. Thus less than 1% of the calcium present in ECF is present as Ca 2+ [32, 33].

Non-ionized calcium, predominantly found in skeletal bone, provides an important structural function to the body, whereas Ca 2+ is responsible for a variety of physiological and cellular effects that are characteristic of that particular cell type (e.g. secretion, neuromuscular impulse formation, contractile functions etc.).

Normal adult man ingests about 20 mmol of calcium per day of which approximately 40% (i.e. 8 mmol) is absorbed in the duodenum and upper jejunum, although absorption varies from 10-90% (2-18 mmol). About 10% of intestinal calcium absorption occurs passively via independent paracellular, non-saturable pathways while the major part is absorped via specialized vitamin D dependent saturable, transcellular pathways [31, 32]. Skeletal bone liberates and reabsorbs approximately 500 mmol of calcium per day from an exchangeable pool of 100 mmol [32]. About 250 mmol of ionized calcium is filtered per day in the kidneys by glomerular filtration. About 65% (i.e. 165 mmol) is reabsorbed passively together with sodium in the proximal tubule, 20-25% (i.e. 60 mmol) is reabsorbed in the thick ascending loop of Henle, and finally 10% (i.e. 25 mmol) is absorbed in the distal convolute tubuli [32]. Parathyroid hormone (PTH) increases calcium absorption in the thick ascending loop of Henle as well as in the distal convoluted tubule, but has no effect on calcium reabsorption in the proximal tubule. The urinary excretion of calcium is 2.5-7.5 mmol per day and represents about 3-5% of the filtered calcium. Approximately 60% (i.e. 12 mmol) of the oral daily intake is excreted with the feces [32].

In healthy man, plasma Ca 2+ does not vary by more than 5% and is maintained constant mainly by the actions of PTH and vitamin D [31, 32, 34]. Calcitonin is largely secreted in hypercalcemic conditions [32, 33].

2.2 Phosphate homeostasis in normal man

An intimate relationship exists between the homeostasis of phosphate and calcium.

Phosphate is needed for mineralisation of bone, for cellular structural components (e.g. phospholipids, nucleotides, phosphoproteins), for energy storage in ATP, for oxygen transport in red blood cell in 2,3-DPG, and in the acid base balance of the organism as cellular and urinary buffers [32]. Phosphates in blood exist as organic (ester and lipid phosphates) and inorganic compounds. Plasma phosphate denotes the inorganic component. Total plasma phosphate in adults ranges from 0.80 to 1.35 mmol/L and is distributed between 15% protein-bound, 45% ionized and 40% in complexed forms with calcium and magnesium. The intracellular concentration of phosphate is approximately 100 mmol/l with 5 mmol/l as inorganic phosphate and 95 mmol/l in an organic form (i.e. bound in ATP, ADP, creatine phosphate, nicotinamide, adenine dinucleotide etc.). These intracellular forms are readily exchangeable [32].

The normal daily oral phosphate intake is approximately 40 mmol of which approximately 60-70% (i.e. 25-30 mmol) is absorbed in the duodenum and upper jejunum. The minimum oral requirement of phosphate is about 20 mmol per day. Normal urinary phosphate excretion ranges between 10 and 40 mmol per day and about 15 mmol per day is excreted with feces. Approximately 180 mmol phosphate is filtered in the kidneys per day of which 70-85% is reabsorbed in the proximal tubule and 15-30% in the distal nephron. PTH induces phosphaturia by an inhibition of the sodium-phosphate cotransport in the proximal tubule. There is no tubular secretion of phosphate. Ca 2+ controls urinary phosphate excretion indirectly via PTH secretion. Tubular reabsorption of phosphate increases up to a maximum (TmPO 4 ); whereafter phosphate is excreted in the urine. PTH, decreases TmPO 4 [32].

Plasma Pi level is maintained within a narrow range through a complex interplay between intestinal absorption, exchange with intracellular and bone storage pools, and renal tubular reabsorption [32, 35]. Most of the factors identified until now, controlling Pi homeostasis, decrease renal reabsorption and intestinal uptake [36]. The key factors are the type 2 Na + -Pi cotransporters NPT2a, b and c [35, 36]. NPT2a is a major molecule expressed in the renal proximal tubular cells and in osteoclasts [37, 38], NPT2c is expressed in the same tissues as NPT2a, but must be different from NPT2a since it cannot compensate for functional defects in the NPT2a transporter [39]. The cotransporter NPT2b is expressed in the small intestine, lung, testis and mammary gland [40, 41]

Recent studies of inherited and acquired hypophosphatemia (X-linked hypophosphatemic rickets/osteomalacia, autosomal dominant hypophosphatemic rickets/osteomalacia and tumor-induced rickets/osteomalacia), which exhibit similar biochemical and clinical features, have led to the identification of novel genes, PHEX and FGF23, that also take part in the regulation of Pi homeostasis [42].

In the last decade, a new hormone, klotho, involved in the ageing process [43] has been shown to play a role in the phosphate/calcium metabolism as well [44]. The klotho protein binds to fibroblast growth factor receptors and regulates FGF23 signalling [45]. Klotho also exhibits enzymatic activity modifying the sugar chains of the transient receptor potential vanilloid-5 channel (TRPV5) thereby regulating its activity which involves a durable calcium channel activity and membrane calcium permeability in the kidney [46]. The klotho protein seems to be negatively controlled by dietary Pi [35, 47]. The possible importance in uremic patients remains to be established.

The first part (1a, b and c)

These studies focused on short - (12 weeks) and long-term (103 weeks) effects of intravenous 1 α (OH)D 3 on plasma PTH and plasma Ca 2+ in relation to the doses of 1 α (OH)D 3 given. Further, it was examined whether the marked suppression of plasma PTH induced by 300 days of intermittent intravenous treatment with 1 α (OH)D 3 could be maintained, when the administration was changed from intravenous to the oral route for 16 further weeks and eventually back to intravenous administration for another 16 weeks.

Patients

In part 1a, 24 patients were included. One patient died from a cerebral thrombosis and 2 patients underwent kidney transplantation during the study. Thus, 21 patients completed the study. One patient was treated outside the protocol with exactly the same schedule. He had not finished 12 weeks of treatment when the results were evaluated and was therefore not included in the publication [1].

In part 1b, a total number of 22, the 21 patients from study 1a plus the extra patient, continued the treatment. A total of 9 patients were withdrawn from the study after 12 weeks of treatment due to severe hypercalcemia (n = 1), death from AMI (n = 1), transferral to CAPD (n = 1), RAT (n = 2) and poor compliance especially to the phosphate binder therapy (n = 4). Thirteen patients were followed for 300 days (10 months). Thereafter, 4 patients were excluded due to RAT (n = 2), death from aortic stenosis (n = 1) and death from severe infection (n = 1). Nine patients were followed for 523 days (14 months), whereafter 3 patients were withdrawn due to severe hyperphosphatemia (n = 1), RAT (n = 1) and severe infection and illness (n = 1). Thus, 6 patients were followed for 720 days (2 years). Twenty patients were treated outside the protocol at exactly the same schedule, but not included in the publication [2].

In part 1c , 6 patients from the previous study and the 20 patients treated after the same schedule were included. Seventeen patients completed 300 days. 6 patients completed the following oral phase and 5 patients all three parts. 21 patients were withdrawn for the following reasons: Hyperphosphatemia (n = 8), RAT (n = 6), death from AMI (n = 2), severe infection (n = 3), hypercalcemia (n = 1) and transferral to CAPD (n = 1).

Methods

Blood samples were drawn immediately before the dialysis at day 0 and then at regular intervals for 1-4 weeks. 1 α (OH)D 3 was given at the end of each dialysis session in increasing doses (maximally 4 μ g per dialysis) under careful control of plasma Ca 2+ . If hypercalcemia developed, 1 α (OH)D 3 was temporarily discontinued or a lower dose given. When necessary, the dosage of oral phosphate binder was adjusted.

