Calcium supplements

Calcium supplementation has long been a key part of osteoporosis treatment and prevention in an ageing population. Calcium replacement, when given without vitamin D, has been shown to result in increased BMD measurements at all sites.Reference Reid, Mason, Horne, Ames, Reid, Bava, Bolland and Gamble^9, Reference Shea, Wells, Cranney, Zytaruk, Robinson, Griffith, Ortiz, Peterson, Adachi, Tugwell and Guyatt¹⁰ However, evidence regarding the efficacy of calcium mono-therapy in fracture prevention is equivocal. Two major meta-analyses have failed to demonstrate a significant anti-fracture effect with calcium supplementation.Reference Shea, Wells, Cranney, Zytaruk, Robinson, Griffith, Ortiz, Peterson, Adachi, Tugwell and Guyatt^10, Reference Bischoff-Ferrari, Dawson-Hughes, Baron, Burckhardt, Li, Spiegelman, Specker, Orav, Wong, Staehelin, O'Reilly, Kiel and Willett¹¹ Bischoff-Ferrari's 2007 meta-analysisReference Bischoff-Ferrari, Dawson-Hughes, Baron, Burckhardt, Li, Spiegelman, Specker, Orav, Wong, Staehelin, O'Reilly, Kiel and Willett¹¹ considered the results of five prospective cohort studies of calcium supplementation or increased dietary calcium intake in men (n = 68,606) and seven such studies in women (n = 170,991), but demonstrated no significant effect upon fracture rates. A 2007 meta-analysis by Tang et al.Reference Tang, Eslick, Nowson, Smith and Bensoussan¹² did, however, demonstrate some anti-fracture effect with calcium supplements. In this analysis, the results of 29 randomized controlled trials (RCTs) were pooled to assess the effect of calcium with or without vitamin D upon change in BMD and fracture rates. The relative risk of fracture with calcium supplementation alone was 0.87, and with combined calcium and vitamin D supplementation it was 0.90 (P = 0.63). Specifically regarding hip fractures, another meta-analysis actually demonstrated an increased incidence with calcium mono-therapy compared with placebo.Reference Reid, Bolland and Grey¹³ The authors postulate that increased circulating calcium levels could suppress parathyroid hormone levels, thereby reducing bone turnover, which may limit the natural expansion in femoral neck size, which could account for an increased risk of fracture at the hip but reduced fracture rates at other sites.

In 2010, a BMJ article by Bolland et al.Reference Bolland, Avenell, Baron, Grey, MacLennan, Gamble and Reid¹⁴ attracted significant medical and media attention when it suggested a link between calcium supplementation and increased rates of serious cardiovascular events. They performed a meta-analysis on the results of 15 studies encompassing a total of 8151 subjects, which had included data on myocardial infarction, stroke, sudden death and all-cause mortality with calcium supplementation. They had a combination of patient level and trial level data but none of the studies included was actually designed to look at vascular events as an end-point. Pooling the trial level data, they reported a statistically significant near 30% increased risk of myocardial infarction with calcium supplements compared with placebo. However, calcium supplementation without vitamin D is now relatively uncommon in the treatment of osteoporosis. Reassuringly, analysis of the results from the Women's Health Initiative in which 36,282 post-menopausal women were randomized to calcium plus vitamin D or placebo for 7 years, showed no significant increase in the risk of vascular events with combined supplementation.Reference Hsia, Geraldo and Hong¹⁵

Vitamin D supplements

Vitamin D deficiency and its management is a topical and contentious issue. Whilst some studies have suggested that more than half of the adult population in the United Kingdom may be vitamin D deficient,Reference Hyponen and Power¹⁶ there is no clear consensus on what levels constitute deficiency. Vitamin D deficiency, in addition to its known effects on bone, is being linked to a wide range of other pathologies including some cancers, metabolic conditions, infectious diseases and autoimmune conditions.Reference Holick¹⁷ This section will review the contradictory body of evidence on the use of vitamin D in the treatment of osteoporosis.

