Bisphosphonates, Statins, Osteoporosis, and Atherosclerosis

Nelson Watts, MD


South Med J. 2002;95(6) 


Both osteoporosis and atherosclerosis increase in frequency with advancing age, both appear worse (or at least more frequent) with estrogen deficiency, and both involve calcification. A number of mechanisms are common to both. Oxidative stress is part of the pathogenesis of both osteoporosis and atherosclerosis.[1] Matrix vesicle formation and mineralization, interleukin-6, and atherogenic diet all cause atherogenesis and also cause bone loss.[2,3,4] Excessive amounts of vitamin D contribute to both atherosclerosis and osteoporosis.[5,6] There are differences, however. Clearly, there is a difference in sex predominance: osteoporosis affects women more often than men, and atherosclerosis affects men more often than women.

At least 2 observational studies suggest a relationship between osteoporosis and atherosclerosis. A 1998 study showed a significant correlation between coronary artery calcification and bone mineral density (BMD).[7] Another study, a year earlier, showed a significant correlation between carotid-plaque score and BMD.[8]

At least 2 clinical trials suggest there may not be a relationship. Vascular events were uncommon in the Multiple Outcomes of Raloxifene Evaluation (MORE) trial for osteoporosis; the rates of coronary events and stroke were extremely low, both in the placebo group and in the raloxifene group.[9] This may be due to a selection bias. The women in the Heart and Estrogen/Progestin Replacement Study (HERS) trial, recruited because they had atherosclerosis, did not have osteoporosis.[10]

From the literature, it is not clear whether there is a strong or significant relationship between these 2 disease processes. It is my impression, from many years of treating patients with osteoporosis, that the likelihood of a heart attack or stroke among women with osteoporosis appears to be low.

Bisphosphonates are the agents that have the strongest evidence for reducing the risk of a variety of fractures in osteoporosis. What is their mechanism of action? The bisphosphonates bind to bone surfaces. When the osteoclasts settle on bone surfaces and seal off the area underneath, they mobilize the bisphosphonate from the surface of bone, and bisphosphonates then enter the cell. Resorptive functions of osteoclasts are then diminished, and eventually the osteoclast undergoes apoptosis (programmed cell death).

The non-nitrogen-containing bisphosphonates interfere with osteoclast function in a straightforward way. Etidronate and clodronate are incorporated into adenosine triphosphate (ATP) as nonhydrolyzable toxic analogs. The enzymes that cleave an oxygen-phosphate bond cannot cleave a carbon-phosphate bond, so these compounds build up and cannot be used for energy. If clodronate is added to macrophage-like cells or human osteosarcoma cells, these toxic compounds are formed. Adding these compound in tissue culture causes both programmed cell death (apoptosis) and necrotic cell death.[11]

The bisphosphonates that are commonly used to treat osteoporosis -- alendronate, risedronate, and other nitrogen-containing bisphosphonates -- do not make compounds like this, so they must work in some other way.

Many G proteins require the addition of either a 20-carbon side chain or a 15-carbon side chain to perform their function. These proteins perform important functions in osteoclasts: Rho is involved with cytoskeletal organization and apoptosis; Rac for membrane ruffling and endocytosis; Rab for membrane trafficking and vesicle transport; Ras for cell proliferation and apoptosis; and Lamin B for organization of nuclear lamina. If these proteins do not function effectively, cells will die. To function effectively, these G proteins need 15-carbon or 20-carbon side chains, which are normally added though a process called prenylation.