Outcome measures

Outcome measures were plasma PTH, plasma Ca 2+ , and the doses of 1 α (OH)D 3 .

The second part (2a and b)

These studies focused on long-term effects (88 weeks in hemodialysis patients and 52 weeks in CAPD patients) of a treatment combining 1 α (OH)D 3 , CaCO 3 phosphate binders (instead of aluminum containing binders) and a decreased calcium concentration in the dialysis fluid (1.25 mmol/l ) in an attempt to avoid hypercalcemia.

Patients

In part 2a, 60 patients on hemodialysis were included of who 54 completed the first 12 weeks. One patient had tertiary hyperparathyroidism and was scheduled for parathyroidectomy, 2 patients were transferred to CAPD, 2 had RAT, and 1 died in septicemia.

Eighteen patients were withdrawn after 12 weeks due to AMI (n = 5), RAT (n = 4), parathyroidectomy (n = 1), severe hyperphosphatemia due to poor compliance (n = 4), transfer to other dialysis units (n = 3), and infection (n = 1). Thus, 36 patients completed 52 weeks of the study. Seven patients were withdrawn after 52 weeks of treatment due to AMI (n = 2), RAT (n = 3) and death of severe infection (n = 2). Thus, 29 patients were followed for 88 weeks.

In part 2b, 41 patients treated by CAPD were included. Thirty-nine of these completed the first 12 weeks of treatment. One patient had a RAT and 1 was excluded due to peritonitis. Further 9 patients were withdrawn after 12 weeks due to AMI (n = 1), RAT (n = 3), exacerbation of chronic obstructive pulmonary disease and inability to perform CAPD (n = 1), transferral to hemodialysis after own whish (n = 1), severe peritonitis (n = 2), and operation for ventral hernia (n = 1). Thus, 30 patients were followed for 52 weeks.

Methods

Two separate sets of blood samples were obtained as basal values. CaCO 3 was then initiated as oral phosphate binder therapy in patients who received aluminum containing phosphate binder at inclusion. Aluminum containing phosphate binders were, however, still allowed, if judged necessary by the investigator.

The calcium concentration in the dialysis fluid was decreased from 1.75 or 1.50 mmol/l, depending on the previous treatment concentration, to 1.25 mmol/l in the patients treated by hemodialysis and from 1.75 mmol/l to 1.25 mmol/I to patients treated by CAPD.

In patients on chronic hemodialysis who already were being treated with oral 1 α (OH)D 3 the administration route was changed to intravenous administration. Patients already treated with intravenous 1 α (OH)D 3 remained initially on an unchanged dose, and patients on CAPD who were already treated by intermittent oral 1 α (OH)D 3 remained initially at unchanged doses.

Blood samples were obtained immediately before the dialysis from patients treated by hemodialysis and in the morning in the outpatient clinic from patients treated by CAPD. Blood samples were drawn at day 0 and then at regular intervals of 1-4 weeks. The maximum dose of 1 α (OH)D 3 was 12 μ g/week given under careful control of plasma Ca 2+ . If hypercalcemia developed, 1 α (OH)D 3 was temporarily discontinued or a lower dose given. If necessary the dosage of oral phosphate binder was adjusted.

Outcome measures

Outcome measures in the two groups of patients (treated by hemodialysis and CAPD) were plasma levels of PTH, Ca 2+ and phosphate, the doses of oral phosphate binder used, and the doses given of 1 α (OH)D 3 .

In patients treated with hemodialysis the biochemical bone markers, osteocalcin, alkaline phosphatases, and procollagen type 1 c-terminal extension peptid (P1cP) were measured. BMC was measured in the lumbar spine, femoral neck and also the femoral shaft, as suggested by Ruedin et al. in uremic patients treated with 1,25(OH) 2 D 3 [196]. Plain X-rays of hands, lumbar spine and hip were performed, too.

Actual plasma Ca 2+ , pH and plasma Ca 2+ adjusted to pH = 7.4 were analysed at the beginning and at the end of two hemodialysis sessions with the 2 different dialysis solutions.

The dialysis fluid used for CAPD was a commercial solution, and simultaneously with the decrease in the calcium concentration, the magnesium concentration was decreased from 0.75 mmol/l to 0.25 mmol/l, and the lactate concentration increased from 35 mmol/l to 40 mmol/l. Therefore, a peritoneal mass transfer (PET) evaluating calcium, phosphate, magnesium, lactate, creatinine, urea, glucose and albumin transport was performed.

The third part (3a and b)

These studies focused upon the pharmacokinetic differences between intravenous and oral administration of 1,25(OH) 2 D 3 and 1 α (OH)D 3 and upon the acute effects of different doses of the two compounds on the plasma levels of PTH, Ca 2+ and phosphate.

Patients

In part 3a, 6 healthy volunteers and 12 terminal uremic patients were included, 7 on CAPD and 5 on hemodialysis. All patients completed the planned schedules.

In part 3b, 11 uremic patients on chronic hemodialysis were included. Eleven patients completed the first 3 schedules. Between the third and fourth part of the study, one patient died due to a severe pulmonary infection, two patients started on regular treatment with 1 α (OH)D 3 and one patient did refused to participate. Thus, 7 patients completed this part of the study.

Method

In part 3a, 6 patients received 4 μ g of 1 α (OH)D 3 intravenously and 4 μ g of 1 α (OH)D 3 orally, while 6 other patients received 4 μ g of 1,25(OH) 2 D 3 intravenously and 4 μ g of 1,25(OH) 2 D 3 orally. For comparison, 6 healthy volunteers passed through all 4 schedules.

Two separate sets of blood samples were obtained as basal values and subsequently at regular intervals for the following 72 hours.

In part 3b, 11 uremic patients on chronic hemodialysis were included. The study was divided into 4 parts. In part one, 10 ml of isotonic NaCl was injected as a bolus and blood samples obtained at specific times for the following 72 hours. With an interval of at least 1 week, the same procedure was performed, but with intravenously administration of 10 μ g of 1,25(OH) 2 D 3 or 10 μ g of 1 α (OH)D 3 in a randomised fashion. With a further interval of at least 3 weeks, the schedule was repeated, but now with administration of 10 μ g of the other vitamin D analog.

An additional fourth part was added to the study, including only 1,25(OH) 2 D 3 in a smaller dose (6 μ g), because no calcemic response was observed during the first 24 hours after injection of 10 μ g of 1 α (OH)D 3 while a significant increase in plasma Ca 2+ was seen after injection of 10 μ g of 1,25(OH) 2 D 3 . The treatment schedule was the same, except that the patients were only followed for 24 and not 72 hours.

Outcome measures

Outcome measures in part 3a were pharmacokinetic parameters as bioavailability, volume of distribution (Vd) and the metabolic clearance rate (MCR), plasma PTH and plasma Ca 2+ in relation to time.

Outcome measures for part 3b were plasma Ca 2+ and plasma levels of PTH measured by one "Whole" and two "Intact" PTH assays in relation to time.

11.1 Plasma Ca 2+

Ionized calcium (Ca 2+ ) is the biological active form of calcium [197-199]. Since the mid-1970s, reliable automatic methods based on ion-selective electrodes have been available for analysis of plasma Ca 2+ [200]. In the studies included in the present review, plasma Ca 2+ , pH and plasma Ca 2+ adjusted to pH = 7.4 were all analyzed by a calcium ion electrode analyzer (ICA1) [1], ICA 2 [2-5], and ABL555 [6, 7], produced by Radiometer, Copenhagen, Denmark.

Based on the knowledge that calcium is the major regulator of PTH synthesis and secretion [60, 66], the intention in these studies was to stabilise plasma Ca 2+ at the upper end of the normal range [1-5] thus minimizing PTH secretion.

Blood samples for determination of plasma Ca 2+ and pH were placed on ice and measured immediately.

