Osteoporosis and the Clinical Endocrinologist: Reducing Fracture Risk through Individual Assessment and Therapy

Exploring Secondary Causes of Osteoporosis

Osteoporosis may be a primary disease resulting from involutional bone losses associated with normal aging and/or natural menopause; a secondary disease resulting from another disease state or related to medications; or an idiopathic disease for which there is no detectable cause.

Although most research has focused on primary osteoporosis, secondary osteoporosis is also very common. A recent study of a large database including 5,604 women and 561 men over age 50 years with osteoporosis reported secondary causes in 41.4% of women and 51.3% of men (Brown JP et al. J Bone Miner Res. 2002: abstr. 289).

Secondary osteoporosis is associated with a wide range of endocrine and hematologic disorders, gastrointestinal and connective tissue diseases, and hypogonadal states. It is also associated with miscellaneous causes such as emphysema, alcoholism, immobiliza-tion, chronic metabolic acidosis, multiple sclerosis, end-stage renal disease, and organ transplantation as well as multiple genetic disorders including cystic fibrosis, Gaucher’s disease, hemochromatosis, and Marfan’s syndrome. Many medications may also induce secondary osteoporosis. Among them are glucocorticoids, anticonvulsants, cyclo-sporines, and chemotherapeutics such as aromatase inhibitors, androgen deprivation therapy, and cytotoxic agents.

Available data suggest that individuals with conditions leading to secondary osteoporosis have a two-to-three-fold increased risk for spine and hip fracture compared with normal individuals of the same age group. For this reason, these patients should be considered for bone density and/or other laboratory testing to identify and treat premature or more advanced osteoporosis in order to prevent skeletal complications of their underlying disease.

While many patients with osteoporosis have known secondary causes, others have occult disorders that can accelerate bone loss and/or undermine the success of osteoporosis therapy. It is essential, therefore, that the clinician look for comorbid conditions in every patient with osteoporosis. To do otherwise is to run the risk of initiating inappropriate and/or ineffective therapy or of failing to recognize underlying pathology that requires specific intervention.

Marjorie M. Luckey, MD (Mount Sinai School of Medicine, New York City) used the case of Rose R. to illustrate the danger of omitting this step in patient management. This healthy woman without known causes of secondary osteoporosis sought counseling on osteoporosis prevention at the age of 55 years, 3 years following meno-pause, because of a strong family history of osteoporosis. Rose R. had T-scores at the femoral head and spine indicative of osteoporosis and low bone densities observed by Z-scores. Her initial treatment consisted of estrogen plus progestin, calcium supplementation, a daily multivitamin, and regular exercise. Over the ensuing 24 months she experienced significant loss of bone density, and calcitonin was added to her regimen. This was subsequently replaced by alendronate, but her bone density continued to decrease to the point of a vertebral fracture approximately 5 years following the initiation of treatment.

Laboratory testing performed at this point indicated low urinary calcium excretion. Additional testing to detect the cause of calcium malabsorption revealed the presence of transglutaminase antibodies, which are highly specific and sensitive for celiac disease. A small-bowel biopsy confirmed this diagnosis, one that is significantly associated with risk for osteoporosis and a 3.5-fold increase in fracture risk (Vasquez H et al. Am J Gastroenterol. 2000;95:183).

The point of critical importance in the case of Rose R. is that she was simply started on antiresorptive therapy on the basis of abnormal T-scores without the benefit of appropriate laboratory evaluations. Because an underlying pathology was not sought, her malabsorption syndrome went unrecognized and untreated. As a result, treatment was ineffective and bone loss continued to the point of fracture.

“How often do patients with osteoporosis harbor occult disorders that affect bone?” In answer to this question, Dr. Luckey cited four small studies of laboratory testing of patients with low bone density undertaken in specialty clinics. The prevalence of new diagnoses ranged from 11% to 63%. The most frequently found new diagnoses were vitamin D deficiency, hypercalciuria, exogenous hyperthyroidism, malabsorption, and malignancy. The studies differed, however, in the extent of laboratory testing performed, and all included populations that might have a higher prevalence of occult abnormalities (e.g., men, premenopausal women, and patients with diseases known to affect calcium and bone metabolism).

