Osteoporosis therapies are designed to improve bone strength and reduce the risk for fracture. When the capacity to reduce fracture is assessed, these therapies reduce risk either expressed as relative risk (RR) or absolute risk (AR) within the context of clinical trials [1,2]. Whether or not risk reduction is expressed as either RR reduction (the ratio of absolute risk reduction in the treated group to absolute risk reduction in the placebo group) or absolute risk reduction (absolute number of fractures over a specified period of time), no osteoporosis treatment abolishes risk. People may still develop fragility fractures even when the osteoporosis-specific pharmacological therapy is altering bone strength in the best biological manner that the agent can induce. In biology, no pharmacological therapy abolishes risk for any specific event. In that regard, for example, patients with high cholesterol still have heart attacks, even though their cholesterol has been normalized by a cholesterol-lowering drug. In the same way, clinicians caring for individual patients will have patients who will suffer fractures while on osteoporosis-specific pharmacological agents; the clinician must then decide if the therapy is “working or not.” Measuring optimal bone biological effect (efficacy) of a bone antiresorptive or anabolic agent is not easy to define—and in that regard, clinicians also use surrogate markers [changes in bone mineral density (BMD) or bone turnover markers (BTMs)] in addition to fracture events as measures of the effectiveness of an agent, since we have no clinical tool that is office-based to measure bone strength [3,4,5,6]. Hence, the efficacy of intervention is measured in four ways:

Waiting for a fracture to occur or not occur is not a highly appealing efficacy end point for either physicians or patients, and, when a fracture does occur during therapy, both physician and patient feel that the specific treatment has failed, even though both parties may know that no treatment abolishes risk. Nevertheless, when a fracture occurs during treatment, the clinician should first review a checklist with the patient to ensure that
everything is being done to guarantee treatment success. This checklist includes queries about adherence and persistence, and queries about following the proper adherence to the dosing instructions.

Specific laboratory tests should also be examined or reexamined that, when found to be abnormal, might contribute to “treatment failure.” A few of these laboratory tests, if abnormal, may contribute to inadequate pharmacological response. These include a low 25-hydroxyvitamin D level; an elevated serum parathyroid hormone level; a low or high 24-hour urine calcium excretion; and a positive serum transglutaminase IgA antibody, the latter test a good screen for asymptomatic celiac disease.

Vitamin D insufficiency is prevalent in all populations, even in the United States, at all latitudes [7,8]. Low vitamin D levels may contribute to inadequate pharmacological response to osteoporosis therapies due, in part, to inadequate calcium absorption or the secondary hyperparathyroidism that often accompanies poor calcium absorption [9]. Celiac disease is also very prevalent; it exists in 1 of every 250 Caucasian Americans and, in certain other ethnic populations, as high as 1 in 50 persons [10,11]. Celiac disease is often completely asymptomatic and yet may lead to the specific malabsorption of calcium or iron, and is one of the most common secondary causes of osteoporosis [12,13]. Serum vitamin D levels are often normal in patients with celiac disease because vitamin D is absorbed in the terminal ileum and celiac disease begins in the proximal duodenum. It may take many years for celiac disease to migrate down the small bowel. As celiac disease advances down the small bowel, patients then may develop clinical symptoms such as bloatedness, diarrhea, and frank malabsorption, and in these more advanced subsets, 25-hydroxyvitamin D levels may finally become low. Most of the fractures or losses of BMD in patients being treated with bisphosphonates who have celiac disease are treatment failures because bisphosphonates are not absorbed in the presence of celiac disease. The transglutaminase IgA antibody is the best screening test for celiac disease, but its sensitivity is not nearly as good in asymptomatic as opposed to symptomatic celiac disease (˜70% vs. ˜90) [14]. Hence, a small bowel biopsy, the gold standard for the diagnosis of celiac disease (histological diagnosis), should still be done for patients with osteoporosis treatment failure who have a negative transglutaminase IgA antibody in two clinical scenarios:

Urine calcium excretions that are consistently low, despite an adequate calcium intake, are strongly suggestive of poor calcium absorption and a high probability of celiac disease in patients with normal renal function. One of the more common conditions that lead to poor calcium absorption is celiac disease. Celiac disease may contribute to osteoporosis by preventing calcium or bisphosphonate absorption.

There are other gastrointestinal (GI) conditions in which oral bisphosphonates are probably not absorbed: gastrojejunostomy, small bowel resections,
hemigastrectomies, and other clinical states in which upper-GI transit time is rapid, including gastric stapling—and all of these GI conditions may be related to bisphosphonate “treatment failure.” The fundamental absorption characteristics of oral bisphosphonates are very erratic and fastidious in the first place, with less than 1% of an oral bisphosphonate formulation absorbed under the best circumstance. Hence, it does not take much of an alteration of GI tract function to abolish bisphosphonate absorption. If clinicians could measure a bisphosphonate blood level, clinical decisions regarding the uncertainty about bisphosphonate absorption could be resolved. Yet there are no clinically available bisphosphonate assays, so clinicians are often left in the dark about the bioavailability of oral bisphosphonates.