11.2 Parathyroid hormone

PTH 1-84 is rapidly cleared from the circulation with a disappearance half-time of about 2 minutes [59]. Removal of PTH from the blood occurs by extraction in the liver for 60-70% and in the kidneys for 20-30% [59, 201]. The Kupffer cells are responsible for both the rapid clearance and the extensive proteolysis of the PTH molecule that occur in the liver. Less than 10-20% of secreted PTH 1-84 is metabolized peripherally to circulating C-terminal fragments. However, 50-90% of the total circulating PTH immunoreactivity is C-terminal fragments because the clearance of C-terminal PTH fragments, occurring mainly via glomerular filtration, is significantly slower than that of PTH 1-84. Furthermore, C-terminal fragments are secreted together with PTH 1-84 by the parathyroid glands [59, 202]. Generated N-terminal PTH fragments are rapidly cleared by the liver and normally difficult to demonstrate in plasma [201].

Radioimmunoassays for analysis of PTH have been available since 1963, but results obtained by different assays varied considerably [203, 204]. The reason was that different assays used polyclonal antisera that were generated against different regions of the PTH molecule. Most antisera were directed toward the middle or C-terminal end of the PTH molecule. Thus, both the PTH 1-84 and the C-terminal fragments of PTH present in the circulation were measured by these assays [203]. Efforts were therefore directed towards developing radioimmunoassays measuring the N-terminal region of the PTH molecule - which is responsible for the known biological effects of PTH. Although such assays had clear theoretical benefits, their poor sensitivity limited clinical use [205]. Nussbaum et al. developed a two-site immunometric method that was offering the possibility of extracting what was thought to be the pure PTH 1-84 from the complex mixture of PTH 1-84 and fragments in the circulation [206]. Most available data that are comparing plasma levels of PTH with bone biopsies are obtained by such first generation immunoradiometric assays. These assays have proved to be adequate screening tools separating high turnover bone disorders with high PTH from low turnover bone disorders with low PTH [89]. The "Intact" PTH assay (Allegro, Nichols Institute Diagnostics, CA,USA) is a commonly used assay, which also was employed in the present studies [1-7]. This immunoradiometric assay makes use of a goat antibody immobilised onto plastic beads binding only to PTH 39-84, i.e. the C-terminal part, and another radiolabelled goat antibody that binds to the N-terminal PTH region - the specific binding-site is, however, not exactly known [207]. Such an assay is supposed to measure all PTH 1-84. Similar assays are available, and in one of our studies [7] the samples were measured simultaneously by another Intact assay, too (Total PTH, a part of the DUO PTH Kit from Scantibodies Laboratories, Santee, CA, USA). This assay uses goat polyclonal anti-PTH 39-84 coated beads as a universal solid phase and a second polyclonal antibody directed against the PTH 7-34 region. To evaluate a possible effect of 1 α (OH)D 3 on the peripheral metabolism of PTH , plasma PTH was further analysed by a C-terminal iPTH 53-84 (MILAB, Malmö Immunlaboratorium AB, Sweden) assay, in three of the studies [1-3], and in one study [3] also by a N-terminal iPTH 1-34 radioimmunoassay (Allegro, Nichols Institute, USA).

The optimal concentration of plasma PTH to prevent extraskeletal complications in chronic renal failure, when measured by the intact PTH assay (Allegro, Nichols Institute Diagnostics, CA, USA) is not precisely known [96, 208, 209]. The K/DOQI- guidelines recommend 150 - 300 pg/ml when analysed by this intact PTH assay [58]. It has previously been shown that a level of PTH 2-4 times the upper normal limit was necessary [96, 208] in order to maintain the coupled processes of bone resorption, formation and normal features of skeletal remodelling in uremic patients. This remarkable finding has been explained as a consequence of the uremic state in it self [210], by "skeletal resistance to PTH" due to down- regulation of the PTH1-receptor (PTHR1) [211], by activation of the RANK-RANKL system and elevated OPG-level [212], and, finally, by decreased levels of growth factors and/or increased levels of inhibitory substances blocking the action of growth hormone such as bone morphogenetic proteins (BMP) [213].

PTH 1-84 acts through a specific PTHR1 receptor. The receptor requires the N-terminal amino acids of PTH to be activated. In plasma from uremic patients, non 1-84 PTH circulating fragments interfering significantly with the commercial "Intact PTH assays" have been demonstrated [214, 215]. This iPTH was co-migrating together with the PTH 7-84 fragment when fractionated by HPLC, and the concentration of the fragments in plasma increased progressively with decreasing GFR [216]. These observations together with reports that there might exist another PTH receptor with binding specificity for the C-terminal region of PTH [217, 218] initiated examinations of the possible relevance of such a potential C-PTH receptor as a modulator of bone cell activity. By infusion studies in rats, it has been demonstrated that the C-terminal fragments of PTH 7-84, 39-84 and 53-84 inhibit the increase in plasma Ca 2+ induced by PTH 1-84 and PTH 1-34 in thyro-parathyroidectomized rats [219, 220]. In vitro studies have further demonstrated that bone resorption, as measured as 45 Ca released from labelled neonatal mouse calvariae, induced by 1,25(OH) 2 D 3 and PTH 1-84 was diminished after incubation with PTH 7-84 [221]. Also impaired formation of osteoclast-like cells in response to 1,25(OH) 2 D 3 in murine bone marrow cultures after incubation with PTH 7-84 and PTH 39-84 was seen [221]. Thus these large C-terminal fragments may play a role in the regulation of osteoclastogenesis and may be part of the mechanisms behind the observed skeletal resistance to PTH in uremia [219]. To overcome the problem of co-measured large C-terminal fragments, newer PTH assays have now been developed - the second generation immunoradiometric assay ("Whole PTH", Scantibodies Laboratories, Santee, CA, USA). This assay uses a specific goat polyclonal PTH antibody directed against the N-terminal PTH 1-4 region and polyclonal anti-PTH 39-84 coated beads as the universal solid phase. The results obtained with this assay were in one of our studies [7] compared to results obtained by two first generation immunoradiometric assays - "Total" PTH, (Scantibodies Laboratories, Santee, CA, USA) and the "Whole" PTH assay (part of DUO PTH Kit). Sufficient knowledge has not been accumulated to establish the predictive power of these newer assays with regard to secondary hyperparathyroidism and renal bone disease.

Blood samples for determination of plasma PTH was immediately placed on ice, centrifuged at 4°C and stored at -20°C until analysed.

11.3 Plasma-1,25(OH) 2 D 3 and 25(OH)D 3

Plasma 1,25(OH) 2 D 3 levels in patients suffering from chronic uremia were previously assumed to be independent of the plasma levels of 25(OH)D 3 [222] although the existence of an extra-renal 1a-hydroxylation and synthesis of 1,25(OH) 2 D 3 in chronic renal failure is well-known [223]. Studies on plasma 25(OH)D 3 in patients with varying degrees of renal failure described a correlation between the concentrations of p-1,25(OH) 2 D 3 and p-25(OH)D 3 [23]. Some authors have suggested that low levels of p-25(OH)D 3 in chronic uremic patients may be a risk factor for the development of secondary hyperparathyroidism and reduced BMD in the hip [224-226]. But there is no evidence from patients on dialysis that supplementation with D 2 or D 3 to raise p-25(OH)D 3 increases p-1,25(OH) 2 D 3 or lowers the elevated levels of PTH. Nevertheless, it is now generally recommended that p-25(OH)D 3 should be kept higher than 30 ng/ml in patients on chronic dialysis [58].

In the present studies, p-1,25(OH) 2 D 3 was measured in different conditions after administration of 1 α (OH)D 3 or 1,25(OH) 2 D 3 . P-25(OH)D 3 was also determined in one study to establish the level in our population of uremic patients [1].

