INTRODUCTION — Many drugs can affect bone metabolism. As examples, heparin, warfarin, cyclosporine, glucocorticoids, medroxyprogesterone acetate, cancer drugs, and thyroid hormone can cause bone loss, whereas thiazide diuretics can minimize bone loss [1,2]. This topic will review the skeletal effects of some of these drugs. The effects of glucocorticoids, aromatase inhibitors, and thyroid hormone are discussed separately. (See "Clinical features and evaluation of glucocorticoid-induced osteoporosis" and "Evaluation and management of aromatase inhibitor-induced bone loss" and "Bone disease with hyperthyroidism and thyroid hormone therapy".)
DRUGS THAT MAY HAVE ADVERSE EFFECTS
Anticoagulants
Heparin — Heparin causes bone loss by decreasing bone formation. The few studies of the mechanisms underlying this bone loss have revealed decreased bone formation [3], increasing bone resorption, or both [4].
Since heparin is usually given for brief periods of time, its adverse effect on the skeleton should be trivial. However, it may be given for a prolonged period during pregnancy since warfarin is relatively contraindicated in the first trimester due to its teratogenic effects [5]. As a result, most of the information concerning the adverse effects of heparin on bone comes from studies of pregnant women requiring anticoagulation.
Chronic heparin therapy reduces bone mineral density (BMD) [6-8]. In one study, for example, in 14 heparin-treated pregnant women and 14 untreated pregnant women, mean hip density fell by approximately 5 percent in the women receiving heparin; in contrast, BMD did not change in the control group [6]. Hip density decreased by more than 10 percent in 5 of the 14 women taking heparin (36 percent) versus no women in the control group. Similar results were reported in another controlled study, which examined the effect of heparin on forearm density [7].
There are many case reports and series of pregnant women with osteoporotic fractures during and after prolonged heparin therapy [9]. One of the largest studies followed 184 pregnant women who were given heparin; four (2.2 percent) had osteoporotic vertebral fractures [9]. Although the incidence of fractures was low, they occurred in young women in whom osteoporotic fractures are extremely rare. (See "Use of anticoagulants during pregnancy and postpartum".)
Bone density increases postpartum after heparin is discontinued [6-8]. It is unclear, however, if recovery is complete.
Low molecular weight heparin — Low molecular weight heparins may have less of an adverse effect on bone than unfractionated heparin [10,11]. However, these trials were small, and a larger, prospective observational study found no difference between the two heparin preparations [12].
Warfarin — Warfarin decreases the tendency of blood to clot by inhibiting the vitamin K-dependent gamma-carboxylation of clotting factors II, VII, IX, and X [5]. Noncarboxylated clotting factors do not bind to calcium and, therefore, cannot participate in the coagulation cascade. Similarly, warfarin inhibits the gamma-carboxylation of osteocalcin, a major protein of bone that is involved in bone formation; noncarboxylated osteocalcin cannot to bind calcium effectively [13]. (See "Vitamin K and the synthesis and function of gamma-carboxyglutamic acid".)
As a result of the latter observation, it has been suspected that warfarin may adversely affect skeletal health. This hypothesis is indirectly supported by the following findings:
●Mean serum vitamin K concentrations in patients who have fractures are lower than in individuals without fracture history [14,15].
●The percentage of noncarboxylated osteocalcin in the serum of older women (who have a greater fracture risk) is higher than in young women [16].
●The risk of hip fracture in women with high serum concentrations of noncarboxylated osteocalcin is higher (relative risk [RR] 5.9) than in women with low concentrations [17].
●Treatment of postmenopausal women with vitamin K reduces urinary excretion of hydroxyproline, a marker of bone resorption [18].
The clinical importance of these observations is uncertain. Some cross-sectional studies have found that mean bone density in warfarin-treated patients was lower than in control patients [19-21], and a retrospective cohort study found that long-term exposure to warfarin was associated with an increased risk of vertebral and rib fractures [22]. Low bone density was also reported in a small case-control study of children on long-term warfarin versus control children [23]. In a retrospective cohort study of hospitalized patients with atrial fibrillation, men, but not women, who had taken warfarin for more than one year were at increased risk for osteoporotic fracture (odds ratio [OR] 1.63, 95% CI 1.26-2.10) [24].
