Tamara Vokes, MD, concludes her two part talk discussing Individualized Approach to Osteoporosis, diagnosis and treatment.
TAMARA VOKES: I am Tamara Vokes. I am an endocrinologist in the section of Endocrinology Department of Medicine at University of Chicago Medicine, and I will talk to you about the individualized approach to osteoporosis. This is part two, while I've discussed diagnosis and treatment primarily as it pertains to post menopausal women and men over 50, I have, in part one, talked more about the pathophysiology of bone loss and a general approach to looking for secondary causes of osteoporosis. So the learning objectives here, again, to individualize approach to osteoperosis, identify clinical risk factors, integrate them with bone density measurement when evaluating fracture risk. And I will focus on what I think is really important nowadays, which is to compare and contrast the risks and benefits of different therapeutic modalities. So the approach, the prevention, and treatment of osteoporosis to postmenopausal women is actually based on more data than anything else we do in osteoporosis. We have guidelines, because we have a lot of studies. Actually, all the studies that form the basis of clinical practice of osteoperosis have been performed in post menopausal women. Then this data was extrapolated to men with relatively fewer studies and to younger people with even fewer studies. So the guidelines that we currently have for approaching this patient population are really aimed at preventing fractures over 5 to 10 years. This is obviously very appropriate when you're dealing with an older person-- let's say 70 plus, who have significant fracture risk. However, it is really important to individualized therapy in these patients, but particularly in young postmonopausal women who may not meet the guidelines for aggressive management. And yet, as I showed in the previous lecture, they could be losing bone, and they can be very significantly increasing their risk of osteoporosis and fractures if left untreated. So that's a caveat when looking at guidelines for osteoporosis management. So in addition to bone density that we saw is an important predictor of fracture risk, age is also very important and an independent predictor. You can see on this slide how at any given age, at any given bone density, the fracture risk is much higher in an 80 year old than in a 50 year old at the same level of bone density. So the importantness of this is that when the patient comes and says, I'm 52 and my bones are that of an 80-year-old woman, that is not correct. Even though her bone density may be close to the mean of an 80-year-old woman, her actual fracture risk is nowhere that high. And that is important to know, because this woman who is 52 shouldn't stop running or being active because she's afraid that she will crumble. Her fragility is not as high, probably because bone quality, which we talked about, is another determinant of fragility, is much better in a younger person than in an old person. As far as school should be screened with [INAUDIBLE]. There are several societies that have come up with recommendations. Generally, it's agreed that women over 65 should be screened, and so that meets the criteria for being on Medicare. Men over 70, according to some societies. But then the disagreement is, when should younger people be screened? Generally, women who are less than 65 years old but have risk factors should be screened. For instance, patients I mentioned in lecture one, patients with eating disorders with anorexia nervosa. If you wait for them to be 65, we've really missed the opportunity to prevent bone loss that otherwise would occur. They will be fracturing long before 65. So generally, those who have risk factors. As far as how to go about that, again, disagreement between different societies. For instance, USPTA's recommendations are to look at FRAX and determine who should be screened among younger women. This however, is not really supported by data. And just last year, ISCD, that's International Society for Clinical Densiometry, looked at that recommendation carefully, and found it would not capture women who actually have low bone mass. So they have come up with a recommendation that's based on more specifying of individual risk factors, and thus, deciding who should be screened among those who are younger than the Medicare age group. It's really important to also screen for low bone density in women who are stopping estrogen therapy. Because when estrogen is stopped, it's like menopause happening all over again. There will again, be significant bone loss. And if women already have low bone density, we should not really allow that bone loss. We might be more cost effective if you want by preventing bone loss that occurs when estrogen is stopped. So there is a clear set of guidelines for approaching postmenopausal women and men over 50. And the most widely used set of guidelines is from National Osteoporosis Foundation, NOF, which gives guidelines