1,25(OH) 2 D 3 and 25(OH)D 3 were extracted from plasma with diethylether and the extracts chromatographed [227]. 25(OH)D 3 was measured by a competitive protein-binding assay. The binding protein was obtained from the cytosolic fraction of rachitic rat kidneys [227]. 1,25(OH) 2 D 3 was measured by a competitive protein binding assay using calf thymus cytosol as the source of binding protein [227]. From 1996 and onwards, the analyses for 1,25(OH) 2 D 3 and 25(OH)D 3 were available in our own laboratory.

Blood samples for the determination of 1,25(OH) 2 D 3 were immediately placed on ice, centrifuged at 4°C and stored at -20°C until analysis.

11.4 Biochemical bone markers
- indirect markers of bone disease

While bone biopsies are the gold standard in directly following on-going bone disease of uremic patients, a few simple biochemical parameters, besides PTH may constitute the first-line in diagnostic algorithms in daily clinical practice.

Indirect markers reflecting bone formation are osteocalcin, procollagen type 1 c-terminal extension peptide (P1cP) and alkaline phosphatases. No useful markers reflecting bone resorption in renal osteodystrophy have so far been identified [88, 100, 228].

Plasma alkaline phosphatase were measured in 6 studies [1-6] by a standard laboratory method which was not discriminating between the bone fraction and other fractions.

Plasma osteocalcin was measured in 3 of the studies [2-4] by an ELIZA method.

Plasma P1cP was measured in one study [4] by a radioimmunoassay (Farmos Diagnostica, Finland) [229].

Aluminum adversely affects the differentiated function of osteoblasts and was previously a major factor in the development of low turnover bone disease [230]. Aluminum toxicity was suspected when p-aluminum was > 100 μ g/l [231]. Serum-aluminum was measured in 5 of the studies [1-5] by electrothermical atomic absorption photometry.

11.5 Other analyses

Total plasma calcium and plasma inorganic phosphate (Pi) were measured by photometry. Hemoglobin, thrombocytes, leucocytes, plasma lactate dehydrogenase, plasma ALAT and plasma ASAT, p-L og D-lactate, plasma PP (factor 2,7 and 10), plasma protein, plasma creatinine, plasma urea, plasma magnesium, plasma glucose and serum albumin were all measured by standard laboratory tests.

Blood Pressure was measured by a manual mercury sphygmomanometer in a sitting position.

Weight was measured on the same electronic weight each time.

The standardized Peritoneal Equilibrium Test [232] was used to calculate mass transfer of calcium, phosphate, magnesium, lactate, creatinine, urea, glucose and albumin.

11.6 Imaging techniques - markers of renal bone disease and extraskeletal calcifications

X-ray examinations and evaluation of bone mineral content (BMC) have been used over time to demonstrate bone disease secondary to prolonged secondary hyperparathyroidism [233]. These investigations may provide further information when combined with measurements of indirect biochemical bone markers, but no combination has so far been able to replace bone-biopsies [233, 234]. The radiological presentation of the bone in advanced secondary hyperparathyroidism is very characteristic. The increased osteoclastic bone resorption secondary to excess PTH results in radiographically evident subperiosteal erosions most often present in the hands along the radial margins of the middle phalanges of the second and third fingers [233]. Osteosclerosis is caused either by excessive accumulation of poorly mineralized osteoid, which radiographically will appear denser than normal bone, or by an exaggerated osteoblastic response following osteoclastic bone resorption. The increased bone density may be generalized or more often found in the axial skeleton where the midplane of the vertebral bodies shows a normal density while the endplate exhibits sclerosis (a so called rugger-jersey spine) [233]. Periosteal reactions may also be seen, most often in the metatarsal, femoral and pelvic bones [235]. These last mentioned radiological features are now rarely seen due to improved therapeutic management. The correlation between radiographic changes and clinical signs and symptoms of bone disease is poor [236].

The diagnosis of adynamic bone disease rests on histomorphometric and histodynamic findings of a low bone turnover combined with a lack of increased thickness of osteoid seams and osteoid. This condition cannot be diagnosed based on plain X-ray pictures [86].

Other complications are now pictured on X-rays such as metastatic and vascular calcifications [237] that in the extreme clinical form may present as calciphylaxis [238]. A score based on calcifications observed on plain x-rays of the abdominal aorta has been developed [239] and showed a good correlation to the more sophisticated Electronic Beam Computer Tomography (EBCT) in determining cardiovascular calcification. Such methodologies may prove useful to guide further therapeutic choices in dialysis patients in the future. In one of the present studies [4], radiographic evaluations including anterior-posterior and lateral views of the columna lumbalis, anterior-posterior view of both hands and both hips including the femoral shaft were performed. At the end of the study, all radiographs of each patient were collected for comparison and evaluated by the same radiologist. The pictures were examined systematically for bone changes only and not for possible vascular calcifications.

The increasing availability of Dual photon- and now Dual-X-ray absorptiometry (DEXA) facilitates measurement of bone mineral density (BMD g/cm 2 )/bone mineral content (BMC) in both healthy people and patients with chronic renal failure. BMD/BMC is useful in diagnosing osteoporosis and predicting the fracture risk in patients with normal renal function, but the method has not been validated in patients with chronic renal failure [240]. In renal osteodystrophy, BMD/BMC can both be decreased and increased when compared to healthy man. This is due to regional differences in the skeleton, and the presence of severe secondary hyperparathyroidism and adynamic bone disease [88, 241, 242]. False elevated high BMD, not reflecting changes in bone, may be due to vascular calcifications [233]. The risk of fractures is increased in patients with chronic renal failure, but the correlation to BMD is poor [243-246]. BMD is, however, only one of several assessment tools in evaluating renal osteodystrophy and should be judged in context with biochemical variables and bone histology if available [246]. In one study [4], we evaluated BMC by dual photon absorptiometry in the lumbar spine, femoral neck and the femoral shaft. All calculations for one particular patient were performed with reference to the previous measurement in that person.

12.1 Benefits of treatment by intravenous
1 α (OH)D 3 on the secondary hyperparathyroidism
in patients on chronic hemodialysis

Oral treatment with 1,25(OH) 2 D 3 and 1 α (OH)D 3 has been used with great benefit as treatment of secondary hyperparathyroidism and renal osteodystrohpy in uremic patients for many years [149, 175, 180, 181, 247]. Only sparse experience with intravenous administration of 1,25(OH) 2 D 3 and 1 α (OH)D 3 was available on these conditions until the end of the 1980s. In 1989, results of our first study on intravenous administration of 1 α (OH)D 3 to patients on chronic hemodialysis were published [1]. In a short-term study (84 days) [1], we found a marked suppression of plasma PTH by 67±6%, Figure 1 .

Our results were in agreement with those obtained in a small Swedish study including 7 patients [248]. In the swedish study, intravenous administration of 1 α (OH)D 3 at the end of each hemodialysis session induced a significant suppression of plasma PTH by 40±20%. Similar results on intravenous administration of other vitamin D analogs have been reported [74, 157, 165, 167]. Although the studies were not quite similar in design and patient material, they were on the whole comparable. The patients included were all treated by chronic hemodialysis, all had elevated plasma PTH levels, and all were treated with a vitamin D analog intravenously with the doses being adjusted to the plasma Ca 2+ level. Slatopolsky et al. administered 1,25(OH) 2 D 3 intravenously for 8 weeks and reported a suppression of plasma PTH by 70.1±3.2% [74]. Akizawa et al. administered 22-oxa-1,25(OH) 2 D 3 intravenously for 12 weeks and reported a suppression of plasma PTH from 905.1±66.7 pg/ml to 590.3±73.8 pg/ml ( ≈ 35%) [167]. Sprague et al. administered 19-nor-1,25(OH) 2 D 2 intravenously for 12 weeks and reported a suppression of plasma PTH from about 700 to 300 pg/ml (ª57%) [157]. Maung et al. administered 1 α (OH)D 2 intravenously for 8 weeks and reported a suppression of plasma PTH by 65±3,9% [165].