There is some evidence from RCTs supporting the beneficial role of intermittent high-dose oral vitamin DReference Trivedi, Doll and Khaw¹⁸ and annual intra-muscular vitamin DReference Heikinheimo, Inkovaara, Harju, Haavisto, Kaarela, Kataja, Kokko, Kolho and Rajala¹⁹ in reducing fracture rates. Meta-analyses have reported that daily oral vitamin D supplementation (700–800 IU daily) reduces the risk of hip fracture by 26% and non-vertebral fractures by 23% when compared with placebo or calcium supplementation alone.Reference Bischoff-Ferrari, Willett, Wong, Stuck, Staehelin, Orav, Thoma, Kiel and Henschkowski²⁰ However, there is increasing evidence casting doubt on these findings. RCTs looking at daily oral vitamin DReference Meyer, Smedshaug, Kvaavik, Falch, Tverdal and Pedersen²¹ and high-dose oral vitamin D every 3 or 4 monthsReference Law, Withers, Morris and Anderson^22, Reference Lyons, Johansen, Brophy, Newcombe, Phillips, Lervy, Evans, Wareham and Stone²³ failed to show fracture reductions in older people residing in institutional care facilities. Some meta-analyses have also failed to demonstrate any significant anti-fracture effect with vitamin D supplementation.Reference Avenell, Gillespie, Gillespie and O'Connell^{24, 25} Additionally, some have actually shown a non-significant trend towards increased rates of fracture with vitamin D supplementation.Reference Lips, Graafmans, Ooms, Bezemer and Bouter^26, Reference Smith, Anderson, Raphael, Maslin, Crozier and Cooper²⁷

In 2010, Sanders et al.Reference Sanders, Stuart, Williamson, Simpson, Kotowicz, Young and Nicholson²⁸ published the results of a double-blind placebo controlled trial in which over 2000 Australian women aged 70 years or older (mean age 76 years) were randomized to high dose (500,000 IU) oral vitamin D every autumn for 3–5 years or placebo. They reported 15% more falls and 26% more fractures in the vitamin D group, both statistically significant. A temporal relationship between the supplementation and falls risk was also noted. The participants in this study were relatively vitamin D replete: median 25-hydroxycholecalciferol levels at baseline were 49 nmol/l, probably due to good sunlight exposure. Dawson-Hughes and HarrisReference Dawson-Hughes and Harris²⁹ postulate that the high-dose vitamin D used in this study may have caused up-regulation of the enzyme CYP24 (25-hydroxyvitamin D-24-hydroxylase), which inactivates 1,25-hydroxyvitamin D, thereby resulting in a paradoxical reduction in the active form.

In summary, evidence for the role of vitamin D alone in osteoporosis treatment is contradictory and we would suggest that, at present, this treatment should be used in a targeted way for those osteoporotic older patients with proven vitamin D deficiency, and in whom additional calcium supplementation is inappropriate.

Combined calcium and vitamin D supplementation

In a seminal study by Chapuy et al.,Reference Chapuy, Arlot, Duboeuf, Brun, Crouzet, Arnaud, Delmas and Meunier³⁰ over 3000 ambulant women in French nursing homes (mean age 84 years) were randomized to calcium (1.2 g as phosphate) plus vitamin D (800 IU) daily or placebo. In the women who completed the 18-month study, rates of hip fracture were 43% lower (P = 0.043) and non-vertebral fracture rates were 32% lower (P = 0.015) in the intervention group. It is likely that the overall high fracture rates and low baseline vitamin D levels in this population enhanced the benefit of combined calcium and vitamin D supplements. Whilst the benefit of this intervention may be less in community-dwelling older people, this study still has relevance for very old institutionalized patients.

The Randomized Evaluation of Calcium Or vitamin D (RECORD) trial³¹ studied calcium (1 g daily) and vitamin D (800 IU daily) in the secondary prevention of osteoporotic fractures. More than 5000 people aged over 70 years who had sustained a recent low-trauma non-vertebral fracture were randomized to one of four arms (calcium plus vitamin D, calcium only, vitamin D only, or double placebo) but there was no significant difference in rates of further fracture between any of the groups. However, when these data were combined with the results of eight other primary prevention RCTs (including outcomes from the large Women's Health InitiativeReference Jackson, LaCroix, Gass, Wallace, Robbins, Lewis, Bassford, Beresford, Black, Blanchette, Bonds, Brunner, Brzyski, Caan, Cauley, Chlebowski, Cummings, Granek, Hays, Heiss, Hendrix, Howard, Hsia, Hubbell, Johnson, Judd, Kotchen, Kuller, Langer, Lasser, Limacher, Ludlam, Manson, Margolis, McGowan, Ockene, O'Sullivan, Phillips, Prentice, Sarto, Stefanick, Van Horn, Wactawski-Wende, Whitlock, Anderson, Assaf and Barad³²), meta-analysis demonstrated significantly lower fracture rates. Hip fracture rates were 18% lower with combined supplementation compared with placebo, and non-vertebral fracture rates were 12% lower.Reference Boonen, Lips, Bouillion, Bischoff-Ferrari, Vanderschueren and Haentjens³³

Bisphosphonates

Oral alendronate and risedronate, given either daily or weekly, are both established treatments with strong evidence bases that include older patients, and demonstrate benefit in hip fracture prevention. This section will concentrate on newer bisphosphonates and recent concerns about adverse effects.