The prevalence of secondary osteoporosis in more typical healthy postmenopausal women with osteoporosis was examined in a study conducted at the Mount Sinai Medical Center by Dr. Luckey and colleagues. Healthy postmenopausal women (ages 46 to 87 years) with T-scores of -2.5 or lower underwent laboratory tests consisting of complete blood count, serum chemistry profile, 24-hour urine calcium concentration, parathyroid hormone, and 25-hydroxy vitamin D. Most patients were also evaluated with thyrotropin and serum protein electrophoresis. Overall, 44% of these apparently healthy subjects had one or more new unexpected diagnosis that have potential adverse effects on bone (Tannenbaum C et al. J Clin Endocrinol Metab. 2002;87:4431; Luckey MM, Tannenbaum C. J Clin Endocrinol Metab. 2003;88:1405). The most common findings were vitamin D deficiency (i.e., <20 ng/ml) (20%), hypercalciuria (10%), malabsorption (7%), hyperparathyroidism (3%), and exogenous hyperparathyroidism (2%). Interestingly, the prevalence of occult disease was similar in younger patients and those older than 65 years, and in those with and without known risk factors for osteoporosis. In addition, the Z-score (a comparison of bone mineral density to one’s peers), which is often cited as a reliable indicator of secondary disease and a basis for further testing, was of no predictive value.

In this study, a minimal screening laboratory panel consisting of a serum chemistry, complete blood count, 25-hydroxyvitamin D, and 24-hour urine calcium measurement in all patients, plus a TSH evaluation in those on thyroid replacement, identified 92% of those with underlying pathology at a cost of under $150 per diagnosis. Dr. Luckey noted that for evaluating men, some clinical endocrinologists add testosterone testing, because occult hypogonadism is a frequent contributor to secondary osteoporosis in this population.

Chemical screening for secondary osteoporosis has become so reliable that bone biopsy is now relatively infrequent, and is generally reserved for patients who have idiopathic disease or in whom secondary osteoporosis was highly unexpected. The exceptions include patients for whom osteomalacia, mastocytosis, or marrow infiltration such as multiple myeloma and Gaucher’s disease need to be ruled out; patients with fragility fractures despite normal BMD; and individuals with osteodystrophy for whom antiresorptive therapy is contemplated.

Interpretation and Application of Clinical Trial Data

The fundamental objective of osteoporosis therapy is to prevent increased bone fragility, with the ultimate goal of reducing fracture risk. Significant progress has been made in the last 15 years in the design of antiresorptive agents and one anabolic agent that have been shown to increase BMD. However, because bone strength is a product of overall bone quality including microarchitecture, the assumption that improved BMD alone will reduce fracture risk requires clinical verification.

Until the 1990s, most data on the reduction of fracture risk resulted from small studies. With the advent of bisphosphonates and their extensive clinical use, however, larger databases made large clinical studies feasible. Dennis M. Black, PhD (University of California at San Francisco) reviewed the data from some of these landmark trials in order to assess their applicability to decision-making for individual patients.

The first large study was the randomized and placebo-controlled Fracture Intervention Trial (FIT), a study of alendronate in approximately 6,400 women ages 55 through 80 years. One trial arm consisted of women with baseline vertebral fractures and the other of women with nonvertebral fractures. Subjects randomized to the alendronate arm were initially treated with 5 mg daily, a dose that was increased after 24 months to 10 mg on the basis of clinical data arising from other sources. After 3 to 4 years of followup, the mean increase in BMD was 6%, although the 10 mg regimen was associated with an increase of 10% compared with placebo (Ensrud KE et al. Arch Intern Med. 1997;157:2545). Importantly, in the course of the trial, women in the vertebral-fracture arm experienced a 50% reduction in risk for subsequent vertebral fractures and a 25% risk reduction for other fractures (Black DM, Thompson DE. Int J Clin Pract. 1999(suppl);101:46). There
was also a significant reduction in hip fractures.