However, in other studies, warfarin had no adverse effect on bone density [25,26] or fracture rates [27-29]. In a prospective observational study of women aged 65 years and older, for example, those taking warfarin (n = 149) and those not taking it (n = 6052) had similar rates of bone loss at the hip (1.1 and 0.8 percent, respectively) over two years and similar fracture rates over 3.5 years of follow-up [27].
Other oral anticoagulants do not work on the vitamin K pathway and would not be expected to have an adverse effect on BMD. Some [30-33], but not all [34], observational studies suggest that direct oral anticoagulants (DOACs) are associated with lower risk of fractures than warfarin. The reason for this difference in fracture risk, if real, is unknown. Nor is it clear whether it represents an increase in risk with warfarin or a decrease in risk with DOACs. If there is a causal association between warfarin and fracture risk, the effect is small. When anticoagulation is indicated, generally fracture risk is not a factor in the selection of an agent.
Cyclosporine — Abundant animal data suggest that cyclosporine adversely affects bone. Administration of cyclosporine to rats, for example, causes an increase in both bone resorption and bone loss, or "high-turnover" osteoporosis [35-37].
There is some evidence that cyclosporine may increase bone turnover in humans as it does in rats [38,39]. However, the effect of cyclosporine on bone metabolism in humans is less clear, being confounded by the presence of other illnesses or drugs that affect bone, particularly glucocorticoids (see "Osteoporosis after solid organ or stem cell transplantation"). As an example, one study evaluated BMD and bone histology in 20 living, related donor kidney transplant recipients treated with azathioprine, cyclosporine, and low-dose prednisone [38]. These patients lost 6.8 percent of their initial bone mass during the first six months after transplantation due to a low-turnover bone disorder resembling that induced by glucocorticoids. By 18 months, bone density had decreased 9 percent from baseline, and 60 percent of the patients had BMD values below the fracture threshold. (See "Kidney transplantation in adults: Persistent hyperparathyroidism after kidney transplantation".)
In contrast, another randomized study compared cyclosporine with placebo in patients with primary biliary cholangitis [39]. Patients receiving cyclosporine had evidence of increased bone turnover; however, cyclosporine decreased the bone loss that normally occurs in this disorder. (See "Evaluation and treatment of low bone mass in primary biliary cholangitis (primary biliary cirrhosis)".)
Thus, bone loss is common in many of the illnesses treated with cyclosporine, but the role of cyclosporine alone has not been determined. We currently do not recommend any specific treatment directed toward bone in patients taking cyclosporine.
Medroxyprogesterone acetate — The low doses (5 to 10 mg/day) of medroxyprogesterone acetate that are typically used in combination with estrogen as part of a regimen of postmenopausal hormone therapy have no effect on the ability of estrogen to prevent bone loss. This issue was best addressed in the Postmenopausal Estrogen/Progestin Interventions (PEPI) Trial in which 875 women were randomly assigned to placebo or one of four treatment arms consisting of conjugated estrogens (0.625 mg/day) alone or with three different progestin regimens [40]. At three years, BMD increased similarly in all treatment groups and fell in the placebo group. (See "Menopausal hormone therapy in the prevention and treatment of osteoporosis".)
In contrast, the higher doses of medroxyprogesterone acetate that have been used to treat gynecologic disorders or that have been given for contraception have been associated with increased bone loss, presumably due to the induction of estrogen deficiency. In 2004, the US Food and Drug Administration (FDA) required additional labeling recommending that depot medroxyprogesterone acetate be used as a long-term birth control method (eg, longer than two years) only if other birth control methods are inadequate and, if used long-term, BMD should be evaluated. This topic is reviewed in detail elsewhere. (See "Depot medroxyprogesterone acetate (DMPA): Efficacy, side effects, metabolic impact, and benefits", section on 'Reduction in bone mineral density'.)
