Role of biochemical markers in the management of osteoporosis

Best Practice & Research Clinical Rheumatology Vol. 15, No. 3, pp. 385±400, 2001

doi:10.1053/berh.2001.0156, available online at http://www.idealibrary.com on

4 Role of biochemical markers in the management of osteoporosis Peter R. Ebeling

MD, FRACP

Associate Professor Department of Diabetes and Endocrinology, The Royal Melbourne Hospital, Parkville, Victoria 3050, Australia

Kristina AÊkesson

MD, PhD

Associate Professor and Consultant Department of Orthopedics, MalmoÈ University Hospital, S-205, MalmoÈ, Sweden

Several serum and urine biochemical markers of bone resorption and formation have been developed. Biochemical bone markers have been used as intermediate end-points in all major studies of anti-osteoporotic therapies. Bone resorption markers, in particular, may add an independent, predictive value to the assessment of bone loss and fracture risk. There are also potential advantages in monitoring anti-osteoporotic treatment in the short-term in addition to bone densitometry, to rapidly identify non-responders to therapy, or non-compliance. Despite these recent advances, until now bone markers have simply been very useful research tools, with their clinical utility being limited by intra-individual and diurnal variability. However, the probability of the true bone mineral density response to hormone replacement therapy for the individual patient may be predicted using algorithms based on a spectrum of cut-o€ bone marker levels with varying false positive and negative rates. Thus, the transition of biochemical bone markers into everyday clinical practice may be rapidly approaching. Key words: biochemical bone markers; fracture risk; therapy; bone density; non-responders.

INTRODUCTION Osteoporosis is one of the most common and serious diseases of the musculo-skeletal system. The lifetime fracture risk of a female aged 60 years is 56%, compared with 29% in a man of the same age. As a result of fragility fractures, the individual experiences pain, decreased mobility and a reduced quality of life. The onset of the disease is silent and insidious with the diagnosis often only being made after there is irreversible damage resulting from a fragility fracture. Thus, it would be highly desirable to develop sensitive and reliable methods for the early diagnosis and monitoring of treatment for this debilitating disease. The e€ective monitoring of treatment ecacy might also aid in long-term compliance with anti-osteoporotic therapy. 1521±6942/01/030385‡16 $35.00/00

c 2001 Harcourt Publishers Ltd. *

386 P. R. Ebeling and K. AÊkesson

Advances in bone biochemistry and physiology have provided important insights into the pathogenesis of osteoporosis. Both post-menopausal and age-related bone loss are due to increased bone turnover with a relative imbalance between bone formation and bone resorption, favouring the latter. The goal of the majority of the currently available treatments for osteoporosis is to normalize the increased bone turnover and to stabilize or increase bone mineral density (BMD), reducing the risk of subsequent fragility fractures. Early biochemical changes in osteoporosis include the increased release of bone collagen degradation products by the action of osteoclasts and also the increased production of molecules released from either osteoblasts or by the bone matrix. The rate of release of these products of bone cells or their actions indicate abnormalities of bone and mineral metabolism. Biochemical assays for these molecules may be used to detect increased bone turnover and to monitor treatment ecacy. A number of biochemical assays readily detect these molecules released from the bone matrix and bone collagen degradation in both serum and urine. Several of these biochemical markers of bone turnover have the potential to serve as both aids in the decision to treat patients with low bone density and as e€ective and early indicators of the response to anti-osteoporotic therapy. In this chapter, we will discuss the potential utility of biochemical markers of bone turnover in the management of patients with osteoporosis. To facilitate the understanding of the rationale and the value of biochemical bone turnover markers in osteoporosis management, a summary of the biochemistry of bone turnover and the markers currently available will be included.

BIOCHEMICAL CONSIDERATIONS IN OSTEOPOROSIS Previously, the only way to accurately assess bone turnover was by bone histomorphometry following double-labelling using tetracycline.1,2 However, this technique is invasive and expensive and may not re¯ect bone turnover rates at other skeletal sites. Nevertheless, it does provide information regarding static and dynamic rates of bone formation and bone resorption, and the percentage of the bone undergoing active bone formation and bone resorption. Using this technique, the overall normal skeletal bone turnover rate is relatively low and estimated to be about 10% per year, on average, with a higher rate in trabecular bone (15±20% per year) and a lower rate in cortical bone (3±5% per year).3 Bone tissue is very vascular and susceptible to systemic regulation by factors that include hormones. Bone remodelling is activated by factors such as: parathyroid hormone (PTH), 1,25-dihydroxy vitamin D, growth factors, cytokines, mechanical loading and bone micro-trauma. Bone resorption relies on osteoclasts creating a suitable environment by locally increasing the acidity of the bone surface as well as secreting proteinases. Calcium and phosphorus are liberated during bone resorption, and osteoclasts also digest bone type I collagen (Figure 1), which is released as a series of peptides that can be measured in the serum and urine to quantify the bone resorption rate of the entire skeleton. Similarly, new bone formation leads to the production of collagen cleavage products and other bone matrix proteins that are elaborated during di€erent stages of osteoblast di€erentiation (Figure 2).

Biochemical markers and osteoporosis 387

NTx

ICTP Cat K

a2(l) JYDGKGVG

Cat K

CTx Cat K

GPP*SAGFDFS FLPQPPQ EKAHDGGR a1(l)

N

C

Figure 1. The cathespin K (Cat K) proteolytic sites and bone marker epitopes at the N- and C-termini of type I bone collagen. NTx, N-telopeptide cross-links; CTx, C-telopeptide cross-links; ICTP, pyridinoline cross-linked carboxyterminal telopeptide of type I collagen. Reproduced from Nishi et al61 with permission.

Peptide chains

C-terminal propeptide

N-terminal propeptide Procollagen molecule

Osteoblast cell membrane

Collagen molecule

Collagen fibril Figure 2. Schematic illustration of collagen formation.