The long-term effect of intravenous 1 α (OH)D 3 was addressed in our following studies [2-4]. When plasma phosphate and plasma Ca 2+ were carefully controlled, it was possible, in our group of patients with elevated PTH, to keep plasma PTH stable for more than 2 years [2-4], Figure 2 .

In advanced chronic renal failure, where the proliferation of the parathyroid glands appears to be more polyclonal [76], the number of receptors for 1,25(OH) 2 D 3 and calcium are reduced in the glands [68, 75, 76]. The enhanced growth of parathyroid cells in secondary hyperparathyroidism has been attributed to hyperphosphatemia per se [64] and a relative or absolute deficiency of 1,25(OH) 2 D 3 with or without concomitant hypocalcemia [69]. In vitro studies in primary cultures of bovine parathyroid cells have shown that 1,25(OH) 2 D 3 completely inhibit cell proliferation [249]. The need for parathyroidectomy in chronic uremic patients increases with the time on dialysis [250] maybe as a consequence of the nodular growth [251] and resistance to treatment with 1,25(OH) 2 D 3 . Regression of parathyroid hyperplasia, as shown by ultrasound, in patients on hemodialysis following oral 1,25(OH) 2 D 3 treatment has been reported by some authors [252], but not by others [253]. The reason for this discrepancy could be differences in the nodular hyperplastic stage of the parathyroid glands at initiation of treatment in the two studies.

The long-term effect of intravenous treatment by 1 α (OH)D 3 observed in our study may indicate a delay in the proliferation of the parathyroid cells although no imaging methods to monitor parathyroid gland size were used. The direct effect of intravenous 1 α (OH)D 3 was, in our studies, a marked decrease in plasma PTH before any increase in plasma Ca 2+ occurred [1-4]. The reduction in plasma PTH was dependent on changes in plasma Ca 2+ and the doses of the vitamin D-analog used when evaluated by multivariate analysis [1, 2], Figure 3 .

The long-term sustained effect achieved by intravenous administration of 1 α (OH)D 3 may be due to an upregulation of the numbers of VDR-receptors in the parathyroid glands in response to 1,25(OH) 2 D 3 [254]. An influence of 1,25(OH) 2 D 3 on the calcium sensing receptor gene transcription in the parathyroid cells cannot be excluded either [255]. A normalisation of the set point of the parathyroid hormone-Ca 2+ relation curve [256] and a shift of the sigmoidal curve to the left when compared to that observed in untreated chronic renal failure [257] have also been reported.

Long-term studies of intravenous administration of a vitamin D analog in patients on chronic hemodialysis have only been sparsely reported. Rodriguez et al. reported that intravenous administration of 1,25(OH) 2 D 3 for 42 weeks at the end of each dialysis resulted in a suppression of plasma PTH from 890±107 to 346±119 pg/ml [258]. Akizawa et al. reported that intravenous administration of 22-oxa-1,25(OH) 2 D 3 , at the end of each dialysis for 12 months kept plasma PTH at the level achieved after 12 weeks [167]. Finally, Lindberg et al. reported that intravenous administration of 19-nor-1,25(OH) 2 D 2 at the end of each dialysis for 13 months resulted in a suppression of plasma PTH from 628.3±27.65 to 409±35 pg/ml [159].

12.2 Limitations in the clinical use
of intravenous 1 α (OH)D 3

As for 1,25(OH) 2 D 3 the use of 1 α (OH)D 3 in doses that effectively suppress plasma PTH may be limited by the development of hypercalcemia, hyperphosphatemia and an increased Ca × P product due to increased intestinal absorption and mobilization of calcium and phosphate from the bone [27]. The toxicity may be further aggravated by the concomitant use of large doses of oral phosphate binders containing calcium. This problem has large implications and is very serious as evidence is increasing that cardiovascular mortality is correlated to abnormalities in the Ca × P product in uremia [8, 56]. In our studies [1-4], all patients had low to normal p-Ca 2+ at the entrance. During the treatment period it was necessary to reduce the doses of 1 α (OH)D 3 due to an increase in plasma Ca 2+ . In our study, the accepted levels of plasma Ca 2+ were in the upper normal range or slightly elevated. With reference to the recent K/DOQI guidelines [58], plasma Ca 2+ should be in the mid-normal to low-normal end of the range.

As expected, all patients in our studies had plasma phosphate levels above the normal range despite treatment with oral phosphate binders [1-4]. During treatment, many patients experienced increases in plasma phosphate to > 2 mmol/l [1-4]. This problem was handled by an increase in the doses of phosphate binder. Severe hyperphosphatemia per se has been reported to cause failure of active vitamin D treatment to suppress PTH [258, 259]. Data showing an increased mortality due to hyperphosphatemia and an increased Ca × P product in uremic patients were not available until 1998 [55]. Although the Ca × P product was not calculated, the obtained mean values of plasma total calcium and plasma phosphate in our studies [1-4] indicate that at least part of the patients must have had elevated Ca × P product values above the now recommended 4.4 mmol2/l2 [58].

No direct comparison has been made between intravenous use of 1 α (OH)D 3 and other vitamin D analogs. However, intravenous administered 19-nor-1,25(OH) 2 D 2 has been directly compared to intravenous administered 1,25(OH) 2 D 3 in patients on chronic hemodialysis [157]. Although animal experiments have demonstrated a lower potency of 19-nor-1,25(OH) 2 D 2 than 1,25(OH) 2 D 3 in stimulating intestinal calcium and phosphate absorption [260], a direct comparison between 19-nor-1,25(OH) 2 D 2 and 1,25(OH) 2 D 2 in hemodialysis patients was not able to demonstrate any differences in the frequency or severity of hypercalcemia and hyperphosphatemia. Fewer episodes of sustained hypercalcemia and increased Ca × P product were, however, observed in the 19-nor-1,25(OH) 2 D 2 group [157]. In a dose equivalence study between 19-nor-1,25(OH) 2 D 2 and 1 α (OH)D 2 the incidences of hypercalcemia, hyperphosphatemia and elevated Ca × P product were similar for doses with a comparable PTH suppression [261]. In a crossover study between 19-nor-1,25(OH) 2 D 2 and 1,25(OH) 2 D 3 , lower serum calcium, lower Ca × P product and lower PTH levels were obtained with 19-nor-1,25(OH) 2 D 2 [160] than with 1,25(OH) 2 D 3 . 22-oxa-1,25(OH) 2 D 3 has been reported to have less hypercalcemic effect than 1,25(OH) 2 D 3 in uremic rats by some [262], but not by others [263]. In a clinical study, intravenous administration of 22-oxa-1,25(OH) 2 D 3 induced an increase in plasma Ca 2+ , and a reduction in dose was necessary [166]. 1 α (OH)D 2 was found less calcemic than 1 α (OH)D 3 in rat studies [264], but still a dose reduction was needed in clinical studies [164, 165]