A high 24-hour urinary calcium excretion might also contribute to treatment failure. Hypercalciuria might be associated with bone loss because the high urinary calcium excretion might be related to a disease that directly contributes to bone loss (e.g., primary hyperparathyroidism, renal tubular acidosis), or the hypercalciuria might, per se, contribute to a negative calcium balance [15,16]. Hypercalciuria does not necessarily mean the patient is in negative calcium balance (calcium balance studies are not pragmatic to do in clinical practice) because hyperabsorption may be contributing to the hypercalciuria without the excess urinary calcium coming from the large bone reservoir of calcium. Nevertheless, a patient who is hypercalciuric and has a fracture or is losing BMD while on pharmacological treatment may benefit by lessening urine calcium excretion [17,18].

Hypercalciuria is best lowered by a thiazide diuretic, and this author prefers either a hydrochlorothiazide-amiloride combination or chlorthalidone (50 mg/day). After 6 to 8 weeks of the diuretic, the 24-hour urine calcium and serum calcium should be remeasured for two reasons: to see if (a) the thiazide is effective in normalizing the urinary calcium excretion (<300 mg/day) and (b) if hypercalcemia develops when normocalciuria is achieved. If hypercalcemia develops after achieving normocalciuria, then one may have unmasked primary hyperparathyroidism—the “thiazide-challenge test” [19]. Patients with mild primary hyperparathyroidism may be normocal-cemic because they are hypercalciuric; for example, the kidney is protecting the blood (and brain) from hypercalcemia. Hence, if sustained hypercalcemia develops once normocalciuria is maintained, then additional testing for primary hyperparathyroidism may be indicated. If the patient does not develop hypercalcemia after becoming normocalciuric, then the clinician may continue the thiazide to see what effect normocalciuria has on the previous loss of BMD. If the BMD loss is reversed with a thiazide-induced normocalciuria, then the thiazide may be continued with or without combination therapy for the osteoporosis [20]. If there is no effect of normocalciuria on BMD, fractures, or kidney stone history (or if noncontrast CT of the kidneys has been done to rule out asymptomatic nephrocalcinosis), then the thiazide should be discontinued, because many of these hypercalciuric patients are so-called “healthy hypercalciurics”—they exhibit elevated urinary calcium excretion of no clinical relevance [21].

In those patients who are “treatment failures” defined by sustaining fractures or losing BMD while on therapy and who also have a persistently high
bone resorption marker while on oral bisphosphonates where GI absorption may be the cause of treatment failure, the clinician has two options:

If an intravenous bisphosphonate is chosen, there are three available. Two of these are currently being used off-label [i.e., they do not have a Food and Drug Administration (FDA) registration for an osteoporotic indication (pamidronate and zolendronic acid)], and the other is an intravenous bisphosphonate recently approved by the FDA for osteoporosis (ibandronate) [22,23,24]. The dose of pamidronate is usually 30 mg diluted in 250 mL D5W given over 2 hours, with an occasional patient requiring 60 mg every 3 months. The dose determination is decided by monitoring the bone resorption marker (NTx/CTx) between doses. This author measures the first marker 1 month after the first IV bisphosphonate infusion and the second marker 2 weeks before the second bisphosphonate infusion. The resorption marker should be maintained within a level consistent with sustained suppression of bone turnover levels. Urine NTx values in the 30s or lower [nm bone collagen equivalents (BCE)/nm creatinine], or serum CTx values between 0.5 and 3.2 pmol/L—the so-called “therapeutic” range—are seen in most treated patients in the clinical trials on the registered doses of bisphosphonates [25,26]. Sustained turnover marker suppression to some consistent level seems to be important to the improvement in bone strength, particularly with intermittent bisphosphonate dosing regimens [26,27]. In the early intermittent intravenous bisphosphonate studies (1 mg every 3 months of ibandronate), in which the spine BMD increased to levels seen in other bisphosphonate studies that did have fracture risk reduction, fracture risk reduction was not seen, probably because of inadequate and nonsustained suppression of bone turnover between dosing schedules at this dose [27]. The FDA-approved ibandronate dose of 3 mg every 3 months kept bone resorption markers more consistently suppressed between dosing intervals, which, in theory, should improve bone strength similarly to the fracture reduction—proven daily ibandronate dosing regimen [28].