Studies into ibandronate were the first to show that the benefits of bisphosphonates were maintained with intermittent dosing regimens. The BONE study groupReference Chesnut, Skag and Christiansen^34, Reference Delmas, Recker, Chesnut, Skag, Stakkestad, Emkey, Gilbride, Schimmer and Christiansen³⁵ showed that intermittent oral ibandronate in a less than weekly regimen reduced the rates of vertebral fractures by 50% (P < 0.001), which was comparable to the 62% reduction with daily ibandronate. Intermittent ibandronate has also been shown to improve BMD measurements in post-menopausal women with osteoporosis.Reference Stakkestad, Lakatos, Lorenc, Sedarati, Neale and Reginster³⁶

Both zoledronate and ibandronate are licensed for intravenous use in the treatment of osteoporosis, but the greatest wealth of evidence looks at zoledronate, hence it will be the focus of this section. In 2007 the HORIZON group published the results of two major trials looking at I.V. zoledronate post hip fracture and in post-menopausal osteoporosis.

The HORIZON-RFT (recurrent fracture trial)Reference Lyles, Conon-Emeric, Magaziner, Adachi, Pieper and Mautalen³⁷ took 2127 women aged 50 years or older (mean age 74.5 years) post hip fracture repair and randomized them to receive intravenous zoledronic acid (5 mg) within 90 days of surgery or placebo, with both groups receiving supplemental calcium and vitamin D. The follow-up period was relatively short (median follow-up 1.9 years) but significant benefit in terms of fracture risk reduction and BMD increase was demonstrated. The zoledronate group showed a 35% risk reduction for all clinical fractures, and vertebral fracture rates were also significantly reduced. This was the first osteoporosis trial to show a significant mortality benefit (28% reduction), which may be in part due to play of chance, since it was not due to prevention of hip fracture-related mortality or other explicable mechanisms. Eriksen et al.Reference Eriksen, Lyles and Colon-Emeric³⁸ re-analysed the data on timing of the infusion, and concluded that the optimal time was 2 weeks or more after the fracture.

The HORIZON-PFT (pivotal fracture trial)Reference Black, Delmas, Eastell, Reid, Boonen, Cauley, Cosman, Lakatos, Leung, Man, Mautalen, Mesenbrink, Hu, Caminis, Tong, Rosario-Jansen, Krasnow, Hue, Sellmeyer, Eriksen and Cummings³⁹ compared yearly zoledronate with placebo in almost 8000 post-menopausal women with osteoporosis with or without evidence of existing osteoporotic fractures. At 3 years the relative risk of morphometric vertebral fracture was reduced by 70% in the yearly zoledronate group and hip fracture risk was reduced by 41% compared with placebo, both of which were statistically significant differences.

Eastell et al.,Reference Eastell, Black, Boonen, Adami, Felsenberg, Lippuner, Cummings, Delmas, Palermo, Mesenbrink and Cauley⁴⁰ in a post hoc analysis, found that zoledronate had a greater effect on vertebral fracture risk reduction in younger women, those with normal creatinine clearance and those with a body mass index ≥ 25 kg/m². A more recent sub-group analysis looked specifically at the effect of zoledronate infusions in those aged 75 years or older.Reference Boonen, Black, Colón-Emeric, Eastell, Magaziner, Eriksen, Mesenbrink, Haentjens and Lyles⁴¹ They found a significant reduction in fractures with zoledronate compared with placebo (hazard ratio: 0.65, 95% CI: 0.54–0.78, P < 0.001). Thus zoledronate is highly effective in both secondary and primary prevention of fractures in all age groups. It is important to treat any calcium or vitamin D deficiency before giving the infusion.

In recent years, case reports suggesting potential links between bisphosphonate usage and certain potentially serious conditions has led to renewed analysis of both old and new bisphosphonate data. In this section we will cover the adverse effects that have received most attention over the last 5 years.