METABOLISM OF TYPE I COLLAGEN IN BONE Type I collagen is formed in bone from the combination of two a-1 and one a-2 collagen polypeptides containing hydroxylated proline and lysine residues. This structure

388 P. R. Ebeling and K. AÊkesson

is known as procollagen. As procollagen is secreted from the osteoblast, the aminoterminal and carboxy-terminal regions are cleaved. These propeptides are released into the extracellular ¯uid although a proportion of the amino-terminal propeptide is also incorporated into bone. Type I collagen is helical and the non-helical domains at the amino- and carboxy-terminii are known as the N-telopeptide and the C-telopeptide regions. The side chains of three hydroxylysine residues from di€erent type I collagen molecules condense to form a pyridinium ring so that pyridinium cross-links are formed connecting three di€erent collagen molecules and stabilizing the structure of type I collagen. Pyridinoline cross-links result from the combination of three hydroxylysine side chains (hydroxylysylpyridinoline)4 while deoxypyridinoline cross-links result from the combination of two hydroxylysine side chains with one lysine side chain (lysylpyridinoline).5 The pyridinium cross-link in the N-telopeptide region (NTx) joins a-1 type I collagen polypeptides to a-2 type I collagen polypeptides.6 By contrast pyridinium cross-links in other tissues join a-1 type I collagen polypeptides to ÿ1 type I collagen polypeptides (e.g. C-telopeptide). This makes the N-telopeptide relatively bone-speci®c. In addition, two-thirds of deoxypyridinoline cross-links in bone type I collagen are N-telopeptide cross-links and only one-third are C-telopeptide cross-links. Assuming that bone turnover under normal circumstances remains at a fairly steady state, products released during bone resorption could be used as markers for the catabolic events. Similarly during the anabolic phase osteoblasts are actively producing collagen and other proteins to be incorporated in the newly synthesized matrix. Surplus products or fragments released into serum may then be used as markers of bone formation. In disease states, alterations of marker levels should preferably provide insight into the metabolic disturbances in a particular bone disorder. In the following section, each one of the potentially useful markers for bone will be separately reviewed (Table 1) and the current utility of these bone markers in osteoporosis management will be discussed. BONE FORMATION MARKERS The predominant product of osteoblasts is type I collagen, which comprises 95% of the extracellular non-mineral bone matrix. Other proteins such as osteopontin, osteonectin Table 1. Potential biochemical markers of bone turnover in osteoporosis. Bone formation

Bone resorption

Serum Osteocalcin Bone speci®c alkaline phosphatase Procollagen type I C-/N-extension peptide (PICP, PINP)

Serum Pyridinoline cross-linking telopeptides (C- and N-telopeptides, CTx, NTx, ICTP) Free pyridinoline and deoxypyridinoline Tartrate-resistant acid phosphatase Bone sialoprotein Urine Pyridinoline cross-links: Pyridinoline Deoxypyridinoline Pyridinoline cross-linking telopeptides (C- and N-telopeptides, CTx, NTx)

Biochemical markers and osteoporosis 389

and osteocalcin are also secreted to form the osteoid or organic substrate in which mineralization occurs. Osteoblasts stain positively for the enzyme alkaline phosphatase, which is attached to their cell membranes. This alkaline phosphatase is functionally similar but antigenically di€erent to hepatic, intestinal or placental alkaline phosphatases.7

Osteocalcin Osteocalcin is one of the most abundant non-collagenous proteins of bone matrix, synthesized by osteoblasts.8 The serum concentration of osteocalcin re¯ects the rate of osteoblast synthesis of osteocalcin. However, only approximately 50% of newly synthesized osteocalcin is released into the circulation while the remaining 50% is incorporated into hydroxyapatite. Osteocalcin is regarded as bone-speci®c, since the small amount present in dentine is negligible, but its precise function remains unknown. Osteocalcin contains three g-carboxylated glutamic acid residues and the degree of carboxylation (i.e. number of residues carboxylated) seems to in¯uence mineralization. Osteocalcin in serum is susceptible to rapid enzymatic cleavage and, subsequently, both intact and fragmented osteocalcin, including a large N-mid-molecular fragment, are detectable in serum at any given time. In addition, this degradation occurs most rapidly at room temperature, which means that specimens should be frozen as soon as possible after collection. For these reasons, assays providing a combined measure of intact and Nmid-fragment are more robust and may improve the sensitivity of the assay.9

Serum bone alkaline phosphatase The skeletal isoform of alkaline phosphatase (ALP) is a membrane-bound protein with enzymatic activity. Despite this, its release mechanism from osteoblasts remains to be elucidated. However, ALP plays an important role in bone mineralization. Serum ALP exists in several isomeric forms depending on its tissue origin. In children and adolescents up to 80% of ALP measurable in serum is derived from bone, mirroring bone growth. In adults, only about 50% is bone-derived, while the remainder mostly emanates from liver tissue. Recently developed immunoradiometric or enzyme immunoassays have facilitated the determination of serum levels of bone ALP.10 Despite a residual and signi®cant cross-reactivity with the liver iso-enzyme, bone ALP provides a higher speci®city for bone than the other bone formation markers.

Carboxy-terminal propeptide of type I procollagen During the synthesis of type I collagen by the osteoblasts, the amino- (N-) and carboxy- (C-) terminal extension propeptides are cleaved. The intact terminal fragment is released into the circulation in its entirety prior to extracellular ®bril formation.11 The serum concentration of the carboxy-terminal propeptide of type I collagen (PICP) re¯ects changes in the synthesis of new collagen, both by osteoblasts in bone and by ®broblasts in other connective tissues.12 Theoretically, utilization of the propeptides as markers of bone formation is appealing, since their synthesis should be directly proportional to the amount of type I collagen produced. However, PICP appears to lack sensitivity for identifying subtle alterations in bone turnover.13

390 P. R. Ebeling and K. AÊkesson

Amino-terminal propeptide of type I procollagen The serum concentration of the amino-terminal propeptide of type I collagen (PINP) also re¯ects changes in the synthesis of new collagen, both by osteoblasts in bone and by ®broblasts in other connective tissues.14,15 However, unlike PICP, a proportion is also incorporated into bone as non-dialysable hydroxyproline and, thus, a component of the measured fragments might represent bone resorption. Nevertheless, PINP appears to be a more dynamic and sensitive marker of changes in bone formation than PICP, but there has been less clinical evaluation of PINP.