about treatment. Basic care-- and I'll talk about that, is suitable for all, regardless of bone density. And then beyond, it's important to assess risk factors and bone density if the person has risk factors. Pharmacologic therapy is indicated in people, men or women, postmenopausal women, if they have hip or vertebral fractures. Hip fractures and vertebral fractures are a hallmark of osteoporosis. So if the person has these fractures, unless they're from trauma or for underlying pathology such as metastasises to the bone, if they have these fractures, they've declared themselves as having osteoporosis, and they warrant treatment. And the other category where treatment is indicated is based on bone density measurement with a T-score of less than minus 2.5 at the spine or the hip. For those who have better bone density than that, we advised to look at FRAX. So if the T-score is in so-called osteopenia minus 1 to minus 2.5, we advise to look at FRAX. And I'll show you the appearance of that calculation too, with the idea of that treatment would be applicable to those who have 10 year fracture probability of 20% or more for major osteoporotic factories, or 3% or more for hip fractures. FRAX is a publicly available website. And I actually advise patients to go to that and play with entering their data. It is developed by WHO and the Sheffield, and this is the website. This is a calculation tool that gives you the fracture probability, taking into account bone density measurement, as well as clinical risk factors. It is calibrated for different populations. For instance, in the US, we have Caucasian, Black, Asian, and Hispanic. Other populations have also been captured with the understanding that the data is really good for Caucasian women from the Western Europe and the US, it is not as good for other populations. And it will actually calculate the fracture probability relative to the competing risk of dying, because it takes into account mortality statistic. So it will actually give you the probability of a patient sustaining a fracture versus the probability of dying. So you will see that actual probability of fracture might be higher in an 8 year old and in an 85 because the probability of dying is higher. So with that caveat, what the FRAX calculation takes into account are these various clinical risk factors and bone density measurement as well. Here, it's entering the T-score, but even better ways to actually enter bone density from the printout. And it's calibrated for both main manufacturers of bone density software. And then it will calculate the probability of major osteoporotic fracture and hip fracture. And I gave you the cut-offs that are recommended for pharmacologic therapy. Now, I think FRAX is a big improvement over treating T-score. We used to say, oh, anybody who has this T-score gets treated, and who doesn't have that T-score basically wouldn't know what to do. So it's an improvement in that we are taking into account clinical risk factors. However, there are really some deficiencies in FRAX. One is that FRAX was developed for assessment of fracture risk. Treatment should ultimately be at the discretion of a clinician. So we shouldn't have very strict guidelines that are based on the stool. But the other limitations are that all the variables here are dichotomized. An example of that is glucocorticoid use. It's very different if the patient is on 7 milligrams for a couple three months versus 60 milligrams. Obviously, the fragility is much higher with the higher dose, and there is no provision too great the fragility increment versus the dose of glucocorticoids. Spine BMD is not considered only a hip measurement. There is a patch that has been developed by Bill Leslie from Canada, which allows to recalibrate the fracture probability taking into account the spine measurement. There is no provision for nonskeletal risk factors such as risk of falling. You could take that into account and say, oh, because you have high fall risk, I'll be a little bit more aggressive. But there is no numerical approach. And then previous fracture is not clearly defined. It's very well known that some fractures are more predictive of future fractures than others. For instance, if somebody's had an ankle fracture, that may not be as significant. In contrast, vertebral fractures that are shown on this slide are very strong predictors of future fractures. And in FRAX, they would be entered the same. Vertebral fracture are very important to think about when evaluating a patient with osteoporosis, because they predict future fractures independent of BMD. So having information on both vertebral fractures and bone density measurement provides a much better assessment of fracture risk. Prevalent vertebral fractures, meaning those that are found on an image at the time of evaluation, you don't know when they happened-- could be any time during life-- they're clinically recognized in only 1/3 of patients will have them on imaging. And patients are often really surprised. It's like, it's