Vitamin A and synthetic retinoids — Vitamin A is required for normal growth, vision, reproduction, cell proliferation, and cell differentiation. However, excess intake appears to increase the risk of hip fracture in women. The mechanism by which this may occur has been studied in animals, in which vitamin A inhibited osteoblast activity, stimulated osteoclast formation [41], and counteracted the ability of vitamin D to maintain normal serum calcium concentrations [42], thereby leading to accelerated bone resorption and fractures. The data in humans have been conflicting, but at least four studies have found an association between excess vitamin A intake and osteoporotic fracture:
●The first was a nested case-control study in Sweden, where both vitamin A intake and the incidence of hip fracture are high [43]. In this report, there was a 10 percent reduction in femoral BMD and a doubling in the risk of hip fracture (OR 2.1, 95% CI 1.1-4.0) among women who consumed more than 1500 mcg retinol per day compared with women who consumed less than 500 mcg per day. The risk was graded; for every 1 mg increase in daily retinol intake, the risk for hip fracture increased by 68 percent (95% CI 18-140 percent).
●The same group in Sweden reported similar findings in men. In a prospective cohort study of 2322 men followed for up to 30 years, the risk of any fracture or hip fracture was increased in those whose serum retinol concentrations were in the highest quintile as compared with the middle quintile (RR 1.6, 95% CI 1.1-2.4 and 2.5, 95% CI 1.1-5.3, respectively) [44].
●In a prospective analysis from the Nurses' Health Study, 603 women had incident hip fractures resulting from low or moderate trauma during 18 years of study [45]. After controlling for confounding factors, women in the highest quintile of vitamin A intake (≥3000 mcg/day of retinol equivalents) had an increased risk of hip fracture (RR 1.48, 95% CI 1.05-2.07) compared with women in the lowest quintile of intake (<1250 mcg/day of retinol equivalents). The association between high vitamin A intake and hip fracture was not seen with beta-carotene intake and was attenuated in women taking postmenopausal hormone therapy.
●In the Rancho Bernardo Study, a community-based cohort study of older individuals, both low and high vitamin A intakes were associated with a lower BMD. Bone density was optimal when vitamin A intake was 2000 to 2800 international units per day (0.6 to 0.9 mg/day) [46].
Not all studies find a relationship between vitamin A intake and bone loss. In an analysis from the National Health and Nutrition Examination Survey (NHANES) study, high serum retinyl ester concentrations (a marker of vitamin A intake) were not associated with a decrease in BMD [47]. Similarly, a case-control study of patients with incident osteoporotic fractures, serum indices of vitamin A exposure were not predictive of hip fracture or of lower BMD [48].
Based upon the available bone density and fracture data, individuals in Western countries should be cautioned against excessive vitamin A intake. Common food sources of vitamin A include liver, milk, egg yolk, butter, and some fruits and vegetables; those with diets high in these foods should avoid supplements containing vitamin A. (See "Overview of vitamin A", section on 'Requirements' and "Vitamin intake and disease prevention".)
On the other hand, vitamin A deficiency is still fairly common in African and Asian countries, and inadequate vitamin A intake may also increase fracture risk [49]. For populations in which vitamin A deficiency is endemic, supplementation is recommended. (See "Overview of vitamin A", section on 'Deficiency'.)
Loop diuretics — Loop diuretics increase calcium loss by impairing reabsorption in the loop of Henle (see "Diuretics and calcium balance", section on 'Loop of Henle and loop diuretics'). The ensuing negative calcium balance has been associated with a decrease in BMD and an increase in risk of hip fracture in some [50-52], but not all [53], studies. As examples:
●In a clinical trial of 87 healthy, postmenopausal women randomly assigned to receive bumetanide, a loop diuretic, or placebo for one year, urinary calcium and serum parathyroid hormone (PTH) increased (17 and 9 percent, respectively), and BMD decreased at the hip (-2 percent) and whole body (-1.4 percent) in the treatment group compared with placebo [50]. This detrimental effect on bone occurred despite supplementation with calcium and vitamin D.
●In a cohort study of 3269 men, the use of loop diuretics was also associated with a greater decline in total hip BMD compared with nonuse (annual rate of decline -0.78 versus -0.33 percent, respectively) [51].
●In a case-control study, the adjusted RR for hip fracture was 3.9 (95% CI 1.5-10.4) in patients taking furosemide compared with patients not taking furosemide [52].
●In contrast, in the Women's Health Initiative (WHI) observational study, there were no significant associations between use of loop diuretics and changes in BMD, falls, and fractures (hazard ratio [HR] for hip fractures in users versus never users 1.21, 95% CI 0.91-1.60) in models adjusted for confounding variables [53]. Women who used loop diuretics were in poorer health and had more risk factors for fracture than never users. However, one-third of loop diuretic users were taking hormone therapy, and mean calcium intake was 1000 mg daily, interventions that may have decreased fracture incidence.