BONE RESORPTION MARKERS Bone resorption is initiated by osteoclasts, which contain acid phosphatase.1 Although acid phosphatase activity is present in other tissues such as the prostate gland and blood cells, the type Vb enzyme is speci®c to osteoclasts. Osteoclasts attach to the bone surface and secrete acid and hydrolytic enzymes that resorb bone, releasing bone minerals and collagen fragments. The N-telopeptide epitope is preferentially liberated from type I collagen in bone by osteoclastic hydrolysis with cathepsin K.16 Cathepsin K is located in osteoclast intracellular vacuoles and in the subosteoclastic space. Within type I collagen telopeptide domains there are three proteolytic sites for cathepsin K. In addition, both serum NTx and C-telopeptide cross-link (CTx) concentrations are reduced in pycnodyostosis, a disorder of reduced cathepsin K activity, while serum pyridinoline cross-linked carboxyterminal telopeptide of type I collagen (ICTP) levels and urine free deoxypyridinoline (Dpd) excretion are elevated in this disorder. N-telopeptide cross-links may be degraded further in the liver and kidney in particular, where free Dpd may be generated. Serum NTx concentrations are elevated in chronic renal failure. Some of the collagen is completely digested by osteoclasts to its smallest units, free pyridinoline and Dpd residues, which are excreted in the urine. The majority however, appears to be incompletely digested, resulting in the formation of pyridinium cross-links bound to fragments of the NTx a-1 and a-2 polypeptides; peptide-bound cross-links are also excreted in the urine.6,17

Hydroxypyridinium cross-links Urinary excretion of the hydroxypyridinium cross-links of type I collagen re¯ects osteoclastic bone resorption, and not dietary calcium or collagen intake. Thus, it is a more precise indicator of bone resorption than urinary calcium or hydroxyproline excretion. The distribution of hydroxypyridinium cross-links varies depending on tissue type; pyridinoline is primarily present in cartilage, but also in bone collagen, while Dpd is predominantly found in bone and dentine.5,18 Theoretically, Dpd is the preferred marker of bone resorption and it correlates positively with other indices of bone resorption, and inversely with bone mass in osteoporosis.19 Although Dpd is also found in relatively high concentrations in vascular tissue (such as the aorta) and in skeletal muscle, the metabolic turnover rate of these tissues is far lower than that of bone. As a result these tissues contribute little to the circulating pool of Dpd. Free Dpd can be measured in the urine or serum.

Biochemical markers and osteoporosis 391

Peptide-bound N-telopeptide and C-telopeptide cross-link assays Recently, the cross-linking telopeptide regions, C-telopeptide to helix and Ntelopeptide to helix, have emerged as potential markers of bone resorption, measurable either in serum or urine by immunoassays. These particular markers have been widely employed in studies of osteoporosis, including therapeutic trials, and are currently considered to be the preferred markers of bone resorption because of their convenience and dynamic response to therapeutic intervention. Peptide-bound NTx can be measured in the urine by an immunoassay6 and CTx can be measured in serum by two immunoassays.20,21 Only the b-isomer of the CTx is measured in the serum assays, while both the a- and b-isomers of CTx are measured in the urine assay. The serum CTx assays are a€ected by the non-fasting state, whereas the urine assay is not. The CTx immunoassay does not measure cross-links, but measures a sequence (octapeptide) of the C-telopeptide region of the a-1 chain of type I collagen in the urine.21 The NTx immunoassay has also been automated and modi®ed to measure the same epitope in serum. Bone sialoprotein Bone sialoprotein (BSP) is a phosphorylated glycoprotein, containing sialic acid, which has cell adhesive properties. Bone sialoprotein is speci®c for bone and it is enriched in the immediate subchondral bone. Serum BSP is increased in states of high bone turnover and correlates well with other bone resorption markers.22 Tartrate-resistant acid phosphatase Osteoclasts produce acid to dissolve the mineral phase of bone and various enzymes to degrade the matrix. Tartrate-resistant acid phosphatase (TRAP) is produced by osteoclasts during bone resorption and the type Vb (V ˆ ®re) isoform has been identi®ed as a potential marker of osteoclast activity through recently developed assays using two-site monoclonal antibodies.23 It may provide di€erent information from the pyridium cross-links that measure osteoclast action, rather than cellular activity of osteoclasts. VALIDATION AND LIMITATIONS IN THE USE OF BIOCHEMICAL BONE MARKERS Three major criteria need to be satis®ed before a biochemical test can be established as a biological marker of bone turnover. Firstly, the marker must change in parallel with changes in bone turnover measured by bone histomorphometry or calcium kinetics.2 Secondly, the serum concentration or urinary excretion of the substance must be high in conditions characterized by high bone turnover, such as Paget's disease of bone. Finally, the serum concentration or urinary excretion of the substance must be low in conditions characterized by low bone turnover, e.g. after the administration of antiresorptive drugs. In the early phase of bone marker development, high analytical variability was a major problem. Improved analytical techniques, from high pressure liquid chromatography (HPLC) to fully automated immunoassays and the addition of serum versus urine assays, have substantially diminished these problems. However, a continuing major concern