Osteonecrosis of the jaw (ONJ) is a condition that appears to be similar to osteomyelitis, but the exact pathophysiology remains unclear. There have been many case reports of ONJ in patients receiving bisphosphonate therapy, and it seems largely related to patients receiving high-dose bisphosphonates for malignancy, and is probably linked to excessive suppression of bone turnover. Of the 7765 women who took part in the HORIZON-PFT study,Reference Black, Delmas, Eastell, Reid, Boonen, Cauley, Cosman, Lakatos, Leung, Man, Mautalen, Mesenbrink, Hu, Caminis, Tong, Rosario-Jansen, Krasnow, Hue, Sellmeyer, Eriksen and Cummings³⁹ there were two reported cases of ONJ: one in the zoledronate group, the other in the placebo arm of the trial. Thus even with this potent bisphosphonate, there is no clear indication that it increases the risk of ONJ when given for osteoporosis.

Atrial fibrillation (AF) is particularly common in the older population. The potential association between bisphosphonate use and increased incidence of atrial fibrillation was unexpectedly first seen in the HORIZON-PFTReference Black, Delmas, Eastell, Reid, Boonen, Cauley, Cosman, Lakatos, Leung, Man, Mautalen, Mesenbrink, Hu, Caminis, Tong, Rosario-Jansen, Krasnow, Hue, Sellmeyer, Eriksen and Cummings³⁹ trial, where the incidence of arrhythmia in the zoledronate group was 6.9% compared with 5.3% in the placebo group (P < 0.003). The results of two major trials (FIT I⁴² and FIT II⁴³) have been pooled (total of 6459 women) to assess the impact of alendronate on rates of AF, and no significant difference in incidence was demonstrated between the treatment and placebo groups (2.5 vs. 2.2%).

In 2008, Heckbert et al.Reference Heckbert, Li, Cummings, Smith and Psaty⁴⁴ published a study looking at 719 patients prescribed alendronate and 966 controls over a 3-year period. Rates of atrial fibrillation were higher amongst those prescribed alendronate (6.7%) compared with the placebo group (4.1%). However, there was no demonstrable relationship between dose or duration of alendronate use and incidence of AF. Additionally, amongst those taking alendronate, AF was shown to be more common in people also prescribed statin therapy or medications for diabetes. Similarly, an observational study from DenmarkReference Abrahamsen, Eiken and Brixen⁴⁵ found a higher incidence of AF amongst people taking an oral bisphosphonate compared with controls. However, those on bisphosphonates who had episodes of atrial fibrillation had higher rates of anti-hypertensive usage and more prescriptions for anti-thrombotic medications, leading the authors to conclude that confounding factors may play a large part in the apparent increased incidence of AF with bisphosphonates.

The acute phase reaction associated with the use of nitrogen-containing bisphosphonates is a constellation of symptoms, typically pyrexia, headache, flu-like symptoms, myalgias and arthralgias. It is likely that these symptoms are the result of the release of pro-inflammatory cytokines into the peripheral blood stream and the use of paracetamol to address this is commonplace. Acute phase reactions have been a recognized effect of intravenous bisphosphonate use since they were first introduced to the market. The HORIZON studyReference Black, Delmas, Eastell, Reid, Boonen, Cauley, Cosman, Lakatos, Leung, Man, Mautalen, Mesenbrink, Hu, Caminis, Tong, Rosario-Jansen, Krasnow, Hue, Sellmeyer, Eriksen and Cummings³⁹ reported that, within 3 days of the first intravenous zoledronate infusion, 31.6% of participants experienced symptoms consistent with an acute phase reaction, but with subsequent infusions the incidence of such symptoms was lower. Acute phase reactions are also known to occur with high-dose oral bisphosphonates.Reference Reginster, Adami, Lakatos, Greenwald, Stepan, Silverman, Christiansen, Rowell, Mairon, Bonvoisin, Drezner, Emkey, Felsenberg, Cooper, Delmas and Miller⁴⁶

Atypical femoral fracture. In 2007, a group from Singapore published a retrospective review of thirteen patients who sustained low-trauma subtrochanteric fractures, nine of whom were taking alendronate. In six of the nine cases of subtrochanteric fractures on alendronate, there was lateral thickening of the femoral cortex. A subsequent case series identified 15 similar cases of low-trauma subtrochanteric or proximal diaphyseal fractures in post-menopausal women taking alendronate.Reference Lenart, Lorich and Lane⁴⁷ These and other reports have led to concerns that long-term bisphosphonates may over-suppress bone turnover, limiting the repair of microfractures or otherwise altering bone architecture, and predisposing to spontaneous fractures.