12.3 The optimal mode of administration

The first report on the use of high doses of intravenous 1,25(OH) 2 D 3 in 20 long-term chronic hemodialysis patients (8 weeks) was published by Slatopolsky at al. [74]. Subsequently, it was shown that intravenous administration of 1,25(OH) 2 D 3 in uremic patients effectively promoted healing of osteitis fibrosa resistant to oral 1,25(OH) 2 D 3 treatment [265]. A higher peak concentration of 1,25(OH) 2 D 3 and a greater "area under the curve" were found using intravenous instead of oral administration [74]. Following these findings, intravenous treatment with active vitamin D was preferred in patients treated by chronic hemodialysis instead of the oral route. We have shown [6], that the peak concentration of plasma 1,25(OH) 2 D 3 and the "area under the curve" is higher when 1 α (OH)D 3 is administered intravenously instead of orally. In general, however, significantly lower concentrations were found after administration of 1 α (OH)D 3 than after administration of 1,25(OH) 2 D 3 both when given intravenously and orally [6]. The administration of active vitamin D was changed from daily to intermittent - mainly 3 times a week given at the end of each dialysis [193]. Several short-term ( < 24 weeks) comparative studies have been performed both as cross-over and parallel studies including 1,25(OH) 2 D 3 , 1 α (OH)D 3 , and 1 α (OH)D 2 [194, 253, 266, 267]. No convincing evidence for the intravenous route has been found in these short-term studies. Initially a faster suppression of PTH was seen in two studies, when the active vitamin D was given intravenously, but after 4 months of treatment no difference existed be-tween the two groups [194, 253]. Further long-term studies are needed to clarify whether the intravenous and oral routes are equal regarding efficacy and side effects. The available studies indicate that the intermittent administration rather than the intravenous route is of importance. When comparing daily oral and intermittent intravenous administration in patients on chronic hemodialysis, no significant differences were found regarding hypercalcemia and hyperphosphatemia [195]. A study in rats suggested, however, that the dose of calcitriol delivered directly to the target organs such as the parathyroid glands might be of importance for the suppression of PTH [69]. An interesting question was whether it is possible to maintain the initial suppression of PTH found after intravenous administration of 1 α (OH)D 3 by subsequent intermittent oral administration. We addressed this question in one study [3] and although the number of patients completing the study was small, we showed that the marked suppression of plasma PTH obtained by treatment with intravenous 1 α (OH)D 3 could indeed be maintained after shifting to intermittent oral treatment. After 16 weeks of oral treatment the route of administration was shifted back to intravenous and no further suppression of PTH was induced, Figure 4 .

In patients treated by CAPD, repeated intravenous administration of 1,25(OH) 2 D 3 is not practical. Therefore high pulse doses of oral and intraperitoneal administration of 1,25(OH) 2 D 3 one to three times a week became the standard treatment in these patients [268-273]. Intraperitoneal 1,25(OH) 2 D 3 is practicable, effective and safe in the treatment of secondary hyperparathyroidism in CAPD patients, but has no advantages when compared to oral treatment [274]. We studied the effect on secondary hyperparathyroidism of intermittent oral 1 α (OH)D 3 in combination with "low calcium dialysis fluid" in CAPD patients [5]. The doses of 1 α (OH)D 3 were lower in our study than in other CAPD studies which have observed an improvement of the secondary hyperparathyroidism by the use of "low calcium dialysis fluid" and oral intermittent 1,25(OH) 2 D 3 [275, 276]. This was also the case when compared to results obtained in our previous study on hemodialysis patients, who had intravenous administration of 1 α (OH)D 3 [4]. It is notable that the prescribed weekly doses of 1 α (OH)D 3 in our present CAPD study were lower after 52 weeks of attempted optimal treatment (1.4 ±0.31 μ g/week), than in our study performed in patients treated by hemodialysis (2.88± 0.52 μ g/week) including patients with exactly the same degree of secondary hyperparathyroidism (PTH in CAPD patients 151± 41pg/ml [5], and PTH in hemodialysis patients 151± 44 pg/ml [4]). A transient increase in plasma PTH took place when calcium was changed to low calcium dialysis fluid in the CAPD patients, but no further aggravation of the secondary hyperparathyroidism was observed thereafter [5].

Moe et al. demonstrated in 1998 in a randomized trial that pulse therapy and daily administration of calcitriol are similarly effective and safe for the treatment of mild to moderate secondary hyperparathyroidism in CAPD patients, despite higher peak levels of 1,25(OH) 2 D 3 found with pulse therapy [277].

12.4 Introduction of a non-calcium containing phosphate binders and of low calcium dialysis fluid (1.25 mmol/l) - Consequences

Changes in plasma Ca 2+ during dialysis
and possible effects on renal bone disease

In order not to increase the total calcium load to the patients when the phosphate binder therapy was changed from aluminum to calcium containing binders calcium containing, an introduction of "low calcium dialysis fluid" was necessary [152, 278-281].

Sperschneider et al. showed that pre-existing microcalcifications progressed and new appeared in vessels and soft-tissue areas of the hand during a 3-year study on uremic dialysis patients when the calcium concentration in the dialysis fluid was 1.75 mmol/l ("high calcium dialysis fluid"), no active vitamin D was administered, and CaCO 3 was given as an oral phosphatebinder [281]. When "low calcium dialysis fluid" was introduced it was feared that a decrease in plasma Ca 2+ , would aggravate secondary hyperparathyroidism and thereby accelerate bone turnover [282]. We addressed this problem in our unit in patients treated by both hemodialysis [4] and CAPD [5]. In patients treated by hemodialysis, we found a decrease in plasma Ca 2+ during the dialysis [4], Figure 5 , while others have reported no change [283].

Contrary to patients treated by hemodialysis, patients treated by CAPD are exposed to "low calcium dialysis fluid" 24 hours a day and not only for 4-5 hours 3 times a week. We found in this group of patients [5] both a decrease in plasma Ca 2+ and a change from positive to negative calcium mass transfer (MT) when the calcium concentration in the dialysis fluid was reduced from 1.75 mmol/l to 1.25 mmol/l (as shown In Figure 5 and Figure 6 ). These findings are in accordance with others [284-286]. These changes increase the risk of aggravation of an existing secondary hyperparathyroidism and often require an adjustment of the doses of both phosphate binding CaCO 3 and 1 α (OH)D 3 [5, 287, 288]. The increased plasma Ca 2+ , observed secondary to the extra oral calcium intake when CaCO 3 was used as the principal oral phosphate binder, was found in some studies to reduce the secondary hyperparathyroidism [278, 279, 289]. In other studies, however, an increase in plasma PTH was observed and treatment with active vitamin D necessary, especially in long-term studies [280, 281]. Saisu et al. combined increased doses of calcium containing phosphate binder with unchanged doses of active vitamin D in a group of uremic patients treated by hemodialysis using "low-calcium dialysis fluid" [290]. After 1 year of treatment they observed a trend towards increased plasma alkaline phosphatases and plasma osteocalcin [290]. While high turnover bone disease may be less frequent than previously found, low turnover bone disease (adynamic bone disease) unrelated to aluminum has been described increasingly often in dialysis patients [291, 292]. Adynamic bone disease is presumably more often observed in patients treated by CAPD than in patients treated by hemodialysis [293]. Patients with adynamic bone disease are often asymptomatic and whether the condition is deleterious on a long-term basis is still a matter of debate [86, 245, 294]. The exact pathogenesis of the disorder is unknown, but it is found in patients with prevailing low PTH which may therefore be a pathogenetic factor [86, 294].

Treatment with active vitamin D, which may depress PTH in dialysis patients where PTH is normal or only slightly elevated, is therefore generally avoided. Oversuppression of PTH is reversible, and elevated plasma PTH induced by low calcium dialysis may be an advantage if oversuppression of PTH is suspected [295]. No bone biopsies were performed in our studies and whether the patients with normal or decreased plasma PTH included [4, 5], suffered from low turnover bone disease is an open question. The prevalence of adynamic bone disease in a Danish dialysis population is unknown. Looking at usual risk factors for the development of adynamic bone disease - age, sex, dialysis duration, prevalence of diabetes mellitus, parathyroidectomy or failed transplanted [291], the existence of the condition in our patients cannot be excluded.