impossible that I have broken my spine, I have never had anything to suggest that. But they are often happening not as a discrete event, but just a slow collapse of the bone. So no wonder they're so predictive of fracture, they're telling you the bone is really soft, and it's really very much prone to fracturing. So we need imaging, because they are so often not recognized. We need imaging to detect them. As a result of that, vertebral fracture assessment has been developed as an application on the most common densitometer models which allows to get the rhetorical lumbar spine image at the time of bone density, and integrate that information with your bone density measurement. And it's called Vertebral Fracture Assessment, abbreviated as VFA. But of course, you can get the same information from an X-ray, or a CT, or an MRI. This is just something that can be done when the patient is on the densitometer. So there have, again, been guidelines as far as who should have imaging for detection of vertebral fracture. And [INAUDIBLE] just last year added vertebral imagining as an important aspect of osteoporosis management. Generally, it's based on age. Again, 65 for older for women, 70 or older for men if they have T-score less than minus 2.5. They also include people who have no bone density. Just if they're old, there is a debate on whether that is a really good idea. And then those who have other risk factors-- ISCD has actually also looked at that in a more quantitative way by actually looking at the available data and developing guidelines that are confirmed by analysis of the data. And this is, I would say, a little bit more substantiated recommendation, which is women over the age of 70, men over age 80, T-score less than minus 1 for any of these considerations. Historical height loss, self-reported history of vertebral fracture, and then glucocorticoid therapy. As you know, glucocorticoids tend to be associated with a significantly high risk of vertebral fractures. Now, why are vertebral fractures important? Well, because they constitute the diagnosis of osteoporosis, and they could lead and should lead to more aggressive therapy. The examples of that would be to use pharmacologic therapy in a patient without osteopenia, who otherwise does not qualify. We may also consider anabolic, which is a more aggressive therapy, if we find vertebral fractures. And also, deciding about the drug holiday's very important nowadays. And in somebody who has vertebral fractures, we shouldn't as easily, or not that's all, consider a drug holiday. I mentioned that there's some aspects of increased bone fragility that FRAX doesn't capture, and one of them is diabetes. This is a relatively new finding. It's been traditionally believed that diabetes is not associated with increased fracture risk, partly because bone density is generally higher in patients with Type 2 diabetes, primarily due to their higher weight. Recent data, however, suggests that there is an actual increase in fracture risk that's associated with diabetes. Different studies, but probably about 1.3 to 1.4 times higher fracture risk just due to diabetes. And patients with diabetes tend to fracture at higher level of bone density than controls. Given the epidemic of diabetes, this becomes an important thing to know, because we will perhaps need to address these patients more aggressively. Why patients with diabetes have higher fracture risk is not really clear. It's not fully explained by the higher fall risk. They have neuropathy. They tend to fall, they might get hypoglycemia. That decreases fall risk. But even when controlling for all of that, we still don't quite get to the fracture risk increase that is observed in studies. There seems to be an alteration in bone quality, and there's some studies that show that there is a difference in architecture. And importantly, it's not captured in FRAX. So we need to perhaps increase upwards our estimate of fracture risk when dealing with diabetes. Mentioning that weight in diabetes is not protective against fracture brings us to another relatively new observation, which is that obesity is not good for bone. Traditionally, we thought, oh. Overweight people, obese people, do not fracture, because we have quite a lot of data that's higher weight is associated with higher BMD, traditionally believed to be protective against osteoporosis. Interestingly, higher than expected fracture rates have been observed in obese individuals. And it's not exactly the same as in lean people. Some fractures are increased-- particularly ankle, leg, excluding the hip, humorous, and spine. But not hip, wrist, or pelvis. Why exactly is it the pattern of fall? Not clear. Within an obvious population, risk factors are similar to non-obese. I.e., higher bone densities, protective lower bone density gives high risk. The treatment is not very clear. It seems that the efficacy's a little bit slower, although there are very few studies that include any proportion of patients who are obese. And perhaps the treatment