392 P. R. Ebeling and K. AÊkesson

limiting their wider clinical use relates instead to the relatively high intra-individual variability. This biological variability, together with the residual analytical variability, limits the interpretation of results for the individual patient both at baseline and during monitoring of treatment. The intra-individual variation is related to factors that include age, with high values during childhood and growth and high values again after the menopause in women. The larger skeleton of men may increase bone marker variability. Biological variability is also increased in women with post-menopausal osteoporosis. Fractures also confound the interpretation of biochemical bone turnover marker results. Hip and other peripheral fractures cause a rapid increase in bone turnover that can last for up to a year.24,25 The impact of vertebral fractures on bone turnover is uncertain, but could be considerable given the generally higher turnover of trabecular bone. Physical inactivity or immobility because of fracture, illness or for other reasons also lead to rapid increases in bone resorption as seen in studies of bed rest.26 In addition, bone turnover is subject to a circadian variability, with peak values in the early morning and with a nadir in the early afternoon and evening.27 Moreover, the degree of variation is not uniform for all markers, but is most pronounced for serum bone resorption markers, reaching 20±50% depending on time of day. To this the intraindividual day-to-day variability should be added. Seasonal changes with increased bone resorption have also been described, with a corresponding pattern of PTH with an inverse curve for 25-hydroxyvitamin D.28,29 The seasonal interference may be greater in Northern latitudes with fewer hours of daylight and sun exposure during winter. In order to minimize these ¯uctuations, the timing of sample collection is critical and should be similar for each patient and at each time point. Urine samples for measurements of bone turnover markers can be collected either as random, non-fasting, 2-h post-voiding or as 24-h urine collections. The results in 2-h early morning postvoiding samples correlate well with those in 24-h samples and rates of bone loss correlate with 2-h urinary Dpd values better than with values from 24-h urine collections.30 Short-term variability of 2-h post-voiding cross-link excretion (9±13%) is less than that of 24-h collections (26%), but di€erences in long-term variability may be lower. Similarly, awareness of collection times needs to be taken regarding serum assays since, for example, both serum NTx and CTx may be more sensitive to diurnal variation than urine assays. When ordering these tests through a local laboratory, it is important that samples are collected according to their instructions so that results can be compared with the normal reference ranges established by that laboratory. Exposure of the sample to ultraviolet light should be avoided. Standardized curves adjusted for age and sex should also be developed to facilitate and improve the interpretation of the result by the manufacturer. The treating physician needs to be aware of the limitations of bone marker assays to be able to accurately interpret the results. BONE MARKERS IN THE MANAGEMENT OF OSTEOPOROSIS Biochemical bone turnover markers do not replace dual energy absorptiometry (DEXA) for the diagnosis of osteoporosis. However, bone markers may give some additional indication about the future risk for bone loss and fractures. They may also be useful in monitoring the ecacy of anti-resorptive therapy in patients with osteoporosis. Biochemical indicators of disease are available in several disease areas. The most successful markers identify patients at risk or in the early stages of disease thereby enabling the ®nal outcome of the condition to be altered by therapeutic intervention.

Biochemical markers and osteoporosis 393

Enhancement of prediction of future risk of bone loss Although advancing age is the strongest predictor for low bone mass and fracture, a patient's current bone mineral density (BMD) is an important predictor of fracture risk.31 However, a single measurement indicates only current BMD, not the anticipated rate of bone loss. Thus, patients with a BMD value corresponding to osteopenia may have a greater risk of osteoporosis and fracture if they are losing bone at a rapid rate. Recent studies have demonstrated that markers of bone turnover may be useful in predicting rates of future bone loss and may, therefore, provide independent information about fracture risk beyond that available from BMD measurements alone. However, the diagnostic utility of a single bone turnover measurement is limited because individuals with low levels of bone turnover have rates of bone loss that range from 0±10%/year32; the test, therefore, has a low speci®city. Nevertheless, a person with a high value for a bone turnover marker is generally at greater risk of bone loss than a person with a low value. In most studies there is a highly signi®cant correlation between bone turnover markers and subsequent rates of bone loss.33±37 The rate of bone turnover explains up to 50% of the variance in bone mass in 80 year old women.19 After the age of 30, women with high bone turnover have bone mass values at the hip, spine or forearm that are 6±11% lower that those of women with a low bone turnover as measured using osteocalcin and CTx.38 In post-menopausal women the di€erences in bone mass between those with high versus low turnover were larger (8±14%). In women aged 67 years or more, an increased value for all bone resorption markers was associated with an increased rate of bone loss from the hip after 4 years.39 In women with a yearly femoral neck bone loss of 85% above average, 56±73% of the resorption marker levels were above the median, however signi®cant overlap occurred. Bone markers are not usually adjusted for age, sex, height or weight, thus the premenopausal mean is often used as a reference value. Post-menopausal women with a bone turnover rate that is 2 SD above the pre-menopausal mean lose 2±6 times more bone in the distal forearm over a 4-year period, depending on the marker chosen.40 Bone turnover increases at the time of menopause and it has been suggested that some women are more susceptible to metabolic changes related to oestrogen de®ciency than others. However, it remains uncertain whether bone markers measured early in the menopause predict post-menopausal bone loss in the individual patient.37,41 Taken together, it would seem reasonable to measure BMD at menopause if preventive therapy is being contemplated for women at risk for osteoporosis. If the BMD is in the osteopenic range, the ®nding of a urine or serum bone turnover marker that is elevated above the upper limit of normal for pre-menopausal women would add further impetus to the recommendation for preventative therapy. Bone markers in fracture prediction Identi®cation of prospective fracture patients is a key goal in osteoporosis. Ideally the bone marker should predict all types of fracture, or at least major fractures such as hip fracture. Retrospective studies have suggested that there is a di€erence in bone marker levels between fracture patients and controls. However, because the fracture preceded the bone marker measurement, it is uncertain whether the di€erence in turnover was caused by the fracture or existed prior to the fracture.42 Prospective population-based studies using a nested case-control design, despite a low number of fractures even in large studies, have shown consistent relationships between