Changes in plasma Ca 2+ , plasma PTH and doses of 1 α (OH)D 3

After introduction of "low calcium dialysis fluid" in CAPD, a significant initial decrease in plasma Ca 2+ was observed together with a significant increase in PTH [5]. The doses of 1 α (OH)D 3 were initially increased and subsequently adjusted under careful control of plasma Ca 2+ . After 12 weeks of treatment, plasma PTH decreased to the initial levels [5]. In patients treated by hemodialysis no changes in p-Ca 2+ were observed initially when measured just before the beginning of the next dialysis. Patients with an initial normal PTH demonstrated a temporary increase in PTH after 64 weeks while no changes were observed in patients with initial elevated PTH. In both groups, it was possible to increase the dose of 1 α (OH)D 3 initially. The dose was subsequently adjusted, mostly reduced, with consideration to plasma Ca 2+ . After 88 weeks, PTH did not differ from initial levels in the patients with an initial normal PTH, and was close to normal in the patients with initially elevated PTH [4].

Changes in type of phosphate binder

It was possible to increase the dosages of the oral phosphate binder CaCO 3 after introduction of "low-calcium dialysis fluid" for patients treated by both CAPD and hemodialysis [4, 5]. The CaCO 3 doses used were however higher, and plasma Ca 2+ and plasma phosphate lower in patients on hemodialysis than in patients treated by CAPD. As reported on 50 patients treated by CAPD by Hutchison et al. [296], it was possible to reduce, but not totally exclude aluminum containing phosphate binders. A negative mass transfer for phosphate of -1.9 mmol/exchange and contemporary changes in the dialysate/plasma ratio of phosphate were demonstrated in our study. These changes were not affected by the calcium concentration in the dialysis fluid, in agreement with other studies [284, 297]. Despite negative MT for phosphate, inadequate removal of phosphate by CAPD is well-known. A positive phosphate balance will be the inevitable outcome in a well-nourished CAPD patient unless diet prescription and oral phosphate binders are used [297].

Changes in biochemical bone markers

The influence on biochemical parameters, that indirectly may reflect bone turnover, in hemodialysis patients treated with "low calcium dialysis fluid" combined with intravenuous 1 α (OH)D 3 and CaCO 3 as the principal oral phosphate binder was addressed in one study [4]. Plasma P1cP, a bone marker which has been introduced as a noninvasive index of bone formation in different metabolic bone diseases [298-302], was evaluated. We observed an initial increase in plasma P1cP followed by a decline to basal levels after 6 weeks, which is in accordance with others [300, 301]. Coen et al. compared bone biopsies and different biochemical bone markers in predialysis patients with secondary hyperparathyroidism on treatment with active vitamin D and demonstrated a positive correlation between plasma P1cP and bone formation. However, no significant correlations were found between plasma P1cP and plasma osteocalcin, plasma PTH or alkaline phosphatases [300]. In our study, only a weak but significant correlation was seen between P1cP and osteocalcin. This finding is in accordance with results from Hamdy et al. [301]. Plasma osteocalcin levels have been shown to reflect bone formation in patients on chronic dialysis [303]. Different assays for osteocalcin measure different absolute values due to different antibodies used [304]. Despite the marked suppression of intact PTH levels, no suppression of osteocalcin was observed in our investigation. The future use of biochemical bone markers, individually or in combination with other methods, will undoubtedly improve the diagnosis and the treatment of the complex disorder, renal osteodystrophy [100].

Changes in Bone Mineral Content

Bone lesions and reduced BMC are frequently seen in patients with chronic renal failure. The reduced BMC correlates to age, sex, PTH, duration of the uremia, and number of transplantations [89, 305]. Oral administration of 1 α (OH)D 3 to uremic patients has previously been shown to bring the decrease to a stop and in some patients even to increase BMC of the forearm [306], femoral head and neck [307], and lumbar spine and femoral shaft [196]. In our study, prevention of a further decrease of BMC was observed in the lumbar spine, and femoral neck and shaft in hyperparathyroid patients while a small, but significant decrease was still found in the femoral shaft of patients with normal PTH levels [4].

None of the studies in this thesis have focused directly on the effect of "low calcium dialysis fluid" and intravenous 1 α (OH)D 3 on cardiovascular events. Although only seen in a few patients, soft-tissue calcification disappeared or was reduced as was a single bone cyst on the "low calcium dialysis regime" [4].

Taken together, dialysis using "low calcium dialysis fluid" combined with a change of oral phosphate binder to mainly calcium containing phosphate binders and the introduction of intermittent intravenous/oral 1 α (OH)D 3 can be used without aggravating secondary hyperparathyroidism and without decreasing BMC in uremic patients when plasma Ca 2+ is carefully controlled. "Low calcium dialysis fluid (1,25 mmol/l)" has been recommended as "normal (standard) calcium dialysis fluid" in the USA since 2003 [58] and in Denmark since 2005 [308].

12.5 Plasma PTH measurements depend on the assays used - Are better tools for management of secondary hyperparathyroidism obtainable?

The actually measured changes in plasma concentrations of PTH induced by intravenous and oral administration of 1 α (OH)D 3 and 1,25(OH) 2 D 3 depend upon which PTH assay is used. Plasma PTH was measured in the three studies concerning short-term and long-term intravenous administration of 1 α (OH)D 3 both by the intact PTH assay and by a C-terminal PTH assay measuring PTH 53-84 [1-3]. The initial levels as measured by the C-terminal PTH assay were about 10 times higher than those obtained with the intact PTH assay. Also the suppression of plasma PTH measured by the intact PTH assay was more pronounced (about 60%) than the PTH suppression measured by the C-terminal PTH assay (about 40%). Although we found great differences in the exact levels, there was a highly significant correlation (R 2 = 0.72) between the concentrations measured in the two assays. In one of the studies, we further evaluated whether the route of administration (oral or intravenous) of 1 α (OH)D 3 affected the circulating levels of N- and C-terminal fragments in plasma [3]. The data are shown in Figure 4. As expected, we found a reduction in plasma PTH when measured by both N-terminal-, C-terminal- and Intact PTH assays. The suppression of plasma PTH, when analysed by the Intact PTH, was more pronounced than when analysed by the N-terminal assay, which was again more pronounced than when analysed by the C-terminal PTH assay. A very high correlation between changes in plasma Intact PTH and plasma N-terminal PTH was found which may reflect that the N-terminal PTH assay mainly measures intact PTH, as has been shown in vitro [309]. A less marked correlation was found between changes measured by the intact PTH and C-terminal PTH assay ( Figure 7 ), and between the N-terminal and C-terminal assays, presumably reflecting the accumulation of C-terminal fragments secondary to the reduced renal clearance [202, 207].

The route of administration of 1 α (OH)D 3 did not change the correlations. A new PTH assay (the whole PTH assay) detecting only the bioactive whole PTH 1-84 [219, 310] was hoped to offer a better tool for the management of secondary hyperparathyroidism. This whole PTH 1-84 assay was used in one study where the suppression of plasma PTH was followed after administration of one large intravenous dose of either 1,25(OH) 2 D 3 and 1 α (OH)D 3 and the measurements compared to that of two intact PTH assays [7]. In accordance with previously published studies on chronic uremic patients, the specific PTH values were significantly lower when analysed by the whole PTH assay than when analysed by the intact PTH assays [311]. The results obtained by the whole PTH and intact PTH assays were highly correlated within the whole range investigated, Figure 8 .

The higher levels of plasma PTH found when analysed by an intact PTH assay were ascribed to accumulation of large C-terminal PTH fragments (non PTH 1-84) [7, 219, 310]. A possible modulating effect of large C-terminal PTH fragments on the stimulatory effects of PTH 1-84 on bone turnover was postulated and gave rise to the hypothesis that a ratio between PTH 1-84 and C-terminal PTH had a higher predictive power in distinguishing low- and high-bone turnover than the concentrations measured [209]. Other groups were unable to confirm this hypothesis [311-315]. The whole PTH assay gives more exact measurements of PTH, but whether it will change the understanding of secondary hyperparathyroidism in chronic uremia remains an open question.

12.6 1 α (OH)D 3 - a precursor to 1,25(OH) 2 D 3
or a free active vitamin D analog per se?