The effect of loop diuretics is in contrast to that of thiazide diuretics, which reduce calcium excretion and may increase BMD. (See 'Thiazide diuretics' below.)
Chemotherapeutic drugs — Most cases of osteoporosis in patients treated for cancer are due to either the hypogonadism that results from chemotherapy and radiation therapy (from the blockade of estrogen synthesis with aromatase inhibitors), or to glucocorticoid therapy [54]. (See "Ovarian failure due to anticancer drugs and radiation" and "Evaluation and management of aromatase inhibitor-induced bone loss".)
However, direct negative skeletal effects have been described for certain chemotherapeutic drugs:
Methotrexate — High-dose methotrexate regimens (such as those used for osteosarcoma or acute lymphoblastic leukemia) are associated with an increase in bone resorption and an inhibition of bone formation, resulting in both osteoporosis and fractures [54,55]. However, this has only rarely been observed with methotrexate in the dose range used for rheumatic disease [56,57]. (See "Major side effects of low-dose methotrexate", section on 'Others'.)
Ifosfamide — Ifosfamide can damage the proximal tubules, causing metabolic acidosis, renal phosphate loss, hypercalciuria, and, in severe cases, hypophosphatemic osteomalacia. (See "Ifosfamide nephrotoxicity".)
Imatinib — Imatinib mesylate, a drug used in the treatment of chronic myeloid leukemia, gastrointestinal stromal tumors, and other malignancies, may be associated with changes in bone turnover [58,59]. In one preliminary report, patients receiving imatinib had increases in markers of bone formation (serum osteocalcin and N-terminal propeptide of type I procollagen [PINP]) but no change in a marker of bone resorption (beta isomer of C-terminal telopeptide of type I collagen) within three months of initiating therapy [59]. In some instances, changes in bone turnover were associated with low-normal serum calcium, secondary hyperparathyroidism, renal phosphate wasting, and hypophosphatemia [58,59]. The mechanism and clinical consequences of these findings are unclear. (See "Hypophosphatemia: Causes of hypophosphatemia".)
Antiseizure medications — Enzyme-inducing antiseizure medications (ASMs) such as phenobarbital, phenytoin, carbamazepine, and primidone increase the activity of P450 enzymes. Induction of the cytochrome P450 system by ASMs leads to increased catabolism of vitamin D to inactive metabolites and a subsequent rise in PTH, which increases the mobilization of bone calcium stores and subsequent bone turnover. In ambulatory patients, long-term antiseizure therapy has been associated with low bone density and an increase in fractures. The increase in fracture rate is due to both seizure-related injuries and the adverse effects of ASMs on bone strength. This topic is reviewed in detail separately. (See "Antiseizure medications and bone disease".)
Proton pump inhibitors — Insoluble calcium, such as calcium carbonate, requires an acid environment for optimal absorption. As a result, drugs that reduce stomach acid secretion (proton pump inhibitors [PPIs] and H2 blockers) may reduce calcium absorption. Because calcium absorption decreases with aging, a further reduction in calcium absorption with the addition of such drugs may have an adverse impact on skeletal health, particularly in older individuals.
Some [60], but not all [61,62], studies with PPIs show that fractional absorption of calcium is reduced in postmenopausal women. A possible explanation for the different results among studies is a difference in study conditions, including selection of isotope-labeled calcium (calcium carbonate capsules versus calcium chloride solution), method of measurement (fasting serum isotope-labeled calcium levels versus dual isotope measurements with a meal), duration (eg, 7 versus 30 days), type of PPI treatment (omeprazole versus a different PPI), and patient characteristics (mean age 76 versus 58 years). Dietary calcium (milk and cheese) absorption was not reduced in healthy individuals treated with omeprazole [63,64], suggesting that a meal induces a sufficient amount of acid secretion for calcium absorption despite PPI therapy.
The more important clinical question is whether PPIs affect fracture risk. In meta-analyses of case-control and cohort studies, the risk of hip, spine, and any-site fractures was modestly but significantly increased in patients taking PPIs (RRs 1.30, 1.56, and 1.16, respectively) [65-68]. In some studies [67,69], but not another [70], the risk was highest in long-term users of high-dose PPI therapy. In one analysis, the risk was confined to patients with at least one other risk factor for hip fracture [71], and in another, to current or former smokers [67].