Abrahamsen et al. used data from the Danish National Hospital Discharge Register and National Prescription Database to look at the rates and types of fracture in relation to alendronate use.Reference Abrahamsen, Eiken and Eastell⁴⁸ They compared patients with previous non-hip fractures who redeemed prescriptions for alendronate with a matched group of non-hip fracture patients who did not receive alendronate. They found that the hazard ratios for subtrochanteric or diaphyseal fractures and for ‘typical’ hip fractures between alendronate-treated patients and controls were the same, with no evidence of an increase in atypical fractures in the alendronate group. In May 2010, Black et al.Reference Black, Kelly, Genant, Palermo, Eastell, Bucci-Rechtweg, Cauley, Leung, Boonen, Santora, de Papp and Bauer⁴⁹ published the results of a secondary analysis of atypical fracture rates in three previous large RCTs of alendronate or zoledronate. Of the 14,195 participants, there were a total of 284 reported hip or femoral fractures, of which only 12 were subtrochanteric or diaphyseal fractures. These numbers are too small to conclude with any certainty a link between bisphosphonate use and atypical femoral fractures.

In conclusion, whilst long-term bisphosphonate use may cause atypical femoral fractures, these events are very rare and the overall fracture prevention benefit outweighs the potential small risk of atypical fractures. Nonetheless, the optimum duration of bisphosphonate treatment merits further study.

Gastrointestinal side-effects, including oesophageal ulceration, are known to be associated with oral bisphosphonate use. These were identified from post-licensing monitoring, and were not seen in the major trials of oral alendronate,Reference Black, Cummings, Karpf, Cauley, Thompson, Nevitt, Bauer, Genant, Haskell, Marcus, Ott, Torner, Quandt, Reiss and Ensrud^42, Reference Cummings, Black, Thompson, Applegate, Barrett-Connor, Musliner, Palermo, Prineas, Rubin, Scott, Vogt, Wallace, Yates and LaCroix^43, Reference Bauer, Black, Ensrud, Thompson, Hochberg, Nevitt, Musliner and Freedholm⁵⁰ risedronateReference Harris, Watts, Genant, McKeever, Hangartner, Keller, Chesnut, Brown, Eriksen, Hoseyni, Axelrod and Miller^51–Reference Taggart, Bolognese, Lindsay, Ettinger, Mulder, Josse, Roberts, Zippel, Adami, Ernst and Stevens⁵⁴ and ibandronate.Reference Chesnut, Skag and Christiansen³⁴ An American analysis of a healthcare database looked at the incidence of gastrointestinal side-effects amongst patients taking oral alendronate and oral risedronate and found no significant difference between the two drugs.Reference Cadarette, Katz, Brookhart, Stürmer, Stedman, Levin and Solomon⁵⁵ Although relatively minor gastrointestinal side-effects such as nausea are still sometimes reported with intravenous bisphosphonates such as zoledronate,Reference Black, Delmas, Eastell, Reid, Boonen, Cauley, Cosman, Lakatos, Leung, Man, Mautalen, Mesenbrink, Hu, Caminis, Tong, Rosario-Jansen, Krasnow, Hue, Sellmeyer, Eriksen and Cummings³⁹ it has been demonstrated that, in terms of gastrointestinal tolerability, intermittent intravenous bisphosphonate therapy is superior to weekly oral bisphosphonates.Reference Blumentals, Harris, Cole, Huang and Silverman⁵⁶

Oesophageal cancer. In 2009 the US Food and Drug Administration (FDA) reported 23 cases of oesophageal cancer in patients taking alendronate.Reference Wysowski⁵⁷ The median time from initiation of alendronate to diagnosis was 2.1 years. One of the patients had Barrett's Oesophagus, a pre-cancerous condition, prior to the commencement of alendronate. They also referred to 31 such cases having been reported in Europe and Japan over the same period. Clearly these are uncontrolled data on a relatively common cancer, and this does not prove a causal link. Several subsequent reports have studied this further.

