Report of the 3rd Bandolier Conference - on Osteoporosis

January 16/17, 1997 at Eynsham Hall Conference Centre, Witney, Oxon


Osteoporosis is a major clinical problem. In the U.K. around 2 million people suffer from osteoporosis, resulting in 60,000 hip, 50,000 wrist (Colles') and 40,000 vertebral fractures a year. However, demographic changes mean that the problem presented by osteoporosis will grow during the early years of the next century. As a result, effective, evidence-based use of osteoporosis services is a priority for clinicians, managers and politicians.

Against this background, more than 70 delegates from all over the UK met in Oxford in January at the third Bandolier conference, to discuss osteoporosis. Presentations reviewed the burden of illness, the diagnosis of osteoporosis, prevention and treatment options as well as strategies for developing an osteoporosis service. This newsletter brings you some of the meeting's highlights.

Osteoporosis: the burden of illness

Chairing the session Dr Martin Lawrence , General Practitioner and Lecturer in the Department of Primary Health Care, Oxford University, noted that the burden of illness imposed by osteoporosis has risen rapidly in recent years.

Dr Lawrence remarked that doctors now face a similar problem managing osteoporosis as they did eight or nine years ago with hypercholesterolaemia. For example, both conditions:

Dr Lawrence added that a strategy for hypercholesterolaemia management by GPs is now becoming clear. Moreover, the burden of illness means that osteoporosis management is likely to become a primary care problem.

Dr Ashok Bhalla , Consultant Rheumatologist, the Royal National Hospital for Rheumatic Diseases in Bath, discussed the definition of osteoporosis. He told delegates that until recently there was no way to assess which patients will develop osteoporotic fractures. This changed following the introduction of imaging systems. Now bone scans reliably detect low bone mineral density (BMD), the precursor of an osteoporotic fracture.

Dr Bhalla reminded delegates that osteoporosis is a progressive systemic skeletal disease characterised by low BMD and micro-architectural deterioration. This leads to an increase in bone fragility and a susceptibility to fractures. Currently, dual-energy X-ray absorptiometry (DXA) represent the `gold standard' for diagnosis and predicting fracture risk. However, the risk gradient is continuous and studies have been unable to set a fracture threshold. Nevertheless, the fracture risk increases between 1.5 and 3 fold for each 1 standard deviation (SD) fall in BMD.

Broadly, two approaches are employed to diagnose osteoporosis. The World Health Organisation defines T scores - levels below the young, normal mean. According to the WHO definition:

Statistically, the WHO definition means that 15 per cent of the population are osteopenic and 0.5 per cent suffer from osteoporosis.

The alternative approach uses Z scores, which are corrected for age. As the population ages, the normal distribution curve shifts to the left and the proportion of people with osteoporosis increases. During questioning, Dr Mike Kirby, a Letchworth GP, pointed out that there was some debate about whether to use T or Z scores in clinical practice. Dr Bhalla suggested that either reading was inadequate without interpretation and a "good and tight" reporting system. Later in the conference, delegates heard that some groups have implemented strategies that meet this need.
Several speakers pointed out that age is probably the pre-eminent risk factor for osteoporotic fractures. "In the elderly, it isn't a question of who's got osteoporosis, but who hasn't," Dr Bhalla said. As a result of the age-related decline in BMD there is a marked rise in fracture risk (see Table 1), which translates into the lifetime risks in Table 2.

Per cent of patients with:
Range (years)
any fracture
hip fracture
50 - 59
60 - 69
70 - 79
80 plus
Table 1: Age-related increase in fracture risk table

Lifetime risk of fracture (%)
proximal femur
distal forearm
Table 2: Lifetime risk at age 50 years of fracture by site

Dr Bhalla then reminded delegates of the natural history of osteoporosis. Bone mass increases as the skeleton matures. Men tend to reach a higher peak BMD than women. Moreover, Black people's peak BMD tends to be higher than either Caucasian or Asians. BMD then declines steadily with age. The fall is specially marked among post-menopausal women. However, as Figure 1 shows, a wide variety of factors influence BMD and fracture risk. On the up-side, this also provides a number of targets for intervention.

Moreover, bone loss does not occur at a consistent rate throughout the skeleton. Wrist fractures occur relatively early - during the patients' 50s and 60s. Spinal and hip fractures tend to occur in later life. The pattern of osteoporotic fractures may reflect the distribution of trabecular and cortical bone in the skeleton.