The largest prospective cohort study (the WHI Study) did not find an association between PPI use and hip fracture (HR 1.00, 95% CI 0.71-1.40) [72]. However, PPI use was associated with an increased risk of clinical vertebral (HR 1.47, 95% CI 1.18-1.82), wrist, and total fractures. There was a smaller number of hip fracture events compared with wrist, clinical spine, or total fractures. The lower number of events, combined with the fact that PPI users were more likely than nonusers to be taking hormone therapy, may have reduced the ability of the study to detect an increased risk of hip fracture in PPI users.
H2 blockers were associated with an increased risk of hip fracture in some reports (assumption of risk [AOR] 1.23, 95% CI 1.14-1.39) [69,71,73] but decreased [70] or unchanged [68,72] risk in others.
In a subsequent analysis of a national prescription database, concurrent use of PPIs and alendronate compared with alendronate alone was associated with loss of protection against hip fracture (fracture risk reduction with alendronate 39 versus 19 percent in non-PPI versus PPI users) [74]. Concurrent treatment with H2 blockers did not modify the treatment response to alendronate.
Although the association between PPIs and fracture is plausible, these observational studies do not prove causality. One possible mechanism by which PPIs and H2 blockers adversely affect bone is through impaired absorption of calcium carbonate due to achlorhydria, which increases bone loss and reduces BMD. In a prospective cohort study, chronic PPI use was associated with lower baseline BMD at the femoral neck and total hip, but use over 10 years was not associated with accelerated BMD decline [75]. In addition, other studies have not found a decrease in BMD in PPI users compared with nonusers [76,77], albeit in one of these there was an increased risk of falls and fractures in PPI users [76]. Thus, factors independent of BMD (eg, frailty) may contribute to fracture risk, and PPI use may be a marker of frailty since PPI users are, as a group, sicker than controls [75]. Further studies investigating the relationship between PPIs and fracture are required.
In the interim, because omeprazole was shown to reduce the fractional absorption of calcium carbonate in fasting postmenopausal women in some studies [60], we advise that postmenopausal women taking long-term PPI or H2 blocker therapy increase dietary calcium and, when necessary, use calcium supplements that do not require acid for absorption, such as calcium citrate. The treatment of postmenopausal women with or at risk for osteoporotic fracture is reviewed separately. (See "Calcium and vitamin D supplementation in osteoporosis", section on 'Supplements' and "Overview of the management of osteoporosis in postmenopausal women", section on 'Pharmacologic therapy'.)
Antidepressants — Both tricyclic antidepressants (TCAs) and selective serotonin reuptake inhibitors (SSRIs) have been associated with an increased risk of fragility fracture in observational studies [78-80]. In meta-analyses of case-control and cohort studies, there was an increased risk of fracture (predominantly hip and nonvertebral fracture) in patients who had taken an SSRI (RR 1.72, 95% CI 1.51-1.95) [81,82]. Similarly, in a separate meta-analysis of case-control and cohort studies examining the association between TCAs and risk of fracture, there was an increased risk of fracture in users of TCAs compared with nonusers (adjusted RR 1.36, 95% CI 1.24-1.50) [83]. The rise in fracture risk was very rapid and decreased with time, suggesting that the effect was not due to a reduction in BMD. In both meta-analyses, the increased risk persisted after confining the analysis to studies that adjusted for important risk factors for fracture (age, comorbidities, medications known to increase fracture risk, BMD).
The results from these studies do not prove causality, and other confounders not controlled for during statistical analysis may explain the observed association. However, one mechanism to explain the possible association between antidepressants and fracture is the adverse effects of the drugs, including sedation and postural hypotension, which may increase risk of falls. Nevertheless, in some large cross-sectional and prospective cohort studies, use of SSRIs, but not TCAs, has been associated with reduced BMD in older men [84] and increased rates of bone loss at the hip in older women [85]. Thus, an alternative possibility is a direct effect of SSRIs on bone metabolism, although studies in this regard are preliminary and not supported by limited randomized controlled trials [86,87].