Women in the clinical-fracture arm experienced similar increases in BMD and vertebral fracture, although the absolute incidence was low. However, the prevention of hip and other nonvertebral fractures in this group of subjects correlated with initial hip bone density: Only women with baseline T-scores of less than -2.5 at the femoral neck experienced a statistically significant reduction in risk for hip and other non- vertebral fracture. These results were confirmed in the Fosamax International Trial (FOSIT), a year-long randomized trial involving approximately 2,000 women randomized to receive either alendronate 10 mg daily or placebo. Once again, the incidence of nonvertebral fractures correlated with low initial BMD.

A Hip Intervention Program (HIP) study enrolled only women over the age of 70 years and was designed to evaluate hip fracture prevention utilizing risedronate. The results were somewhat controversial, primarily because although women under 80 years of age were admitted only if their T-scores were very low—less than -3.0—or if they had one or more additional risk factors, most of the older women had not had BMD measurement. Overall, however, risedronate in this trial was associated with a 30% reduction in risk for hip fracture, but results varied between the two age groups. Among women 70 to 79 years of age with very low bone density, the mean risk reduction was 40%, while there was no significant reduction among older women. Dr. Black attributed this discrepancy to the confounding absence of BMD data for the older women. Nevertheless, the trial did confirm the preventive value of bisphosphonate therapy in women in their 70s with very low bone density.

The Women’s Health Initiative (WHI), a multicenter study involving 16,608 unselected women ages 50 to 79 years taking combination hormone replacement therapy, was discontinued early (mid-2002 compared with a projected completion in 2005) because of lack of a global benefit and evidence of increased risk for breast cancer, heart disease, and thromboembolic morbidity and mortality. During the course of the trial, however, there was a significant reduction in hip fracture and in all fractures associated with estrogen plus progestin (Cauley JA et al. JAMA. 2003;290: 1729).

With the exception of the WHI, all of the clinical trials on fracture risk reduction due to bisphosphonate therapy enrolled selected patients, as do most randomized and placebo-controlled clinical trials. Because the environment of a clinical practice differs considerably from that of a clinical trial, Dr. Black questioned the generalizability of these results to unselected individual patients. He concluded, nonetheless, that the results appear to be generalizable across age groups of postmenopausal women, particularly among patients with low baseline bone density with histories of vertebral fracture.

Dr. Black next turned his attention to a comprehensive meta-analysis of osteoporosis drugs conducted by the Osteoporosis Research Advisory Group (ORAG). The objectives of the study were (i) to summarize and translate evidence from randomized clinical trials into systematic reviews than can be used by clinicians, and (ii) to learn the optimum treatment of postmenopausal osteoporosis based on meta-analysis. The treatments studied were calcium, vitamin D, calcitonin, hormone replacement therapy (HRT), three bisphosphonates (alendronate, etidronate, and risedronate), and raloxifene (Cranney A et al. Endocrine Rev. 2002:23:570).

The analysis found that alendronate (especially at higher doses), etidronate, risedronate, HRT and raloxifene were all associated with significant increases in BMD, with the highest scores for high-dose alendronate and HRT. All treatments except calcitonin, calcium, and HRT significantly decreased risk for vertebral fractures. With regard to nonvertebral fracture, however, only alendronate at doses of 10 mg and higher and risedronate were associated with statistically significant risk reduction, 49% and 26%, respectively. Reduction in risk for hip fracture was not studied comprehensively, but the data for alendronate appeared to be consistent with those for nonvertebral fracture.