Solomon et al.Reference Solomon, Patrick and Brookhart⁵⁸ looked at data from US Medicare records and found no difference in rates of oesophageal cancer between patients receiving oral bisphosphonates and those receiving other osteoporosis treatments (e.g. raloxifene or calcitonin). A Danish group compared the incidence of gastric and oesophageal cancer between 13,678 fracture patients who had collected at least one prescription for an oral bisphosphonate with 27,356 matched controls, who had also had a fracture but had not received a bisphosphonate. There was no difference in gastric cancer, and a lower incidence of oesophageal cancer in those taking oral bisphosphonates (hazard ratio: 0.35, 95% CI: 0.14–0.85, P = 0.02).

Two studies using the UK General Practice Research database have yielded conflicting results. Green et al.Reference Green, Czanner, Reeves, Watson, Wise and Beral⁵⁹ conducted a nested case-control study of patients aged over 40 years old with a diagnosis of oesophageal cancer (n = 2954), gastric cancer (n = 2018) and colorectal cancer (n = 10,641) and identified five matched controls for every case of malignancy. Bisphosphonate use was not significantly associated with an increased risk of gastric or colorectal cancer but was shown to be associated with higher rates of oesophageal cancer. For those patients who had received ten or more prescriptions for an oral bisphosphonate, the incidence of oesophageal cancer nearly doubled when compared with controls (relative risk: 1.93, 95% CI: 1.37–2.70). There was also a significant increase in the incidence of oesophageal cancer in those patients who had taken an oral bisphosphonate for 5 or more years when compared with those not on a bisphosphonate. Cardwell et al.Reference Cardwell, Abnet, Cantwell and Murray⁶⁰ also undertook a study using the UK General Practice Research Database and identified a cohort of 41,826 patients prescribed an oral bisphosphonate at some point during a designated 10-year catchment period (1996–2006). They found no significant difference in the rates of oesophageal or gastric cancer between those taking bisphosphonates and matched controls not receiving this medication. Additionally, this study did not identify any association between length of bisphosphonate use and risk of oesophageal or gastric malignancy.

In summary, the available evidence is inconclusive. The UK Medicines and Healthcare products Regulatory Agency (MHRA) has advised that ‘the benefits of bisphosphonate treatment are still considered to outweigh the risks’.⁶¹

Strontium

Strontium ranelate is often referred to as a ‘dual action bone agent’ as it inhibits osteoclasts but has an additional osteoblast-stimulating effect. Strontium is a more radio-absorbent material than calcium, and therefore rises in BMD measurements with strontium may be disproportionate and higher than the true increase in bone mineral content.

There have been two major trials looking at the role of strontium in the treatment of osteoporosis. The Spinal Osteoporosis Therapeutic Intervention study (SOTI) took a total of 1649 women with osteoporosis and at least one vertebral fracture and randomized them to strontium ranelate (2 g/day) or placebo, with both groups receiving additional calcium and vitamin D supplementation. After 4 years they demonstrated a significant 33% reduction in risk of a new vertebral fracture.Reference Meunier, Roux, Ortolani, Diaz-Curiel, Compston, Marquis, Cormier, Isaia, Badurski, Wark, Collette and Reginster⁶² The treatment of peripheral osteoporosis study (TROPOS) randomized a larger group (n = 5091) of older women (mean age 77 years) with osteoporosis but without previous vertebral fractures. After 3 years, the risk of a non-vertebral fracture was reduced by 16% (P = 0.04) in the strontium group compared with placebo.Reference Reginster, Seeman, De Vernejoul, Adami, Compston, Phenekos, Devogelaer, Curiel, Sawicki, Goemaere, Sorensen, Felsenberg and Meunier⁶³ Analysis of the 53% of participants continuing to 5 years demonstrated that results were maintained with significant reductions in non-vertebral, vertebral and hip fracture rates after 5 years of strontium use.Reference Reginster, Felsenberg, Boonen, Diez-Perez, Rizzoli, Brandi, Spector, Brixen, Goemaere, Cormier, Balogh, Delmas and Meunier⁶⁴

Pooled analysis of 3-year data from SOTI and TROPOS looking only at those aged 80 years or older (n = 1488) showed significant reductions in the risk of vertebral and non-vertebral fractures.Reference Seeman, Vellas, Benhamou, Aquino, Semler, Kaufman, Hoszowski, Varela, Fiore, Brixen, Reginster and Boonen⁶⁵ More recently, Rolland et al.Reference Rolland, Van Kan, Gillette-Guyonnet, Roux, Boonen and Vellas⁶⁶ took the 3-year results from SOTI and TROPOS and assessed the effects of strontium in those participants who at baseline were deemed frail (n = 264), intermediate (n = 2472) or robust (n = 2346). Rates of vertebral fractures were significantly reduced in all three groups of patients but the effect was most marked in those with frailty, where strontium treatment reduced the risk of vertebral fracture by 58% (relative risk: 0.42, 95% CI: 0.24–0.74, P < 0.01).