Prospective trials are needed to definitively determine the relationship between antidepressants and fracture. In the interim, we recommend that the risks of potential fracture be balanced against the benefits gained from treating depression, especially in older persons who are already at increased risk for osteoporosis and fracture. The decision to prescribe SSRIs versus TCAs should not be driven by fracture data, since there are only observational data that either may increase risk of fracture.
We also recommend counseling regarding fall prevention, adequate calcium and vitamin D supplementation, and smoking cessation, which are important lifestyle changes that may prevent fracture. BMD testing should be considered in patients receiving SSRIs, especially when other risk factors for fracture are present, such as advanced age or prior history of fragility fracture. (See "Osteoporotic fracture risk assessment" and "Overview of the management of osteoporosis in postmenopausal women" and "Prevention of osteoporosis", section on 'Calcium and vitamin D'.)
Thiazolidinediones — There is evidence suggesting that thiazolidinediones have an adverse impact on skeletal health, including an increased risk of fractures in women (humerus, hand, foot). (See "Thiazolidinediones in the treatment of type 2 diabetes mellitus", section on 'Safety'.)
Antiretroviral therapy — There is an increased risk of reduced bone density in human immunodeficiency virus (HIV)-infected individuals. Low BMD among patients with HIV infection is usually multifactorial. This topic is reviewed elsewhere. (See "Bone and calcium disorders in patients with HIV".)
DRUGS THAT MAY HAVE BENEFICIAL EFFECTS
Thiazide diuretics — Thiazide diuretics stimulate distal tubular reabsorption of calcium, leading to a decrease in urinary calcium excretion. This effect is used clinically to reduce the frequency of stone formation in patients with recurrent nephrolithiasis and hypercalciuria. (See "Diuretics and calcium balance" and "Kidney stones in adults: Prevention of recurrent kidney stones".)
As a result, thiazide diuretics may have a beneficial effect on bone mineral density (BMD). This is in contrast to loop diuretics, which increase calcium excretion and may reduce BMD. (See 'Loop diuretics' above.)
Bone mineral density — Long-term thiazide therapy may increase bone density and decrease fracture risk. In cross-sectional studies, patients taking a thiazide had a higher bone density than control patients [88,89]. A report of 9704 White women, for example, found that the forearm and calcaneus density was 4 to 5 percent higher in those taking a thiazide, even after correcting for other confounding variables such as body mass index and calcium intake [88].
There are also clinical trial data that demonstrate a beneficial effect of thiazide diuretics on BMD:
●In a controlled trial of 184 normal postmenopausal women, hydrochlorothiazide (50 mg/day for two years) resulted in a significant increase in forearm but not lumbar spine or hip density [90].
●A three-year trial of 320 healthy individuals aged 60 to 79 years examined the effect of hydrochlorothiazide (12.5 or 25 mg/day) versus placebo. In women, the hip and spine density was 1.4 and 1.3 percent higher, respectively, in the 25 mg-group compared with the placebo group [91]; the effect of the 12.5 mg dose was smaller. Hydrochlorothiazide also prevented bone loss in men, although the effect was less than in women (0.7 and 0.4 percent versus placebo in the hip and spine, respectively) [91].
Fracture risk — Most [88,92-98], but not all [52,99], observational studies have shown a beneficial effect of thiazides on fracture risk. A 2011 meta-analysis of 21 observational studies involving almost 400,000 individuals noted a significant reduction in risk of hip fracture with long-term thiazide therapy (RR 0.76, 95% CI 0.64-0.89) [96].
Although a randomized clinical trial has not been performed, we conclude that thiazide therapy attenuates bone loss and may reduce fracture risk, but the effect is modest. We do not routinely recommend the administration of thiazides to prevent or treat osteoporosis, but a thiazide diuretic is a reasonable choice if a patient with osteoporosis has hypertension or nephrolithiasis.
Statins — It has been thought that 3-hydroxy-3-methylglutaryl coenzyme-A (HMG-CoA) reductase inhibitors (statins), which are used for the treatment of hypercholesterolemia, may have beneficial effects on the skeleton, but data are conflicting. In some studies, statin therapy was associated with a decrease in fractures or an increase in bone density [100-104], but in others, there was no decrease in fractures [105-107]. Many of the studies included men (sometimes in a majority), the particular statin varied, and the design of the studies varied. Most were case-control or retrospective studies.