efficacy is lower, because the volume of distribution is larger so that the dosing may not be appropriate for these individuals. Coming to treatment of osteoporosis, there are several components. One is to improve the environment for the bone by snot smoking, limiting alcohol intake to less than two drinks per day, proper nutrition activity. And then finally, pharmacologic therapy. I actually spent quite a bit of time with my patients discussing nutrition and activity. Because often, that is an area where they can use some information and support. But also, I find that at this time when patients are so afraid of medication, if I spend some time on overall lifestyle management, I'm much more likely to have cooperation from the patient even when I recommend pharmacologic therapy than initially when they come full of fear about medication. Nutrition has several aspects; one is adequate caloric intake. If there is caloric deficiency, there will be bone loss. If there is weight loss, there will be bone loss, and this is important. I have seen patients come to me because they have decrease in bone density on appropriate therapy. You take the history. And turns out, they went on a crash diet, they lost 10, 15% of their body weight, they will invariably have bone loss, even if they're taking appropriate pharmacologic therapy. So weight loss is not good for the bone. Beyond that, it's important to have adequate protein, calcium, and vitamin D intake. This is a real issue, especially in older, frail population. Often, they're calorically deprived. And one of the definitions of frailty's actually weight loss. They have tea and toast diet. They really do not get enough protein. There is sometimes excessive fear of animal protein being acidifying. And while that is correct, these little old ladies are very unlikely to have enough protein to have a problem with too much acid in their diet. In terms of calcium and vitamin D, this is a huge controversy. And I could spend the whole hour presenting studies on whether calcium is harmful to cardiovascular outcomes, whether it is preventing fractures, and so on. Given the uncertainty in the data and a lot the fashion in its interpretation, the best advice I can give is to provide what I would call reasonable amount of calcium, which is about 1,000 to 1,200 milligrams for postmenopausal women per day, preferably in food. Because if there is this association with cardiovascular outcomes, it seems to be really for supplements. So if possible, get it from food. And then only supplement as much as necessary to get to the appropriate amount. I also check PTH often. And if I have PTH that's 30, I know they are getting enough, even if it's not necessarily within guidelines. That's not the guideline. And then vitamin D, the level of vitamin D that we should achieve is also a debatable issue currently. There is no question that over 30 is enough. We saw on the graph that when vitamin D gets to be about 30, the PTH gets to be mid-normal. So that, I think, is the best data we have. Some people would say over 20 is already sufficient. I think if we're dealing with an osteoporotic population, especially the elderly, which are so often vitamin D deficient, we should shoot for over 30. How to get there depends on the baseline level. Obviously, if somebody's starting with a level five, they will need high dose vitamin D to get to the goal level. And it also depends on the weight of the person. Because Vitamin D is fat soluble, so it's distributed in the fat tissue. Obviously, a person who weighs 250 pounds is going to need higher daily dose than somebody who weighs 100 pounds, because the volume of distribution is higher. So generally, 800 to 1,000 for a normal weight person will usually be sufficient. In terms of activity, it is very important to counsel people both to not have too much and not have too little. I have seen patients who are very sedentary. And for them, I will advise walking, any kind of weight bearing meaning bearing their own weight. So just getting out of the house and being active is very helpful. On the other hand, I have patients who are incredibly aggressive. For instance, I had a patient who had two fractures while she was rollerblading at age 52. And then she said, should I be rollerblading? No, I don't think you should be rollerblading if you have osteoporosis and you're 52 years old. So it is important to ask what patients are doing, what they like to do, and then design a regimen that's going to be both enjoyable and likely to last, and provide some gravity stimulus. For instance, walking and running would be weight bearing, but biking and swimming are not. Actually, there is data that competitive bikers have lower bone density than controls, perhaps because they spend so much time without gravity stimulus. In addition to this weight bearing activity, other forms of exercise which improves strengths, posture, and balance, and does decrease fall risk also should be recommended. Again, according to the patient's preferences, I'll