In expectation of a more effective inhibitory effect on the parathyroid glands [148], intravenous administration of 1,25(OH) 2 D 3 instead of oral was introduced in 1984 [74]. This expectation was based upon the finding of a higher peak-concentration of 1,25(OH) 2 D 3 after intravenous injection than after oral administration of the same doses [74, 316]. 1 α (OH)D 3 is known to be hydroxylated to 1,25(OH) 2 D 3 by the liver [171, 172]. In many studies regarding intravenous administration of 1 α (OH)D 3 , blood samples for measurements of plasma 1,25(OH) 2 D 3 were drawn immediately before the next dialysis i.e. 2 or 3 days after administration of the last dose. By using this setup we observed [1, 2] a small increase in plasma 1,25(OH) 2 D 3 in accordance with others [189] which correlated with the doses of 1 α (OH)D 3 administered. The peak plasma concentration of 1,25(OH) 2 D 3 achieved after oral administration of 1,25(OH) 2 D 3 is lower than after intravenous administration [6, 74], Figure 9 .

The peak levels of plasma 1,25(OH) 2 D 3 achieved in our study after oral and intravenous administration of 1 α (OH)D 3 were similar [6] to the findings in other studies [74, 185-187, 189] and were only about 50% of the plasma levels achieved after similar doses of 1,25(OH) 2 D 3 . The suppression of PTH was similar after administration of 4 μ g of 1 α (OH)D 3 and 4 μ g of 1,25(OH) 2 D 3 . This may be due to a "threshold level" of 1,25(OH) 2 D 3 in combination with a normal plasma Ca 2+ as has been suggested by Shigematsu et al. [317]. It might however also be due to a conversion of 1 α (OH)D 3 to 1,25(OH) 2 D 3 inside the parathyroid cells [318]. This would explain the benefits found by intermittent oral regimes of both 1,25(OH) 2 D 3 [194, 195, 253] and 1 α (OH)D 3 [3, 5]. Since the peak concentration of plasma 1,25(OH) 2 D 3 after administration of 1 α (OH)D 3 was markedly lower than that obtained after similar doses of 1,25(OH) 2 D 3 [6], another explanation could be that the peak concentration of 1,25(OH) 2 D 3 is of less importance for the direct suppression of PTH secretion than previously assumed. Finally, it's possible that the effect of 1 α (OH)D 3 on the PTH secretion in uremic patients can not solely be explained by the conversion of 1 α (OH)D 3 to 1,25(OH) 2 D 3 . Therefore, the theoretical possibility exists that the 25-hydroxyl group on 1,25(OH) 2 D 3 is not mandatory for the suppression of the PTH gene in the parathyroid glands, but that the 1 α -hydroxyl group is the structural feature required. This assumption is supported by results from an in vitro study from our laboratory where the suppression of PTH secretion from bovine parathyroid cells by 1 α (OH)D 3 was equal to that of 1,25(OH) 2 D 3 [178], and further supported by equal suppression of plasma PTH in uremic patients after administration of single intravenous doses of 4 μ g of 1 α (OH)D 3 and 1,25(OH) 2 D 3 [6], Figure 10 .

No general agreement exists on the acute effects of 1,25(OH) 2 D 3 and 1 α (OH)D 3 on plasma PTH neither in healthy nor in uremic man. In normal subjects, both no change after intravenous administration of 1,25(OH) 2 D 3 [319] and 1 α (OH)D 3 [320] and a decrease of plasma PTH after intravenous administration of 1 α (OH)D 3 [6] have been reported. In uremic patients, both no change after intravenous administration of 1,25(OH) 2 D 3 [321] and 1 α (OH)D 3 [189, 321] and a decrease of plasma PTH after intravenous administration of 1,25(OH) 2 D 3 [6, 190, 322] and 1 α (OH)D 3 [6, 323] have been reported.

In a direct comparative set-up where 1 α (OH)D 3 and 1,25(OH) 2 D 3 were administered intravenously in equally high doses of 10 μ g, we observed that plasma Ca 2+ increased significantly more after administration of 1,25(OH) 2 D 3 than after 1 α (OH)D 3 , in accordance with a previous report [182]. A dose of 6 μ g of 1,25(OH) 2 D 3 caused no increase in plasma Ca 2+ and suppressed plasma PTH by ∼ 60% (p < 0.001) after 24 hours while the suppressive effect of 10 μ g of 1 α (OH)D 3 on PTH was ∼ 20% (p < 0.05), indicating that 1,25(OH) 2 D 3 is a more potent vitamin D analog than 1 α (OH)D 3 . This difference in potency may be due to the kinetics of the conversion of 1 α (OH)D 3 to 1,25(OH) 2 D 3 and does indicate that 1 α (OH)D 3 continuously will be converted to 1,25(OH) 2 D 3 . From our previous studies, it is known that the bioavailability of 1 α (OH)D 3 when measured as 1,25(OH) 2 D 3 is less than 50% of the amount achieved after administration of similar doses of 1,25(OH) 2 D 3 [6]. In a study on the vitamin D receptor in the osteoblast using autoradiography following a single intravenous injection of 3 H-1 α (OH)D 3 and 3H-1,25(OH) 2 D 3 to neonatal rats [324] it was demonstrated that the nuclear uptake in the osteoblasts was delayed, but sustained longer after administration of 1 α (OH)D 3 than after administration of 1,25(OH) 2 D 3 .

This observation can also be explained by the relative long time needed for the conversion of 1 α (OH)D 3 to 1,25(OH) 2 D 3 . In receptor affinity studies in rat intestine comparing 1 α (OH)D 3 and 1,25(OH) 2 D 3 in the presence of 1,25(OH) 2 D 3 the affinity of the receptor for 1 α (OH)D 3 was 900 times less than for 1,25(OH) 2 D 3 [325]. This finding supports the concept that the main effect of 1 α (OH)D 3 takes place after conversion to 1,25(OH) 2 D 3 . Our own study [6] did not exclude a direct effect of 1 α (OH)D 3 on the parathyroid glands, but our results support that the main effect in vivo of 1 α (OH)D 3 on PTH suppression is preceded by a 25-hydroxylation to 1,25(OH) 2 D 3 .

The observed pharmacokinetic differences between 1,25(OH) 2 D 3 and 1 α (OH)D 3 might explain the different biological responses between single-dose studies and long-term continuous studies. Consequently, the pharmacokinetic differences between 1,25(OH) 2 D 3 and 1 α (OH)D 3 in uremic patients and healthy volunteers were analysed in one study [6]. A 57% lower metabolic clearance rate of 1,25(OH) 2 D 3 was found in uremic patients as compared to healthy volunteers in accordance with others [274, 326]. This difference in the clearance of 1,25(OH) 2 D 3 between healthy volunteers and uremic patients may be due to the lack of kidney function in the uremic patients affecting renal metabolism and the elimination of 1,25(OH) 2 D 3 , or due to the presence of toxins in uremic plasma suppressing the degradation of 1,25(OH) 2 D 3 [326]. The bioavailability as 1,25(OH) 2 D 3 after intravenous administration of 1 α (OH)D 3 was only 40% when compared to 1,25(OH) 2 D 3 in both normal and uremic subjects in accordance with results obtained after intravenous administration of 1 α (OH)D 2 [327]. Despite the finding that 4 μ g of both 1,25(OH) 2 D 3 and 1 α (OH)D 3 [6] seemed to have the same suppressive effect on plasma PTH, we observed a concomitant small, yet significant increase in plasma Ca 2+ in the uremic patients after administration of 1,25(OH) 2 D 3 that was not seen after administration of 1 α (OH)D 3 .This means that no exact conclusions can be drawn on the therapeutic equivalence between the two drugs. However, the clinical experience is that 1 α (OH)D 3 and 1,25(OH) 2 D 3 are both effective drugs in the treatment of secondary hyperparathyroidism in uremic patients [2, 4, 184, 193].