Looking for products involved in prenylation, in the control cells there are plenty of these 15- and 20-carbon fragments; in cells treated with clodronate and other non-nitrogen bisphosphonates show adequate amounts of these proteins, but in cells pretreated with alendronate, risedronate, ibandronate, there are few or none of these side chains.[12] What does this have to do with atherosclerosis and statins (hydroxymethylglutaryl coenzyme A [HMG-CoA] reductase inhibitors)? These side chains -- the 15-carbon side chain is a farnesyl side chain, and the 20-carbon side chain is a geraynlgeranyl side chain -- are produced through the HMG-CoA/mevalonate pathway, the same pathway that is inhibited by the statins, drugs that inhibit HMG-CoA reductase. The bisphosphonates that we use to treat osteoporosis work more distally in this pathway to inhibit the production of the 15- and 20-carbon side chains. There is interesting basic science work on this. Rabbit osteoclasts will resorb a certain amount of bone in vitro. Addition of mevalonate (the production of which is blocked by statins) does not change the resorptive capacity of the osteoclasts. Addition of geranylgeraniol (the 20-carbon side chain) does not change the resorptive capacity of the osteoclasts. However, if alendronate, a very potent antiresorptive drug, is added, it decreases the resorptive capacity of the osteoclasts by about 70%.[13] If mevalonate is added, alendronate continues to have an antiresorptive effect, but addition of geranylgeraniol completely reverses the antiresorptive effect of alendronate. In this same tissue system, addition of statin, such as mevalonate, has a striking antiresorptive effect, which can be abolished by the addition of mevalonate or geranylgeraniol.[13] The same is true for the number of osteoclasts in this system. Osteoclasts are happy if left alone. They continue to be happy if some mevalonate or geranylgeraniol is added. If alendronate is added, however, they disappear. If mevalonate is added, a few of them come back. If geranylgeraniol is added, this process is partially reversed. If lovastatin is added, osteoclasts disappear. If mevalonate is added, they come back. So the nitrogen-containing bisphosphonates work to decrease osteoclastic bone resorption by decreasing the resorption by individual cells, decreasing the number of osteoclasts, and increasing apoptosis.[13] Statins have the same effect in vitro, but it is hard to get a bisphosphonate into a liver cell. If bisphosphonates could be introduced into liver cells, they might lower cholesterol, but it might also kill the liver cells. It is hard to get a statin into bone cells, but if it could be done, it should have antiresorptive effects.

Statins are inhibitors of HMG-CoA reductase, which is the rate-limiting step in hepatic cholesterol synthesis. Statins reduce serum cholesterol, and they also have a variety of other putative effects: vasodilatory, antithrombotic, antioxidant, anti-inflammatory, and antiproliferative effects. For the reader intrigued by the lack of a clear relationship in osteoporosis treatment, where bone mineral density and antifracture effects do not seem to correlate, there is a huge body of literature on the effectiveness of statins in reducing cardiac events that seems to be distinct from their ability to change cholesterol levels. For a variety of reasons, statins are some of the most widely prescribed drugs in Western countries, and over 3 million Americans are regular users.

There is an excellent review article on this topic in Bone.[14] Statins mostly exist as prodrugs that have to be converted to active form. They are targeted to the liver, not to bone. They are metabolized by cytochrome P-450; some of the metabolites are active. The original statins were natural extracts, but the recent ones are synthetic. The newer ones are more powerful and have different pharmacologic properties. Most are lipid-soluble, which means that they enter cells easily. From a point of view of possible effect of statins in bone, getting into the bone cell is what it is all about. Some of the newer statins, particularly pravastain and robuvastatin, are water-soluble, and depend on specific carrier mechanisms to enter liver cells. It is unlikely that these same carrier mechanisms are present in the bone cells.

What might statins do to bone? Overall, the uptake of statins by bone is low; they are extracted by the liver. There is a first-pass metabolism; if a statin is taken by mouth, most of it is taken up by the liver, so there is little available to exert a systemic effect. Some statins appear to have little or no action in bone, but one of the newer statins, cerivastatin (atorvastatin), appears to be more effective in bone than lovastatin.