take a history what they've done in the past, look on the physical exam, how fit are they. And based on that suggestion, maybe working with a personal trainer to develop a strengthening regimen-- pilates, yoga. Tai Chi, actually, there is some data that it prevents fractures, because it does improve balance and prevent some falls. Pharmacologic therapies, the last component of our approach to osteoporosis. And I think it's best understood based on understanding the remodeling cycle of the bone. You will remember that bone is being constantly remodeled. That's how we can actually keep the same skeleton for 90 years without crumbling. There is no man-made structure that would be subjected to the loading and the stresses that our skeleton is subjected for so many years without suffering material damage. What is done in the bone to repair that materials damage is that when there is a microcrack or some poor quality bone, osteoclast, which is a bone dissolving cell, will basically chew up the bone and create this resorption lacunae. And then the osteoblasts will come in and generate new bone, which will then replace the previously damaged bone with a new, higher biological quality material. An interesting aside is about osteocyte. These osteoblasts that created new bone become trapped in the new bone as osteocytes. And we thought that it was just a graveyard for osteoblasts. But more recent data shows that these osteocytes are very important in the [INAUDIBLE] transduction in translating the mechanical stimulus to bone formation. And for instance, in glucocorticoids treatment, osteocytes are the primary cell that suffers, and thereby the bone lost. So the important part of this remodeling cycle is that in young people, in everybody actually, it takes about two weeks to create this resorption lacunae. But it takes three to six months to fill it in. So in young people, this is fine, because the activation frequency, which is how often a remodelling cycle is initiated in a given skeletal site, the activation frequency's about one to two per year. So you will basically maintain the constancy of the skeleton. Postmenopausally, however, everything speeds up. So there is certainly two weeks to make the resorption lacunae. But they're not six months available to fill it in. As a result, there is net bone loss because bone starts getting reabsorbed in adjacent area before we have time to fill in the resorption lacunae. Therefore, our entirely resorbative agents which interfere with osteoclast the bone resorption are very useful to prevent postmenopause bone loss. And they were indeed, developed in postmenopausal women. The other class of therapeutic agents that actually stimulate bone formation by directly stimulating osteoblasts, and these are called anabolic agents. And they're useful where bone formation is deficient such as glucocorticoid treatment, or let's say, somebody who has been on bisphosphonates for a very long time and bone formation is impaired. So our pharmacologic agents are then divided into antiresorptives and anabolic agents. Among antiresorptives, we have this phosphate's hormonal therapy with estrogen and serums. We have calcitonin, which is not very potent. So it's not used very much. And finally, denosumab. In terms of anabolic drugs, we only have one, which is teriperatide. Now, the problem is that bond resorption and formation are always coupled. So if we use an antiresorptive drug, we try to block the formation, we block the resorption, but we end up blocking the formation as well. Conversely, if we use an anabolic drug that stimulates formation, we will often stimulate resorption as well. So bisphosphonates are most widely used drugs and the different formulations are given on this slide. The advantages that are bone specific, they increase bone density. They have been documented to reduce fractures, and they're relatively easy to take. The disadvantage is that some patients will have GI side effects from moral bisphosphonates. And then the biggest problem that we have which has generated so much fear is the rare but widely publicized occurrence of ONJ, Osteonecrosis of the Jaw, and atypical femur fractures, which trace the question about over-suppression of bone turnover as being a problem after long them use. So often, necrosis of the jaw is something I actually only saw once over 10 years of practice. But it's very much feared by dentists, and is a source of a lot of fear in patients who avoid taking medication. Because of that, is described as exposed bone of the jaw that's exposed for over eight weeks and not healing. It's usually seen with hide dose intravenous bisphosphonates, much more rare with patients who take oral bisphosphonates for osteoporosis. A typical femur fracture's a low energy fractures below the trochanter. And that, I have seen a few, and they are becoming a fearful consequence of osteoporosis-- so much so that American Society of Bone and Mineral Research has convened a task force that was first reporting the findings in 2010 with