Parathyroid hormone

Parathyroid hormone (PTH) 1–34, or teriparatide, has impressive anabolic effects on bone density and reduces both vertebral and non-vertebral fractures.Reference Neer, Arnaud, Zanchetta, Prince, Gaich, Reginster, Hodsman, Eriksen, Ish-Shalom, Genant, Wang and Mitlak⁶⁷ Human recombinant intact (1–84) parathyroid hormone has evidence for preventing only vertebral fractures in post-menopausal women, with variable, site-specific effects on BMD that include increases at most sites, but a reduction at the forearm.Reference Greenspan, Bone, Ettinger, Hanley, Lindsay, Zanchetta, Blosch, Mathisen, Morris and Marriott⁶⁸ Both preparations are given by daily injection, and are more expensive than other treatments, thus limiting their use.

There were concerns that prior or concurrent bisphosphonate therapy might attenuate the effect of PTH preparations.Reference Black, Bilezikian, Ensrud, Greenspan, Palermo, Hue, Lang, McGowan and Rosen^69, Reference Finkelstein, Hayes, Hunzelman, Wyland, Lee and Neer⁷⁰ Nonetheless, recent observational data suggest that teriparatide improves bone density, reduces fractures and improves pain and quality of life in elderly patients (mean age 70 years) with severe osteoporosis who have already been treated with bisphosphonates.Reference Maugeri, Russo, Luca, Leotta, Mamazza, Sorace, Rizzotto, Manuele, Fiore, Taverna, Castiglia and Calitro⁷¹ Furthermore, there is preliminary evidence of a beneficial synergistic effect of teriparatide and annual zoledronate on bone mineral density at the hip in women of mean age 65 years.Reference Cosman⁷²

Masud et al.Reference Masud⁷³ have presented data on over 600 UK patients treated with teriparatide and showed that about 70% persisted to completion of 18-months therapy. In England and Wales NICE recommends that teriparatide can be used for women at very high risk, with previous fragility fractures, who are intolerant of or unresponsive to bisphosphonates or strontium.⁷

Other hormonal treatments

Selective oestrogen receptor modulators (SERMs) such as raloxifene, lasofoxidine and bazedoxifene have been shown to play a role in the treatment of post-menopausal osteoporosis. The Multiple Outcomes of Raloxifene Evaluation (MORE) study demonstrated increases in BMD at the hip and lumbar spine, and significantly lower rates of new vertebral fracture in the raloxifene group.Reference Delmas, Ensrud, Adachi, Harper, Sarkar, Gennari, Reginster, Pols, Recker, Harris, Wu, Genant, Black and Eastell^74, Reference Ettinger, Black, Mitlak, Knickerbocker, Nickelsen, Genant, Christiansen, Delmas, Zanchetta, Stakkestad, Glüer, Krueger, Cohen, Eckert, Ensrud, Avioli, Lips and Cummings⁷⁵ The study also revealed higher rates of venous thromboembolism with raloxifene compared with placebo, and this has been found with other SERMs also.

Lasofoxifene is another SERM that has been shown to reduce fracture rates in women with post-menopausal osteoporosis. A recently published RCT compared lasofoxifene (0.25 or 0.5 mg daily) or placebo in 8556 post-menopausal women with osteoporosis (mean age 67 years).Reference Cummings, Ensrud, Delmas, LaCroix, Vukicevic, Reid, Goldstein, Sriram, Lee, Thompson, Armstrong, Thompson, Powles, Zanchetta, Kendler, Neven and Eastell⁷⁶ Over a 5-year period, the higher lasofoxifene dose significantly reduced the rates of vertebral and non-vertebral fracture when compared with placebo. Additionally, they demonstrated benefits of lasofoxifene in reducing rates of oestrogen-receptor positive breast cancer and coronary heart disease but, as with raloxifene, rates of venous thromboembolism were significantly greater than with placebo. Bazedoxifene has also been shown to have anti-fracture benefits on a par with those seen with raloxifene.Reference Silverman, Christiansen, Genant, Vukicevic, Zanchetta, Villiers, Constantine and Chines⁷⁷

Denosumab

The receptor activator of nuclear factor-κB ligand (RANKL) is a cytokine essential to osteoclast function. Denosumab is human monoclonal antibody which inhibits this cytokine, and is given by 6-monthly subcutaneous injection, with a cost in the same range as non-generic bisphosphonates or strontium.