Dr Peter Selby , Consultant Endocrinologist at Manchester Royal Infirmary, said, during his talk on the epidemiology of osteoporosis in the UK, "The Colles' is perhaps the Cinderella of fractures". Colles' fractures tend to affect younger, productive people and many sufferers are not hospitalised. Oestrogen is protective. However, oral corticosteroids, which tend to undermine the axial skeleton, alcohol, calcium intake and body mass index do not influence the risk of Colles' fractures.

"Vertebral fractures are the hallmark of osteoporosis," Dr Selby added. However, up to two thirds of vertebral fractures are detected on X-ray rather than by clinical symptoms or signs. Some 20 per cent of elderly patients show some evidence of deformity on X-rays and the fracture risk increases with age.

Finally, Dr Selby commented that most attention falls on hip fractures because of their impact on health and welfare systems. Around 20 per cent of sufferers die within a year and half lose their independence. Overall, treating osteoporosis costs some £750 million each year and hip fractures account for much of the economic burden.

The incidence of hip fractures is relatively low up to the age of 70 years. However, demographic shifts mean that hip fractures are becoming more common and the burden may double between 1985 and the early years of the next century 14 . While the increase is sobering enough in North America and Europe, the rise will be even more marked in the developing world. This may reflect socio-economic factors. For example, the prevalence of hip fractures tends to increase with the per capita Gross Domestic Product. Several other risk factors have been identified for hip fractures (see Table 3).

Positive factors
Negative factors
low bone mineral density
weight gain
previous fracture
regular walking
poor health
moderate alcohol consumption
sedative drugs

low exercise

high caffeine

Table 3: Factors influencing the risk of hip fracture

Despite this burden of illness, relatively few women are treated. Dr Selby has estimated that in Manchester there are 25,000 women over the age of 50 years who are osteoporotic. Prescribing data would suggest that at the very best just under 9,000 of these are receiving bone sparing treatment. The majority of this treatment is HRT with much smaller usage of calcium supplements and bisphosphonates. "Almost two thirds are presently untreated" he said. "This is almost certainly an underestimate. We are only scratching the surface."

When questioned about the duration of HRT needed to protect the skeleton, Dr Selby said that when women stop taking HRT the bone mass begins to decline. He added that women need to take HRT for between 5 and 10 years to retain some increase in BMD in later life. The role of HRT in osteoporosis prevention was considered in more detail later in the conference.

Diagnosis of osteoporosis

Dr Paul Thompson , Consultant Rheumatologist at Poole Hospital NHS Trust, discussed the options for the assessment of bone density and agreed that DXA was the gold standard. However, there are other options including ultrasound, quantitative computed tomography and biochemical markers.

Despite being the gold standard, DXA does not measure BMD. Rather it measures the amount of calcium, which is a surrogate marker for bone density. As a result, DXA is less accurate in elderly patients, particularly those over 70 years of age, who may have arthritic changes that result in bone calcification. Therefore, DXA may under-diagnose osteoporosis in these patients. Dr Thompson also suggested measuring BMD at the site of main clinical interest.

Ultrasound, which usually measures BMD at the heel, is growing in popularity. The system is relatively cheap and portable, raising the possibility of screening in primary care. However, Dr Thompson added that ultrasound's mode of measurement is not fully understood. Moreover, there are a number of different machines, which have not been fully calibrated against each other. Studies suggest that, in the elderly, ultrasound predicts fracture risk as accurately as DXA. Moreover, recent studies show that both methods can detect changes in bone density produced by treatment.

However, both DXA and ultrasound need normal ranges. Dr Thompson suggested that these should come from the local area rather than relying on manufacturers' ranges, which are often based on studies of North American populations. Moreover, he asked whether the range should be set according to the general population or the fittest proportion to "set the standard as high as possible". Currently, these questions are unresolved.

Dr Eugene McCloskey , Senior Clinical Research Fellow at the Royal Hallamshire Hospital, Sheffield, discussed the clinical indications for bone density measurement. He noted that population based screening of post-menopausal women cannot be justified. Moreover, the WHO criteria represent diagnostic guidelines rather than treatment thresholds. Dr McCloskey argued that patients should be selected based on strong clinical or historical risk factors and only if the results are likely to influence management. However, bone density measures may be used to monitor therapy in some patients.

From several sets of guidelines, Dr McCloskey noted a certain consistency about the patients who may be suitable cases for bone densitometry:

However, the question of how frequently women need to be monitored remains unanswered. Dr McCloskey noted that the certainty is proportional to the duration of follow-up and the square-root of the number of measurements. As a result, two measurements over three years is better than three readings over two years. The choice of site, the impact of treatment and the departments' reproducibility also influences the decision.