The Women's Health Initiative (WHI) observational study, the largest observational study to address this issue, reported no reduction in fracture rates in postmenopausal statin users compared with nonusers after four years of follow-up [108]. (See "Overview of the management of osteoporosis in postmenopausal women".)
In a meta-analysis of eight observational studies that reported statin use and documented fracture outcomes in postmenopausal women (including the WHI observational study noted above), statin use was associated with a lower risk of hip fracture (odds ratio [OR] 0.43, 95% CI 0.25-0.75 for users versus nonusers) [109]. However, post hoc analyses of two cardiovascular trials did not support a protective effect [105,107].
In the one clinical trial to date, 82 postmenopausal women were randomly assigned to simvastatin (40 mg/day) or placebo for one year. There were no effects of simvastatin on biochemical bone markers or on BMD at the hip or spine, although an increase in bone density was seen in the forearm. Thus, observational data on statin use and beneficial effects on bone health have been conflicting, but in the one randomized clinical trial to address this, simvastatin did not appear to have an overall beneficial effect on bone [110].
Nitrates — Nitric oxide slows bone remodeling and bone loss in animals [111]. In observational studies in humans, nitrate administration, which increases nitric oxide levels, was associated with an increase in BMD [112] and a reduction in fracture risk [113]. However, a randomized trial did not show benefit. In a three-year trial of nitroglycerin ointment (22.5 mg/day) versus placebo in 186 postmenopausal women without osteoporosis (BMD T-scores between 0 and -2.5 and no prior vertebral or hip fracture), women in both groups had similar reductions in lumbar spine and total hip BMD after 36 months (-2.06 versus -2.48 percent and -4.38 versus -4.03 percent for lumbar spine and total hip, respectively) [114]. The trial was not powered to assess fracture outcomes. Headaches were a common side effect, occurring in 57 percent of women in the nitroglycerin group.
Beta blockers — In animal models, beta blockers stimulate bone formation and inhibit bone resorption [115]. Data in humans are inconsistent, with some studies showing a lower risk of fracture with current use of beta blockers [116,117] and another showing no effect [118]. A meta-analysis of 16 observational studies showed a significant reduction in fracture risk in patients receiving beta blockers, a finding primarily driven by a reduction in risk of hip fracture [119]. The meta-analysis was limited by significant heterogeneity in the results.
Given the limitations of the observational data and the absence of clinical trial data, beta blockers have no current role in osteoporosis management.
SUMMARY
●Drug-related skeletal effects – There are several drugs that have been associated with a decrease in bone mineral density (BMD) and an increase in the risk of fracture, whereas other drugs appear to improve BMD and decrease the risk of fracture. (See 'Drugs that may have adverse effects' above and 'Drugs that may have beneficial effects' above.)
•Drugs that may have adverse effects on bone – In observational studies, commonly used medications, such as proton pump inhibitors (PPIs), antidepressants, anticoagulants, and loop diuretics have been associated with an increased risk of fracture. (See 'Drugs that may have adverse effects' above.)
When these medications are given for a brief period of time, the adverse skeletal effects are of less concern. Because long-term therapy may have detrimental skeletal effects, older patients should be evaluated for other risk factors for osteoporosis and fracture. BMD testing should be considered in some patients, especially when other risk factors for fracture are present, such as advanced age or prior history of fragility fracture. Older patients should receive counseling regarding fall prevention, adequate calcium and vitamin D supplementation, and smoking cessation, which are important lifestyle changes that may prevent fracture. (See "Osteoporotic fracture risk assessment" and "Clinical manifestations, diagnosis, and evaluation of osteoporosis in postmenopausal women" and "Clinical manifestations, diagnosis, and evaluation of osteoporosis in men".)
•Drugs that may have beneficial effects on bone – Long-term thiazide therapy may increase bone density and decrease fracture risk. Data are inconsistent for the effects of statins, nitrates, and beta blockers on BMD and fracture risk. None of these medications are recommended specifically for skeletal protection. (See 'Drugs that may have beneficial effects' above.)
●Glucocorticoid-mediated effects on bone – The skeletal effects of glucocorticoids are reviewed separately. (See "Prevention and treatment of glucocorticoid-induced osteoporosis".)