394 P. R. Ebeling and K. AÊkesson

resorption markers, alone or in combination with BMD, and fracture risk. In the EPIDOS study, urine free Dpd and CTx independently predicted hip fracture (odds ratio ˆ 1.9±2.5).43 Their predictive value increased when combined with low bone mass to a more than four fold fracture risk in these women (odds ratio ˆ 4.1±5.2). In the Rotterdam study, high urinary free Dpd was associated with an increased risk of hip fracture over 4 years. However, the number of women with fractures was low (n ˆ 17) and, in addition, the women were generally more disabled and the marker levels were a€ected by the subsequent immobility.44 Women in early menopause with high bone turnover had a doubled risk of sustaining vertebral or peripheral fractures during 15 years of follow-up, compared with women having normal early menopausal bone turnover.45 The degree of vitamin K-dependent g-carboxylation of osteocalcin may a€ect hydroxyapatite binding and bone mineralization. Because the degree of g-carboxylation is reduced with ageing46, impaired g-carboxylation may increase the risk of low BMD and fracture. High serum levels of under-carboxylated osteocalcin are associated with an increased risk of hip fracture, independent of BMD, while adding BMD enhanced its predictive ability.47,48 In a community-based study of women and men over the age of 70, low levels of carboxylated osteocalcin or a low carboxylated to total osteocalcin ratio conferred a high risk for any type of fracture, including hip fracture (relative risk ˆ 5.32 (3.26±8.68)). The relative risk was greatest in persons over 80 years and the increased risk persisted for 3 years out of a 5 year follow-up period.49 Increased levels of bone resorption markers and a decreased level of carboxylated osteocalcin have the potential to allow for the estimation of fracture risk, particularly hip fracture risk in the elderly, for at least 3±4 years. In combination with BMD measurements the fracture risk prediction is further strengthened. The fracture risk predictions for other peripheral fractures or vertebral fractures are similar. Therefore, bone markers, particularly bone resorption markers, may have a role in the clinical evaluation of the individual woman with osteoporosis together with other risk factors for fracture, in particular BMD. Bone markers and monitoring of treatment Because the skeletal bone turnover rate is low, with less than a tenth of the skeleton exchanged per year, the impact of alterations in bone turnover is only evident after 1±2 years, at the earliest, using DEXA to measure BMD. Given that the precision of spine DEXA is between 1±2%, the smallest change in BMD required for a p 5 0.05 level of signi®cance would be 2.77±5.54%. In the individual patient, BMD measurement is a surrogate for assessing the anti-fracture ecacy of currently used anti-osteoporotic drugs. The majority of these are anti-resorptive agents of varying potency, with bisphosphonates and oestrogen being the most potent. For other treatments, such as selective oestrogen receptor modulating drugs (SERMs) and intra-nasal calcitonin, the change in bone mass may be borderline for detection in the individual within this time frame, despite a demonstrated fracture-sparing e€ect in clinical trials. In addition there is growing evidence that the change in BMD at 2 years may more truly re¯ect the therapeutic e€ect of a drug because of the trend to regress to the mean at year one. Hence, there is a need for a more rapid tool to evaluate the ecacy of treatment in the individual patient. There are non-responders to each treatment regimen and the rate is in the order of 20% and 15% for hormone replacement therapy and alendronate, respectively. In clinical practice there is also a need to detect non-compliance, a common occurrence

Biochemical markers and osteoporosis 395

with these drugs. There is now the option for non-responders to treatments of individualizing the anti-osteoporotic therapy on the basis of the therapeutic response. An alternative to using BMD measurements only for assessing early therapeutic ecacy would be to also measure a bone marker at baseline and after 6 months of treatment. In all clinical trials of anti-osteoporotic drugs, bone markers have been used as intermediate end-points. Bone markers indicate early treatment e€ects on bone metabolism. A decrease in bone resorption markers occurs within 6 weeks to 3 months after initiation of antiresorptive therapy, while the nadir in bone formation markers is delayed until 6 months. After this time, bone turnover is stabilized at a lower level. E€ects on bone resorption markers have been of greater interest, since the therapeutic agents act on osteoclast number and activity. The e€ect on bone formation is secondary and related to the coupling e€ect, which also explains the delayed response. The absolute percentage change from baseline is variable depending on the speci®c properties of each individual biochemical bone marker. Women with the highest bone turnover appear to derive the greatest bene®t from anti-resorptive therapy with alendronate50, oestrogen36, calcium36, or calcitonin51, while by comparison, women with the baseline urinary N-telopeptide values in the lowest quartile are no less likely to lose bone during administration of oestrogen than of placebo.36 Alendronate treatment induces a marked, dose-related 50±65% decrease in NTx after 6±8 weeks of therapy.50,52,53 The decrease in NTx level at 6 months correlates with the change in BMD at 2 years at all sites. In addition the subgroup of women with the largest percentage initial decrease in bone resorption, showed the largest gain in bone mass after 2.5 years of treatment. Further studies have shown similar correlations between short-term decreases in other bone markers following alendronate treatment and the change in bone mass at 2 years, as well as for NTx.54,55 By comparison, Dpd decreases by 10±30% during alendronate treatment. Hormone replacement therapy induces similar changes in the levels of bone turnover markers.32 After 1 year of treatment, the decrease varies between 25±60%, depending on which marker is used, the largest changes being for CTx and NTx.56 The short-term change in bone markers is correlated with the long-term (1±2 year) change in bone mass.32,57 The corresponding changes seen after 6±9 months of raloxifene treatment are a 20±30% and a 30±40% decrease in osteocalcin and CTx, respectively.58 By comparison, calcium supplementation alone decreases bone turnover by 5±15%. The identi®cation of responders versus non-responders to treatment remains problematic. One approach is to calculate the least signi®cant change (LSC) based on the known marker variability. The LSC for bone resorption markers was 46±72% and 30±62% for formation markers in the oestrogen/progestin intervention trial, indicating that large changes in markers are necessary and that predictive values in study populations may include signi®cant overlap in individual values.56 However it should be noted that when expressed as the expected change over long-term variability, the predictive power of BMD and biochemical bone markers is similar. Secondly, bone markers are also independent predictors of fracture, as noted above, and this may relate to deterioration in bone microarchitecture during phases of high bone resorption. Finally, it may not be necessary to reach a level of signi®cance of P 5 0.05 in the individual patient to make clinical decisions. In this regard, two recent large studies of hormone replacement therapy over 2±3 years have assessed monitoring individual responses using serum osteocalcin, bone alkaline phosphatase and CTx, and urine CTx.59,60 As early as after only 2 weeks of treatment, decreases in bone resorption markers, in particular, correlated with the