Professor Chris Price , Professor of Clinical Biochemistry at the St Bartholomew's and Royal London School of Medicine and Dentistry, discussed biochemical markers of bone formation and resorption. Several have been suggested (see Table 4).

Osteoblast activity
measured in serum

Osteoclast activity (resorption)
'a-e '- measured in urine
'f' - measured in serum
bone alkaline phosphatase
galactosyl hydroxylysine
total pyridinium crosslinks

deoxypyridinoline (free and total)

N- and C-terminal telopeptide (also known as "NTX" and "Crosslaps" )

tartate-resistant acid phosphatase (TRAP)
4: Biochemical markers of table bone formation and resorption

Each of these markers has certain advantages and disadvantages. Moreover, levels show within-and between-day variations and can be influenced by nutritional state, steroid use, recent fractures and so on. However, early indicators suggest that bone alkaline phosphatase, deoxypyridinoline and N-terminal macromolecule cross-links are emerging as the leading contenders for use.

Professor Price noted that laboratories need a certain technical sophistication to perform these assays. Currently, St Bartholomew's Hospital, the Royal Liverpool Hospital, St Mary's Hospital Paddington and the Sheffield Northern General Hospital act as supra-regional assay centres and provide advisory services for bone markers.

There is a growing interest in biochemical markers. For example, the pathology measured by the assays may represent independent risk factors for osteoporosis. Moreover, biochemical markers may allow a drug's effectiveness to be assessed in as little as three months.

Professor Price added the assays were becoming easier to perform. Moreover, with more widespread use, the cost of each assay should fall. Professor Price predicted several future roles for biochemical bone markers, including:

Professor David Barlow , Consultant Gynaecologist from the John Radcliffe Hospital, Oxford, who chaired the Department of Health Advisory Group on Osteoporosis, (AGO Report 1994), commented that there is great uncertainty surrounding osteoporosis management. He began by highlighting the immense suffering caused by osteoporosis. Patients endure pain, vertebral deformity, loss of height and limitation on their activities of daily living. In a study of 6,000 women, 120 suffered hip fractures. After six months the hip fracture dramatically affected the women's ability to perform certain tasks (see Table 5).

Per cent able to perform activity
before fracture
6 months after fracture
dress independently
walk across road independently
climb flight of stairs
Table 5: Disability caused by hip fractures

Professor Barlow noted that certain population strategies may be appropriate, such as smoking cessation, moderation of alcohol consumption and encouraging weight bearing exercise. Primary prevention may involve intervening with HRT for premature menopause and prolonged amenorrhoea or targeting people using oral corticosteroids. However, he noted that discovering an evidence base is not the same as ensuring implementation. For example, there is now good evidence that vitamin D and calcium reduce fracture rate in the elderly. So, Professor Barlow asked, why is so little being done in nursing homes?

Secondary prevention could target people who experienced a fracture or those with confirmed deformities. However, Professor Barlow questioned the value of routine bone scans in these patients. "What would densitometry add to decision making?" he asked.

Professor Barlow also expressed concern that the wide availability of bone densitometry, especially ultra-sound, in GP surgeries, could lead to "population screening by the backdoor". He argued that this could distort the appropriate use of bone densitometry. However, estimates of the number of scans needed for defined indications varies widely, one estimate being 180 scans per 100 000 of the population 4 . Professor Barlow told delegates that national guidelines are being prepared to help address some of these questions.
Prevention and treatment options - what is the evidence?

Dr Antony Woolf , Consultant Rheumatologist at the Royal Cornwall Hospital, Truro, examined the evidence for general lifestyle measures. He noted that hard evidence seldom exists for these interventions. However, dietary calcium and vitamin D seem to be effective, especially among elderly women. Holbrook et al reported in 1988 that the risk of hip fractures related to calcium intake 9 . More recently a French study 2 assessed the effects of a combination of 1200mg calcium and 20 micrograms (800 iu) vitamin D3 daily in 3270 elderly women. After 18 months, women who received the supplements had 43 per cent fewer hip fractures and 32 per cent fewer nonvertebral fractures compared to the placebo group. The same group recently reported 3 that this regimen reduced hip fracture risk by 23 per cent over three years.

Exercise is also beneficial. In children, adolescents and premenopausal women, high load exercise produces a small increase in BMD. However, among post-menopausal women high impact exercise is undoubtedly beneficial, although this is lost when exercise stops. Dr Woolf noted that the benefits were less than 10 per cent. In response to a question, he suggested that patients should walk a minimum of seven miles a week - equivalent to three hours hard walking.

Older patients may be unable to perform high impact exercise. However, gentle exercise improves balance and muscular strength and, as a result, prevent falls. For example, one study 15, suggested that Tai Chi produced a 25 per cent reduction in the number of falls.