Evidence for its efficacy in fracture prevention comes from the FREEDOM trial,Reference Cummings, San Martin, McClung, Siris, Eastell, Reid, Delmas, Zoog, Austin, Wang, Kutilek, Adami, Zanchetta, Libanati, Siddhanti and Christiansen⁷⁸ in which 7868 women aged 60–90 (mean age 72) with T-scores between –2.5 and –4.0 were randomized to 6-monthly injections of denosumab or placebo. New radiographic vertebral fractures occurred in 7.2% of controls and 2.3% of actively treated patients, and hip fractures in 1.2% and 0.7%, respectively, over 3 years. The main side-effects were eczema (3 vs. 1.7%) and cellulitis (12 cases vs. 1 case). In men receiving androgen deprivation therapy for prostate cancer, denosumab resulted in improvements in bone mineral density and a reduction in vertebral fractures.Reference Smith, Egerdie, Hernández Toriz, Feldman, Tammela, Saad, Heracek, Szwedowski, Ke, Kupic, Leder and Goessl⁷⁹ In a comparison with weekly alendronate therapy, denosumab showed significantly greater effects on bone density and bone turnover markers.Reference Brown, Prince, Deal, Recker, Kiel, de Gregorio, Hadji, Hofbauer, Alvaro-Gracia, Wang, Austin, Wagman, Newmark, Libanati, San Martin and Bone⁸⁰

The effectiveness in hip fracture, relative freedom from side-effects and convenience of twice-yearly subcutaneous injections makes this a very attractive second-line treatment for older patients unable to tolerate oral bisphosphonates, and it has been approved by NICE⁸¹ for this indication.

Vertebroplasty and kyphoplasty

While management of fragility fractures themselves is beyond the scope of this review, acute vertebral compression fractures often present to physicians with severe pain, which can be disabling, particularly in frail older patients. Treatments include conventional analgesia, bone-acting drugs such as calcitonin, bisphosphonates, parathyroid hormone and vitamin D, physiotherapy exercises and spinal orthoses as well as invasive procedures.Reference Francis, Aspray, Hide, Sutcliffe and Wilkinson⁸²

Injection of bone cement into painful vertebral compression fractures is increasingly popular, appears to relieve pain, and may stabilize the fracture. The cement is usually injected under local anaesthesia and X-ray control via the pedicle. The term ‘kyphoplasty’ refers to a procedure in which a balloon is inflated to create a void and partially re-expand the vertebra before cement injection. ‘Vertebroplasty’ refers to the injection of cement without the balloon procedure. In unblinded RCTs, both procedures offer rapid and sustained pain relief and improvement in function and quality of life.Reference Klazen, Lohle, de Vries, Jansen, Tielbeek, Blonk, Venmans, van Rooij, Schoemaker, Juttmann, Lo, Verhaar, Van Der Graaf, van Everdingen, Muller, Elgersma, Halkema, Fransen, Janssens, Buskens and Mali^83–Reference Wardlaw, Cummings, Van Meirhaeghe, Bastian, Tillman, Ranstam, Eastell, Shabe, Talmadge and Boonen⁸⁵ However, two trials of vertebroplasty in which the control group underwent a sham procedure, which included injection of local anaesthetic, showed rapid relief in both groups, with no significant difference between active and sham procedures.Reference Buchbinder, Osborne, Ebeling, Wark, Mitchell, Wriedt, Graves, Staples and Murphy^86, Reference Kallmes, Comstock, Heagerty, Turner, Wilson, Diamond, Edwards, Gray, Stout, Owen, Hollingworth, Ghdoke, Annesley-Williams, Ralston and Jarvik⁸⁷ Many explanations for this have been suggested, including inappropriate patient selection, inadequate injection procedures in the trials, placebo effect, Hawthorne effect (i.e. the influence of being studied rather than the actual intervention) and therapeutic benefit of local anaesthetic injections.Reference Buchbinder, Osborne and Kallmes^88, Reference Noonan⁸⁹ The potential longer-term effects of these procedures on spinal stability and future vertebral fractures remain uncertain. Further research is ongoing to compare the two procedures,Reference Wardlaw and Van Meirhaeghe⁹⁰ and scientific evaluation of local anaesthetic injection is also merited before clear recommendations can be made.