a revision in 2014. These are the initial major and minor features that have been slightly revised on the more current version. And the big problem with this is that even though these events are rare, they are no good ways of predicting them, which has generated the fear. So because of these and also because of the fact that bisphosphonates stay bound to the skeleton for a number of years, bisphosphonate drop holiday has become a very important area of a clinical care. We actually have two trials that have addressed what happens with bone when bisphosphonates are stopped. The flex trial looked at women who took alendronate for five years and were then randomized to continue it for another five years, or stop take placebo. And the horizon extension trial looked at postmenopausal women who took once a year zoledronic acid for three years, and were then randomized to continue for another three years, or stop the drug and take a placebo. So what is really the important end point here is the fracture rates, not necessarily what happens to bone density. In both of these trials, we see that the overall fracture rates weren't significantly different, which some caveats. Vertebral fractures tended to be higher in those who stopped the drug, suggesting that patients who are very fragile-- because those are the patients who have vertebral fractures-- should not be considered for a drug holiday. And then those with low bone density also seem to do better when they were continuing the drug. So based on this, can we have some general guidelines about a drug holiday? Should be considered in patients who have had bisphosphonates for three to five years, there is some difference in bisphosphonates in that risedronate is less tightly bound than alendronate, and particularly zoledronic acid. So may want to shorten the drug holiday in those that have less tight binding. But generally, consider after three to five years, consider it definitely in patients who have low fracture risks. Let's say somebody was given a bisphosphonate early menopause to prevent bone loss. She's now five years postmenopausal. The bone density has stabilized. She doesn't have very high fracture risk that would be an ideal candidate. Those who have good BMD readings and those who do not have vertebral fractures. Now, when to restart is not really clear. It's also a matter of contention. But one approach that's been proposed is to actually use FRAX. The way you would go about starting the first time, you could use the same FRAX calculation to determine when to restart the bisphosphonate. And although FRAX is initially developed for patients who have not been on medical therapy, Bill Leslie-- and this is the reference in the bottom-- has looked at the ability of FRAX to predict fractures in patients who have been on medication versus those who haven't been and found that FRAX is equal useful in patients who have been on medication. So we can use FRAX. And what to use in those who do not qualify for drug holiday is a rarely data free zone. You could think about going back on a bisphosphonate, you could use maybe an anabolic drug, or another antiresorptive. Again, this is where individualized approach, knowing the patient's history, their bone density, perhaps their bone turnover, is quite useful. Estrogen is actually a good drop for osteoporosis, despite the many negative effects that have been observed in more recent studies. It increases bone density. It has been shown to reduce fractures, and is very useful in women who have menopausal symptoms. The disadvantage is increased breast cancer, cardiovascular [INAUDIBLE]. And the other important thing, it's very different from bisphosphonates. The effect of bisphosphonate stays for a long time after desktop. Estrogen, in fact, is lost immediately, and bone loss will occur unless prevented with another approach. So this is my idea of who should be treated with estrogen. This is by no means a practice guideline. I like to use estrogen in postmenopausal women who have hot flashes but low risk of breast cancer. In hypogonadal young women, estrogen is probably a much better choice than any of the other agents. And finally, it is a consideration in postmenopausal women who are intolerant to bisphosphonates, although we have other options now, as well. Serums are also working on the bone through estrogen receptor. Raloxifene is specifically approved for osteoporosis. Tamoxifen that's used for breast cancer is probably similar. It should be remembered that they're not as potent as estrogen. Estrogen and alendronate in a direct comparison trial are accrue potent. But serums are probably about 1/2 or 2/3 effective in terms of bone density. Now, the advantage is that they increase BMD. They show reduction in vertebral fractures, not non-vertebral. But the main advantage is a decreased risk of breast cancer. The disadvantage is that it's less potent than estrogen. That's for [INAUDIBLE]. And it increases hot flashes and also has thromboembolic risk. Now for estrogen, we can minimize