Why use lovastatin as a comparator? We have talked about the HMG-CoA/mevalonate pathway, by which the bisphosphonates reduce bone resorption. That pathway is also influenced by statins. There is also a possibility that statins may have an anabolic effect in bone, however. Bone morphogenetic protein-2 (BMP-2) is a potent anabolic agent. Mundy et al[15] screened over 3,000 compounds for their ability to promote the synthesis of BMP-2 and found that lovastatin was the most powerful in promoting the synthesis of BMP-2. Lovastatin had a dramatic anabolic effect in vitro and in vivo. The anabolic effect is prolonged after brief exposure. The anabolic effects of statins in bone can be prevented by noggin, which inhibits BMP-2. The evidence is strong that if certain statins get into bone cells, they can have an anabolic effect by increasing production of BMP-2. In Mundy's experiments, the cells that increased in number in animals treated with high doses of lovastatin were active osteoblasts. There was a dramatic increase in bone formation after just a few days. After several weeks, there was no real change in the control animals, but a dramatic increase in bone formation induced by lovastatin.

It did not take long for people to realize that there were repositories of data in humans that would allow some insight into whether statins had an effect on bone mass or fractures. About 6 months after the Mundy paper was published, three studies[16,17,18] suggested that there might be an association between statin use and reduction in the incidence of fracture.

In the first study, Wang et al[16] looked at the risk of hip fracture in elderly patients in a case-control study done in New Jersey residents aged 65 years or older who participated in Medicare, Medicaid, or pharmacy-assistance programs for the aged and disabled. More than 1,200 patients with hip fractures were matched for age and sex with 4 controls. About 40% of the hip-fracture patients were between 75 and 84 years of age and about 40% were over the age of 85 years, 83% were women, 85% to 90% were white, a small number were using estrogen, about 15% were users of thiazide diuretics, and about 4% were users of corticosteroids. They found a statistically significant 70% decrease in the likelihood of hip fracture in current statin users compared with controls. Those who had stopped using statins for 180 days had a 50% reduction in fracture risk, and those who had used statins within the previous 3 years had about a 45% reduction in fracture risk.

There were some other interesting observations in this study. Users of estrogen did not seem to be protected against hip fracture; users of thiazide diuretics had a slight protective effect (not statistically significant). Nonwhite subjects (most of the nonwhite subjects were African American) had a significantly lower rate of hip fracture than whites, which was not unexpected. A high risk of hip fracture was associated with a high comorbidity score. Spending time in the hospital or a nursing home in the previous 180 days was associated with a 50% to 70% increase in the risk of hip fracture.

A second study was a population-based, nested, case-control analysis drawn from the United Kingdom (UK) General Practice Research Database.[17] Of more than 90,000 individuals over the age of 50 years, more than 28,000 were taking lipid-lowering drugs, 13,000 were untreated hyperlipidemic patients, and 50,000 were randomly selected individuals without hyperlipidemia. From this group, almost 4,000 patients with fractures were matched with more than 23,000 control subjects for age, sex, area of residence, and years since enrollment in the health system. About 60% of the subjects were between ages 60 and 79 years, 75% were women, 50% were non-smokers, and 13% were current smokers. Of subjects whose body mass index was known, 37% were overweight and 17% were obese. A small number were current estrogen users. Patients who had received 1 to 4 prescriptions for statins had half the risk of any fracture compared with controls. The reduction in fractures was similar for higher numbers of prescriptions. People currently using statins had the best protection. Those who had stopped them in the past 3 months were still protected. Those who had used them more than 3 months in the past were not protected. There was a reduction in the incidence of all fracture types, most strikingly in femoral fractures and vertebral fractures, less so in upper extremity, foot, and other sites of fracture.

In the third study, Chan[18] found no effect overall, but fewer fractures in patients who had filled their most recent statin prescriptions.

This all sounds good. It did not take long, however, for other investigators to find some other databases with different messages. A retrospective observational study from Van Staa et al[19] used the same UK General Practice research database that was used by Meier et al[17] -- same database, different analysis, different conclusion. These investigators looked at more than 80,000 patients over the age of 50 years who had had a fracture that might be related to osteoporosis between 1987 and 1999. They matched them to an equal number of control subjects without fractures, then looked at the incidence of fractures in current statin users compared with nonusers. There were essentially no differences in any of their comparisons. The use of statins was similar between the cases and controls: 145 of the patients with fractures used statins and 109 of the control subjects used statins, so there was no statistical difference. Current statin use was similar. The duration of statin use, current dose, and cumulative dose were not different. There was no difference in type of statin used; most of the subjects in the study were using simvastatin, second most used was pravastatin, and only a small number were using other statins. In current statin users compared with control subjects, there was no difference for any fracture types. In hyperchlolesterolemic users of statin versus other drugs, there was no difference. The conclusion from these data is that the use of statins is not associated with reduction in the incidence of fractures. The data may be confounded by better reporting of fractures in statin users, protective effect of being overweight, and prescriptions that were written for the patient but not filled.