396 P. R. Ebeling and K. AÊkesson

1100

Sensitivity for 90% specificity

1000 900 68.4%

Actual value

800 700 600 500 400 300

76.4%

200 58.5%

100 0 –100

0

100

200

% change Figure 3. The ability of urine C-telopeptide cross-link levels measured at 6 months to predict bone mineral density (BMD) responses to hormone replacement therapy (HRT) at 2 years. Produced from Delmas et al59 with permission. , non-responder; , responder.

BMD response at 2 or 3 years and this correlation was strongest at 6 months (Figure 3). In these studies, receiver operating characteristic (ROC) analysis allowed a spectrum of cut-o€ levels with varying false positive and negative rates. For the individual patient it would thus be possible to predict the probability of the true BMD response by consulting plots or a computer programme. These studies cannot be generalized to other bone markers or other treatment regimens, however. Based on available evidence, it seems likely that bone markers, especially bone resorption markers, can predict the BMD response to treatment for 2±3 years reasonably well. Regarding the identi®cation of non-responders, it is likely that an estimate of the probability of a response to hormone replacement therapy will soon be possible for some markers and that other markers and treatment modalities will soon be assessed in a similar manner. Further improvements in analytical techniques, including automation, together with careful measures for minimizing sample handling di€erences, may reduce marker variability and optimize the predictive power of these tests. Using bone densitometry it would seem prudent to wait 2 years before changing therapy. It is hoped that by using biochemical bone markers and, in particular, bone resorption markers, that this decision can be made before further bone is lost. Whether monitoring using biochemical bone markers improves long-term compliance with anti-osteoporotic therapy in clinical practice remains to be tested by appropriate prospective studies.

SUMMARY An understanding of the underlying cellular processes of bone turnover in osteoporosis and other bone metabolic diseases has led to the development of biochemical markers

Biochemical markers and osteoporosis 397

Practice points . bone markers, in particular high levels of resorption markers, may indicate progressive bone loss and in addition to bone mass measurement in postmenopausal women, may strengthen the indication for treatment . bone markers, in particular high levels of resorption markers and undercarboxylated osteocalcin, in combination with bone mass measurement, may enhance fracture prediction . changes in bone marker levels may indicate a response to anti-resorptive treatment and BMD change at 2±3 years but, because of limitations related to intra-individual variation, bone markers are not yet an established method for evaluating treatment response, ecacy or compliance in the individual patient

Research agenda . development of tools to enhance the utility of markers in the management of the individual patient, including normality curves for each marker related to age and BMD to facilitate interpretation of values . long-term evaluation (up to 10 years) of the association between bone markers, BMD and fracture, to determine the duration of the predictive ability of bone markers for bone loss and fracture . prospective studies aimed at determining treatment response by bone markers as an intermediate end-point for various treatment modalities in order to assign cut-o€ levels that would be usable in individual patients . bone markers are at present mainly evaluated in post-menopausal women. Future research needs to evaluate bone markers in relation to bone loss and fracture in men

of bone resorption and formation that can be measured in the serum and urine. Previously, bone turnover could only be assessed by invasive and expensive bone histomorphometry or nuclear techniques. Regarding osteoporosis, we have reached a stage where there are strong indications that biochemical bone markers, particularly bone resorption markers, may add an independent, predictive value to the assessment of bone loss and fracture risk. There are also potential advantages for monitoring antiosteoporotic treatment in addition to bone mass measurements, for example to identify non-responders or non-compliance. Despite the rapid advances in the use of biochemical markers in the assessment of disorders of bone metabolism, including osteoporosis, bone markers are currently mainly used as very useful research tools. However, their transition into everyday clinical practice may be fast approaching. REFERENCES 1. Calvo MS, Eyre DR & Gundberg CM. Molecular basis and clinical application of biological markers of bone turnover. Endocrinology Reviews 1996; 17: 333±368. 2. Par®tt AM, Drezner MK, Glorieux FH et al. Bone histomorphometry: standardization of nomenclature, symbols, and units. Journal of Bone and Mineral Research 1987; 2: 595±610.

