For the measurement of bone-forming activity there are several clinically useful markers: alkaline phosphatase, osteocalcin and protocollagen I extension peptides.

Alkaline phosphatase activity is derived from various tissues such as the liver, bone, placenta, etc. Bone and liver isoforms are the most common (90%). Both are found in the same proportion in the healthy individual and differ in glycosylation patterns, and there is a cross-activity of 10%–20%, according to studies with monoclonal antibodies.5 The bone isoform has the advantage of presenting no variation between genders and not influenced by circadian rhythm, so that, despite having low sensitivity and specificity in the study of metabolic bone disease, it is easy to detect in the absence gestation and liver disease.6

Osteocalcin is the most abundant non-collagenous protein of the extracellular matrix. Specific for bone and dentin, it is elevated in situations of increased bone turnover, has a short half-life and is eliminated via the urine, so its levels are increased in situations of renal failure.7 Its exact role in bone remodeling is not well established. Recent work has analyzed the potential role of infracarboxilated osteocalcin in predicting bone mass and risk of fracture.8

Because type I collagen is the main product of synthesis of the osteoblast, the amino-terminal carboxy propeptides would, theoretically, be the ideal marker of bone formation. However, the fact that type I collagen appears in other tissues other than bone limits its use in the study of metabolic bone disease.9

Historically, urinary calcium was the first test used to assess bone resorption. However, the fact that it is influenced by various factors, such as calcium intake, intestinal absorption and renal threshold of excretion of calcium, makes its determination a test with low sensitivity and specificity, and is currently unused. The same applies to urinary hydroxyproline: different tissue origins and the influence dietary intake exerts on it are to be considered in its low yield.1

The collagen molecules in bone matrix are covalently linked by pyridinoline (Pir) and deoxypyridinoline (Dpir) forming fibrils. The former is also found in cartilage, however, Dpir is more specific for bone.10 They do not depend on diet, and are not absorbed via the gut. They express both changes of bone metabolism, rising in childhood, menopause, osteomalacia, hyperparathyroidism and hyperthyroidism, and lowering due to estrogen and bisphosphonate treatment.11

Other elements released during bone resorption are the carboxyterminal (ICTP, CTX) and amino-terminal (NTX) telopeptides of collagen. They have shown a significant correlation with BMD in postmenopausal women, and both CTX and NTX are considered the most clinically useful markers of bone resorption currently available.12,13

Tartrate-resistant acid phosphatase 5b (5b Fatra) is a lysosomal enzyme not only involved in osteoclast bone degradation, but is also present in other tissues. It is poorly specific, and together with the methodological difficulty in identifying it, currently makes it of little use.14

There are newly developed bone turnover markers whose clinical value is still under study. Cathepsin K acts to degrade the collagen and its serum levels reflect the number of osteoclasts; it has been suggested as a marker of bone resorption. Urinary osteocalcin fragments also come from the resorption of bone matrix and appear to be more specific. In the assessment of bone quality posttranslational modifications of bone matrix proteins such as infracarboxylated osteocalcin, native and isomerized forms of CTX and pentosydine have been considered especially important.13,15

Currently represents its best established clinical use. Several studies have shown that after initiating antiresorptive therapy there is a significant decrease in both resorption (within 4–6 weeks), and in bone formation markers (2–3 months). In most cases there is a low value reached 2–3 months after start of treatment and remains constant while the patient continues taking the drug. A significant change would be between 40% and 70% of reduction in markers of bone resorption (CTX in serum and urine NTX and Dpir in urine), when using a potent antiresorptive (bisphosphonates), and more modest reductions (30%–40%) with less energetic catabolic (raloxifene). Therefore, the changes will depend on the therapeutic agent employed and the marker analyzed.

Thus, the failure to observe these reductions indicates poor adherence to treatment by the patient or the improper administration of the drug.18–20

The bone forming treatment initially induces a rapid increase in bone formation markers, followed by a subsequent increase in resorption markers. In this regard, several studies have pointed to markers such as PICP and especially PINP as they have a higher sensitivity to predict changes in BMD in patients treated with teriparatide. A proposed performance algorithm in patients treated with this anabolic agent is related to the variation detected in PINP: increases above 10μg in this formation marker, 3 months after initiation of therapy, suggest an appropriate response to it.21,22

Although it is desirable to have markers of formation and resorption, according to the available evidence the markers of bone turnover that are more sensitive and clinically useful are serum CTX—when using an antiresorptive—and PINP when treating with anabolic drugs. Their determination, after 2 or 3 months of treatment, offers the remarkable advantage of being able to assess the effectiveness of medication and reassuring patients without having to wait 12–24 months to document changes in BMD.23

Several studies have demonstrated an association between changes in bone turnover markers after antiresorptive or anabolic treatment and long-term anti-fracture efficacy. A meta-analysis of 18 clinical trials found that, during antiresorptive therapy, a 70% decrease in resorption markers decreased the risk of vertebral fracture by 40% and a decrease of 50% of formation markers reduced this risk by 44%.24 In this sense, there is evidence that an increase in formation markers, such as PIPC and FAO, a month of starting treatment with teriparatide was associated with improvement in bone structure.25

However, the optimal threshold for each marker is not well established and there is no long-term data when the biomarkers are reduced below the reference range. Some researchers support the hypothesis that it could interfere with physiological bone repair of “microcracks” and, therefore, lead to increased bone fragility.26

It has been suggested that patients with a greater increase in bone turnover markers respond better to antiresorptive therapy. In a 2-year prospective study in women with hormone replacement therapy, patients placed in higher quartiles for urinary excretion of NTX showed a higher bone gain.27

Theoretically, the reference levels of biomarkers could be used to guide treatment decisions more rationally. However, most clinical trials have shown that anti-fracture efficacy is largely independent of the baseline markers of bone remodeling. Therefore, it is not considered currently as clinically useful.28

BMD measured by DXA remains the best predictor of fracture risk. However, BMD does not explain everything and must take into account other variables, including the rate of bone turnover. Data from several prospective studies in elderly postmenopausal women show an association, independent of BMD, between the indices of bone turnover and osteoporotic fracture risk, although the results are discordant, particularly with regard to bone29 formation markers. Elevated bone resorption markers have been associated with an increase of approximately twice the risk of vertebral and nonvertebral fractures compared with women in whom such markers were at normal levels. The EPIDOS study results were consistent with this fact: In older women, increased bone resorption markers in urine (CTX) were associated, after adjustment for BMD, with an increased risk of hip fracture. This suggests that a combined approach, using BMD and biochemical markers of bone remodeling, could improve the prediction of fracture risk in postmenopausal women, and this risk increases in those with low BMD and/or elevated markers of bone remodeling.30

Another prominent use of these biomarkers is the possibility of identifying women who will present a high rate of bone loss in the years following menopause, in order to start a preventive strategy for osteoporosis. As is known, bone densitometry is an important predictor of fracture risk, but a single measurement does not portend the rate of bone loss. Several studies have shown that markers of bone turnover may be useful in determining the rate of bone mass loss,1 with markers of resorption providing greater predictive power than those of formation.5,31

A recent review has analyzed the association of biomarkers with loss of bone mass in different skeletal sites. The most consistent relationship has been detected at the distal radius and calcaneus. In the hip, the relationship is more modest, while in the lumbar spine it is negligible. The latter finding could be due to the high prevalence of vertebral osteoarthritis in the elderly. In summary, markers of bone turnover together with other demographic variables could predict 30%–40% of the variation in bone loss in untreated postmenopausal women. However, the bone turnover threshold from which it is necessary to intervene in order to prevent bone loss32 has not yet been established.