There was a randomized, placebo-controlled, double-blind trial of pravastatin for atherosclerosis. Reid et al[20] analyzed and reported that data from the point of view of fractures. Mean follow-up was 6 years. Fractures were ascertained from adverse-event reports. Baseline characteristics of placebo and treated groups were similar for age, sex, smoking, use of hormone replacement therapy, body weight, height, and body mass index. The numbers of fractures were essentially identical in the placebo group and the pravastatin group. Had there been just a 30% decrease in fracture incidence, there was good statistical power to see it. Pravastatin is water-soluble, which means it does not enter cells easily; of all the statins, it is the least likely to have an effect on bone.

More negative data come from the Women's Health Initiative (WHI). The final report from the main studies will be out in 2005. A group of investigators within the WHI looked at the observational cohort (more than 90,000 women) and found absolutely no difference in fractures comparing women who were receiving statins with women not receiving statins.[21] Statins may have a positive effect on bone: they could be antiresorptive, they could be anabolic. Effects will probably differ from one statin to another. If a particular statin has a positive effect on bone, it might be an interesting side benefit, but whatever effect there is may not be as good as the existing osteoporosis drugs. Randomized, prospective trials are needed, not only to show that statins or a particular statin has an effect on bone, but also that the effect of the statin on bone is on the same order of magnitude as the effect of the drugs that we use specifically to treat osteoporosis.

How do we pull all of this information together? It is interesting, but not something that can be applied in the clinic tomorrow. There may be a relationship between the propensity toward these 2 diseases, osteoporosis and atherosclerosis. This is an area that begs for more research. Bisphosphonates have effects as antiresorptive drugs on the HMG-CoA pathway, where statins also work. In addition to possibly being antiresorptive, statins may also be anabolic for bone by increasing the production of BMP-2. Whether it is possible to deliver sufficient statin into bone in intact humans remains to be seen.

Presented as the Keynote Guest Lecture at the Fifth Annual Southern Medical Association Conference on Osteoporosis, Amelia Island, Fla, February 21-24, 2002.

The print version of this article was originally certified for CME credit. For accreditation details, contact the publisher: Southern Medical Assocation, 35 Lakeshore Dr, Birmingham Alabama 35209, telephone: (205)945-1840.

In publishing this section in Southern Medical Journal, the Southern Medical Association recognizes educational needs of physicians in all specialties, especially those in primary care, for current information regarding the diagnosis and treatment of osteoporosis. In this section, authors may have included discussions about drug interventions, whether Food and Drug Administration approved or unapproved. Therefore, it is incumbent on physicians reading this section to be aware of these factors in interpreting the contents and evaluating recommendations. Moreover, views of authors do not necessarily reflect the opinions of the Southern Medical Association. Every effort has been made to encourage the author to disclose any commercial relationships or personal benefit that may be associated with this section. If the author disclosed a relationship, it is indicated below. This disclosure in no way implies that the information presented is biased or of lesser quality, but allows participants to make informed judgments regarding program content.

Ronald C. Hamdy, MD, FRCP, FACP

Grant/Research Support: Several pharmaceutical companies, including Merck, Procter &Gamble, Aventis, and Novartis

Consultant: Several pharmaceutical companies, including Merck, Procter &Gamble, Aventis, and Novartis

Speaker's Bureau: Several pharmaceutical companies, including Merck, Procter &Gamble, Aventis, and Novartis

Stock Shareholder: Several pharmaceutical companies, including Merck, Procter &Gamble, Aventis, and Novartis

Other Support: Several pharmaceutical companies, including Merck, Procter &Gamble, Aventis, and Novartis


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.