398 P. R. Ebeling and K. AÊkesson 3. Par®tt AM. The physiological and clinical signi®cance of bone histomorphometric data. In Recker R (ed.) Bone Histomorphometry. Techniques and Interpretations, pp 143±223. Boca Raton: CRC Press, 1993. 4. Horgan DJ, King NL, Kurth LB & Kuypers R. Collagen crosslinks and their relationship to the thermal properties of calf tendons. Archives of Biochemistry and Biophysics 1990; 281: 21±26. 5. Eyre DR, Koob TJ & Van Ness KP. Quantitation of hydroxypyridinium crosslinks in collagen by highperformance liquid chromatography. Annals of Biochemistry 1984; 137: 380±388. 6. Hanson DA, Weis MA, Bollen AM et al. A speci®c immunoassay for monitoring human bone resorption: quantitation of type I collagen cross-linked N-telopeptides in urine. Journal of Bone and Mineral Research 1992; 7: 1251±1258. 7. Gomez B Jr, Ardakani S, Ju J et al. Monoclonal antibody assay for measuring bone-speci®c alkaline phosphatase activity in serum. Clinical Chemistry 1995; 41: 1560±1566. 8. Price PA. Vitamin K-dependent formation of bone Gla protein and its function. Vitamins and Hormones 1985; 42: 65±107. 9. Delmas PD. Biochemical markers of bone turnover. Acta Orthopaedica Scandinavica 1995; 66: 176±182. 10. Garnero P & Delmas PD. Assessment of the serum levels of bone alkaline phosphatase with a new immunoradiometric assay in patients with metabolic bone disease. Journal of Clinical Endocrinology and Metabolism 1993; 77: 1046±1053. 11. Prockop DJ, Kivirikko KI, Tuderman L & Guzman NA. The biosynthesis of collagen and its disorders. New England Journal of Medicine 1979; 310: 13±23. 12. Melkko J, Niemi S, Risteli L & Risteli J. Radioimmunoassay of the carboxyterminal propeptide of human type I procollagen. Clinical Chemistry 1990; 36: 1328±1332. 13. Charles P, Mosekilde L, Risteli J et al. Assessment of bone remodeling using biochemical indicators of type I collagen synthesis and degradation: relation to calcium kinetics. Bone and Mineral 1994; 24: 81±94. 14. Ebeling PR, Peterson JM & Riggs BL. Utility of type I procollagen propeptide assays for assessing abnormalities in metabolic bone diseases. Journal of Bone and Mineral Research 1992; 7: 1243±1250. 15. Melkko J, Kauppila S, Niemi S et al. Immunoassay for intact amino-terminal propeptide of human type I procollagen. Clinical Chemistry 1996; 42: 947±954. 16. Atley LM, Mort JS, Lalumiere M & Eyre DR. Proteolysis of human bone collagen by cathepsin K: characterization of the cleavage sites generating by cross-linked N-telopeptide neoepitope. Bone 2000; 26: 241±247. 17. Garnero P, Gineyts E, Arbault P et al. Di€erent e€ects of bisphosphonate and estrogen therapy on free and peptide-bound bone cross-links excretion. Journal of Bone and Mineral Research 1995; 10: 641±649. 18. Eyre DR. The speci®city of collagen cross-links as markers of bone and connective tissue degradation. Acta Orthopaedica Scandinavica 1995; 66: 166±170. 19. Garnero P, Sornay-Rendu E, Chapuy M-C & Delmas PD. Increased bone turnover in late postmenopausal women is a major determinant of osteoporosis. Journal of Bone and Mineral Research 1996; 11: 337±349. 20. Risteli J, Elomaa I, Niemi S et al. Radioimmunoassay for the pyridinoline cross-linked carboxy-terminal telopeptide of type I collagen: a new serum marker of bone collagen degradation. Clinical Chemistry 1993; 39: 635±640. 21. Garnero P, Gineyts E, Riou JP & Delmas PD. Assessment of bone resorption with a new marker of collagen degradation in patients with metabolic bone disease. Journal of Clinical Endocrinology and Metabolism 1994; 79: 780±785. 22. Seibel MJ, Woitge HW, Pecherstorfer M et al. Serum immunoreactive bone sialoprotein as a new marker of bone turnover in metabolic and malignant bone disease. Journal of Clinical Endocrinology and Metabolism 1996; 81: 3289±3294. 23. Halleen JM, Karp M, Viloma S et al. Two-site immunoassays for osteoclastic tartrate-resistant acid phosphatase based on characterization of six monoclonal antibodies. Journal of Bone and Mineral Research 1999; 14: 464±469. 24. AÊkesson K, Vergnaud P, Delmas PD & Obrant KJ. Serum osteocalcin during fracture healing in elderly women with hip fracture. Bone 1995; 16: 427±430. 25. Ingle BM, Hay SM, Bottjer HM & Eastell R. Changes in bone mass and bone turnover following distal forearm fracture. Osteoporosis International 1999; 10: 399±407. 26. Zerwekh JE, Ruml LA, Gottschalk F & Pak CY. The e€ects of twelve weeks of bed rest on bone histology, biochemical markers of bone turnover, and calcium homeostasis in eleven normal subjects. Journal of Bone and Mineral Research 1998; 10: 1594±1601. 27. Schlemmer A, Hassager C, Jensen SB & Christiansen C. Marked diurnal variation in urinary excretion of pyridinium cross-links in premenopausal women. Journal of Clinical Endocrinology and Metabolism 1992; 74: 476±480. 28. Woitge HW, Scheidt-Nave C, Kissling C et al. Seasonal variation of biochemical indexes of bone turnover: results of a population based study. Journal of Clinical Endocrinology and Metabolism 1998; 83: 68±75.