For guidance on the duration of therapy, extension data for the original phase III alendronate trials have been published. These indicate that patients who continue on alendronate therapy for up to 10 years experience maintenance of the BMD improvements and suppression of bone markers achieved in years 1 through 5, whereas those who discontinue treatment after 5 years go through a gradual decrease in BMD over the following 5 years. The decline, however, is not as steep as that of a hypothetical placebo group. Although the data are inconclusive, there is reason to think that with both alendronate and risedronate, fracture rates may continue to decline with long-term therapy-induced maintenance of BMD despite the continuing aging of patients. Data from a randomized long-term alendronate trial involving 1,100 women will be submitted for publication soon. In this trial, all women were treated for 5 years and then randomized to receive either alendronate or placebo for the 5 succeeding years.

Finally, Dr. Black addressed the results of a recent analysis of women in the WHI who, because of prior hysterectomy, used estrogen maintenance only (N=10,739). They ranged in age from 50 to 79 years. (The trial, which was designed for 8 years of followup, was discontinued by the NIH in 2004 after 7 years.) Based on available data, the study concluded that estrogen-only therapy is associated with a decreased risk of hip fracture (Anderson GL et al. JAMA. 2004;291:1701).

Individualized Assessment and Therapy for Patients at Risk

Michael Kleerekoper, MD (Wayne State University) noted that although detection rates for osteoporosis are virtually identical when selecting single sites for measurement, the rate increases as more sites are tested. Detection of low BMD at any single site suffices for a diagnosis of osteoporosis. Patients with osteoporosis detected by BMD (T-score less than -2.5) should be offered therapy, as should patients with fragility fractures irrespective of BMD measurement. (“Fragility fracture” is arbitrarily defined as a fracture resulting from trauma that is equal to or less than a fall from a standing height.)

Despite its large size, the results of the National Osteoporosis Risk Assess-ment (NORA) are not widely recognized or understood. In this study of 149,524 women conducted at primary care sites nationwide, bone density was measured using peripheral technology including WHO-endorsed measurements of the finger and forearm as well as heel ultrasound and heel single-energy x-ray absorptiometry. All devices studied predicted fracture risk (Miller PD et al. J Bone Miner Res. 2002;7:2222). However, as Figure 1 indicates, the osteoporotic fracture rate in this study correlated with patient age.

The trend illustrated here is consistent with femoral-neck-fracture discharge data reported by the Centers for Disease Control and Prevention (CDC) that compare fracture rates among women 45 through 64 years with those of women 65 years and older. For all fractures, there is a three-fold increase in the older group; but for fractures of the femoral neck, the ratio is 1:18. These data and many others underscore the point that postponing osteoporosis screening beyond age 65 may leave patients exposed to serious fracture risk. Although there clearly are indications for looking for potential osteoporosis in younger women, the recommendation of the Preventive Services Task Force that age 65 be considered the critical cut-off point is founded on sound clinical evidence.

Alarmingly, however, many women present with fractures who have BMD T-scores higher than -2.5. In absolute numbers, in fact, fracture occurs more frequently in individuals with osteopenic BMD levels than in patients who are frankly osteoporotic based on T-score. Thus it becomes apparent that the risk factors for osteoporosis or for low bone mass (osteopenia) are not the same as those for fracture. The clinician’s challenge, therefore, is to identify individuals at risk irrespective of T-scores. In order to meet this challenge, the AACE has developed guidelines to assist endocrinologists in identifying those women with BMD scores not below -2.5, but who should be considered for therapy on the basis of clinical history and examination. These guidelines are available for viewing and downloading at www.aace.com.

For identifying patients at risk for fracture, the use of biochemical markers of bone remodeling (turnover) may be helpful, as is indicated by data emerging from Europe’s prospective Epidemiologie de l’Osteoporose (EPIDOS) study. In one analysis, early postmenopausal women were screened for multiple urinary and serum bone markers and identified as having low or high bone turnover based on whether the value was within (“low”) or above (“high”) two standard deviations from the mean for premenopausal women. The risk for subsequent fracture is demonstrated in Figure 2.

These results suggest that if a patient is undergoing rapid bone turnover even in the presence of a normal BMD, the likelihood of rapid bone loss (and subsequent fracture) may be sufficiently great to initiate antiresorptive therapy prior to a decrease in BMD.