thromboembolic risk by using transdermal preparation. Raloxifene is only available as a pill, so we don't have that option. So who would I treat with raloxifene? Postmenopausal women who have high risk of breast cancer but not did they have very low BMD and very high fracture risk, because the potency is not quite the same. And then also, postmenopausal women intolerant to bisphosphonate. Or let's say, on a drug holiday. Finally, the newest form of therapy for osteoporosis is based on RANK ligand. RANK ligand is a protein that's secreted from an osteoblast, which is necessary and sufficient for osteoclass differentiation and proliferation. So anything that leads to bone loss involves RANK ligand pathway. So this is the first biological agent for osteoporosis this was actually developed on the basis of understanding bone biology. So denosumab is an antibody to RANK ligand which binds to RANK ligand, and thereby prevents bone resorption. It's given as a sub-Q injection every six months. Advantage is that it has increase in bone density, it has documented efficacy in terms of fracture reduction. It can be used in renal insufficiency where bisphosphonates are contraindicated. And finally, it's effect wears off quickly when it's stopped, which could be an advantage if you don't want a long term effect, but could also be a disadvantage if the patient is, let's say, non-compliant. The disadvantage is again, the rapid loss of the fact in a non-compliant patient. It's relatively new. So it's been around for three years now, and it looks good. But there is always a concern. And reported in the lay press in particular is the increased risk of infection and malignancy. The signal is actually not very strong, and I'm providing the data here in terms of the increase in infection rates and skin infections in particular. So who should be treated with denosumab? No guidelines. Again, postmenopausal women with high risk of fracture, contraindication or intolerance to other agents, renal insufficiency, would be some of the examples. This slide shows how denosumab differs in terms of its on and off effect. So here we see denosumab given as 30 milligrams every three months, which is the same dose as 60 milligrams every six months. There is an increase in bone density, which is actually a little bit more than alendronate, which is given in the standard dose of 70 milligrams per week. But the difference is when the drug is stopped, there is a rapid decrease of the denosumab is stopped. First, it's a relative preservation when alendronate is stopped. Because as we said, it's retention in the bone is quite long. And this is showing the serum CTX, which is a marker of bone resorption. It's very potently suppressed by denosumab. But when the drug is stopped, it doesn't just come back to baseline. It actually overshoots. There is this rebound increase in bone resorption, which explains why we have this very rapid bone loss when the drug is stopped. This is something that needs to be kept in mind. When we think about this drug, it is not the bisphosphonate. The mechanism is very different, primarily in terms of its on and off effect. Finally, in terms of anabolic therapy, we only have one agent, which is teriparatide given by a daily sub-Q injection. It is the only drug that truly builds a bone, produces large increase in bone density and documented fracture reduction. The disadvantages are it's a daily injection-- the cost is high. There is an association with osteosarcoma in the rats, but really not in humans. And also, after the drug is stopped, there is rapid bone loss. So this drug needs to be followed by an antiresorptive agent to preserve the gains. So who would I treat with teriparatide? Patients with highest fracture risk and patients who are fracturing on therapy with bisphosphonates, institutionally. And then I would even consider it in patients who start treatment with very low BMD perhaps makes more sense to increase their bone mass, and then use an antiresorptive to protect it from being lost. In terms of future of osteoporosis therapy, there is very significant research going on. Two kinds of agents are in clinical trials. Nothing is even close to being submitted to FDA, let alone being clinically available. But the drugs that are currently under most research are inhibitors of cathepsin K. Cathepsin K is an enzyme that's produced by the osteoclast, and digests the collagen. So it's very important in bone resorption. So this is another antiresorptive agent. And you would say, well, why do we need more of the same? The difference is that cathepsin K inhibitors block the resorption but do not block the formation. So that's where their promises are really exciting. And then antibodies, the sclerostin-- one of them, romosozumab, has been described in a New England Journal paper from the beginning of 2014 showing remarkable results in terms of bone density. So this is something that we can look forward to in the future. And this completes my talk. So thank you very much for your attention.