Biochemical markers and osteoporosis 399 29. Woitge HW, Knothe A, Witte K et al. Circannual rhythms and interactions of vitamin D metabolites, parathyroid hormone, and biochemical markers of skeletal homeostasis: a prospective study. Journal of Bone and Mineral Research 2000; 15: 2443±2450. 30. Uebelhart D, Schlemmer A, Johansen JS et al. E€ect of menopause and hormone replacement therapy on the urinary excretion of pyridinium cross-links. Journal of Clinical Endocrinology and Metabolism 1991; 72: 367±373. 31. Cummings SR & Black D. Bone mass measurements and risk of fracture in Caucasian women: a review of ®ndings from prospective studies. American Journal of Medicine 1995; 98 (supplement 2A): 24S±28S. 32. Johansen JS, Riis BJ, Delmas PD & Christiansen C. Plasma BGP: an indicator of spontaneous bone loss and of the e€ect of oestrogen treatment in postmenopausal women. European Journal of Clinical Investigation 1988; 18: 191±195. 33. Bonde M, Qvist P, Fledelius C et al. Applications of an enzyme immunoassay for a new marker of bone resorption (crosslaps): follow-up on hormone replacement therapy and osteoporosis risk assessment. Journal of Clinical Endocrinology and Metabolism 1995; 80: 864±868. 34. Slemenda C, Hui SL, Longcope C & Johnston CC. Sex steroids and bone mass: a study of changes about the time of menopause. Journal of Clinical Investigation 1987; 80: 1261±1269. 35. Cosman F, Nieves J, Wilkinson C et al. Bone density change and biochemical indices of skeletal turnover. Calci®ed Tissue International 1996; 58: 236±243. 36. Chesnut CH III, Bell NH & Clark GS. Hormone replacement therapy in postmenopausal women: urinary N-telopeptide of type I collagen monitors therapeutic e€ect and predicts response of bone mineral density. American Journal of Medicine 1997; 102: 29±37. 37. Hansen MA, Overgaard K, Riis BJ & Christiansen C. Role of peak bone mass and bone loss in postmenopausal osteoporosis: 12 year study. British Medical Journal 1991; 303: 961±964. 38. Ravn P, Fledelius C, Rosenquist C et al. High bone turnover is associated with low bone mass in both pre- and postmenopausal women. Bone 1996; 19: 291±298. 39. Bauer DC, Sklarin PM, Sone KL et al. Biochemical markers of bone turnover and prediction of hip bone loss in older women: the study of osteoporotic fractures. Journal of Bone and Mineral Research 1999; 14: 1404±1410. 40. Garnero P, Sornay-Rendu E, Duboeuf F & Delmas PD. Markers of bone turnover predict postmenopausal forearm bone loss over 4 years: The OFELY study. Journal of Bone and Mineral Research 1999; 14: 1614±1621. 41. Keen RW, Nguyen T, Sobnack R et al. Can biochemical markers predict bone loss at the hip and spine?: a 4-year prospective study of 141 early postmenopausal women. Osteoporosis International 1996; 6: 3999±4006. 42. AÊkesson K, Vergnaud P, Gineyts E et al. Impairment of bone turnover in elderly women with hip fracture. Calci®ed Tissiue International 1993; 53: 162±169. 43. Garnero P, Hausherr E, Chapuy M-C et al. Markers of bone resorption predict hip fracture in elderly women: the EPIDOS prospective study. Journal of Bone and Mineral Research 1996; 11: 1531±1538. 44. van Daele PLA, Seibel MK, Burger H et al. Case-control analysis of bone resorption markers, disability, and hip fracture risk: the Rotterdam study. British Medical Journal 1996; 312: 482±483. 45. Riis BJ, Overgaard K & Christiansen C. Biochemical markers of bone turnover to monitor the bone mass response to postmenopausal hormone replacement therapy. Osteoporosis International 1995; 5: 276±280. 46. Plantalech L, Guillaumont M, Vergnaud P et al. Impairment of gamma carboxylation of circulating osteocalcin (bone gla protein) in elderly women. Journal of Bone and Mineral Research 1991; 11: 1211±1216. 47. Szulc P, Chapuy M-C, Meunier PJ & Delmas PD. Serum undercarboxylated OC is a marker of the risk of hip fracture in elderly women. Journal of Clinical Investigation 1993; 91: 1769±1774. 48. Vergnaud P, Garnero P, Meunier PL et al. Undercarboxylated osteocalcin measured with a speci®c immunoassay predicts hip fracture in elderly women: the EPIDOS study. Journal of Clinical Endocrinology and Metabolism 1997; 82: 719±724. 49. Luukinen H, KaÈkoÈnen S-M, Pettersson K et al. Strong prediction of fractures among the elderly by the ratio of carboxylated and total serum osteocalcin. Journal of Bone and Mineral Research 2000; 15: 2473±2478. 50. Greenspan SL, Parker RA, Ferguson L et al. Early changes in biochemical markers of bone turnover predict the long-term response to alendronate therapy in representative elderly women: a randomized clinical trial. Journal of Bone and Mineral Research 1998; 13: 1431±1438. 51. Civitelli R, Gonnelli S, Zacchei F et al. Bone turnover in postmenopausal osteoporosis: e€ect of calcitonin treatment. Journal of Clinical Investigation 1988; 82: 1268±1274. 52. Gertz BJ, Shao P, Hanson DA et al. Monitoring bone resorption in early postmenopausal women by an immunoassay for cross-linked collagen peptides in urine. Journal of Bone and Mineral Research 1994; 9: 135±141.

400 P. R. Ebeling and K. AÊkesson 53. Braga de Castro Machado A, Hannon R & Eastell R. Monitoring alendronate therapy for osteoporosis. Journal of Bone and Mineral Research 1999; 14: 602±608. 54. Ravn P, Hosking D, Thompson GC et al. Monitoring of alendronate treatment and prediction of e€ect on bone mass by biochemical markers in early postmenopausal intervention cohort of study. Journal of Clinical Endocrinology and Metabolism 1999; 847: 2363±2368. 55. Ravn P, Clemmesen B & Christiansen C. Biochemical markers can predict the response in bone mass during alendronate treatment in early postmenopausal women. Bone 1999; 24: 237±244. 56. Marcus R, Holloway L, Wells B et al. The relationship of biochemical markers of bone turnover to bone density changes in postmenopausal women: results from the Postmenopausal Estrogen/Progestin Interventions (PEPI) trial. Journal of Bone and Mineral Research 1999; 14: 1583±1595. 57. Hannon R, Blumsohn A, Naylor K & Eastell R. Response of biochemical markers of bone turnover to hormone replacement therapy: impact of biological variability. Journal of Bone and Mineral Research 1998; 13: 1124±1133. 58. Delmas PD, Bjarnason NH, Mitlak BH et al. E€ects of raloxifene on bone mineral density, serum cholesterol concentrations, and uterine endometrium in postmenopausal women. New England Journal of Medicine 1997; 337: 1641±1647. 59. Delmas PD, Hardy P, Garnero P & Dain M. Monitoring individual response to hormone replacement therapy with bone markers. Bone 2000; 26: 553±560. 60. Bjarnason NH & Christiansen C. Early response in biochemical markers predicts long-term response in bone mass during hormone replacement therapy in early postmenopausal women. Bone 2000; 26: 561±569. 61. Nishi Y, Atley L, Eyre DR & Edelson JG. Determination of bone markers in pycnodysostosis: e€ect of Cathespin K de®ciency on bone matrix degradation. Journal of Bone and Mineral Research 1999; 14: 1902±1908.