Dr Woolf noted that other lifestyle factors may influence the risk of osteoporosis:

Summarising the evidence base, Dr Woolf scored each lifestyle risk factor on a scale of one to five:
Caffeine: 1
Smoking: 1
Alcohol: 1
Calcium: 2
Exercise: 3
Vitamin D: 3
Hip protectors: 3
Falls: 4

Dr Jonathan Tobias , a Rheumatologist from Bristol Royal Infirmary, discussed the role of HRT in osteoporosis prevention. This is not as clear cut as is often supposed. He noted that most women take HRT to relieve menopausal symptoms. However, oestrogen normalises bone metabolism. For example, studies 8 , suggest that HRT increases BMD at the femoral neck and spine.

However, compliance can be a problem with long-term treatment especially in the light of recent reports 12 , which found that taking HRT for more than 10 years may increase the risk of breast cancer. In one study 5 of 5,743 women who underwent a hysterectomy:

Moreover, an important question remains unanswered. Are the benefits of HRT maintained in later life? Dr Tobias pointed out that bone loss may accelerate in later life (see Table 6). While long-term HRT may prevent bone loss at the spine and hip, the effect may be lost once treatment is discontinued, so that fracture risk among past HRT users may approach that of the untreated population in later life. But this is just the time that women are most at risk from a hip or vertebral fracture.

Age (years)
Bone loss (mg/cm2) femoral neck
over 84
Table 6: Bone loss at femoral neck by age 6

One option is to prescribe HRT to older women. The advent of continuous combined HRT, which no longer requires regular bleeds for endometrial protection, may make this a viable option. However, this approach needs further investigation. Dr Tobias pointed out that as most women take HRT for less than five years, it is not an effective intervention for osteoporosis as currently used.

Dr Richard Keen , ARC Research Fellow in Rheumatology at St Thomas' Hospital, reviewed the randomised control data for vitamin D and calcium in osteoporosis prevention. The results of the study 2 have already been noted. However, other studies reported mixed results.

One study 7 , evaluated the effects of between two and five annual vitamin D injections in 199 men and women over 85 years of age living in their own homes and 142 people aged between 75 and 84 years in residential care. Patients who received the injections were significantly less likely to experience total and arm fractures than controls. The fracture rate was higher among women, and only women benefited from supplements. Nevertheless, the authors recommended that older people should receive vitamin D supplements.

In contrast, another study 11 reported that vitamin D3 (400 iu) once daily for up to 3.5 years did not reduce the incidence of hip or other fractures compared to placebo among 2578 people aged over 70 years. However, the authors admitted that vitamin D supplements may be more effective if patients also take calcium. Nevertheless, the women showed a 2.3 per cent increase in femoral neck bone density. This should translate into a 15 per cent reduction in hip fractures. But the study was only powered to detect a 20 per cent reduction in hip fractures.

In conclusion, Dr Keen suggested that several questions remain unanswered:

Professor David Hosking , Consultant Physician, Nottingham City Hospital, reviewed the role of bisphosphonates, which can be divided into three broad groups. All act by binding to the bone surface thereby inhibiting osteoclast resorption. This prevents bone resorption. However, the three generations differ in potency in animal models of bone resorption. Their relative potencies are:-

First generation: etidronate potency 1
Aminobisphosphonates: alendronate potency 1000
Cyclic bisphosphonates: tiludronate potency 10

Reviewing the evidence to date, Professor Hosking said that bisphosphonates seem to produce a 5 to 8 per cent gain in BMD and halve vertebral fracture rate. There is also a dose response relationship, with most BMD gained during the first year of treatment.

Professor Hosking reviewed the results of the Fracture Intervention Trial (FIT) 1 . In the study, 6459 women, aged between 55 and 80 years with low femoral neck BMD received either alendronate or placebo. The first arm included women with existing vertebral deformity. The monitoring committee terminated this arm after an average 2.9 year follow-up. The second arm enrolled women in whom X-rays did not reveal spinal fractures before entering the study and is due to be completed this year.

Women with existing vertebral deformity taking alendronate were significantly less likely to suffer hip, vertebrae and wrist fractures. For example, the incidence of hip and vertebral fracture was 51 and 55 per cent respectively lower among women receiving alendronate than placebo. The risk of any clinical fracture was 28 per cent lower among women receiving alendronate.

However, bisphosphonates have two problems. Firstly, poor absorption from the upper gastro-intestinal tract (1 to 10 per cent) and this is strongly inhibited by food, particularly that containing calcium. Secondly, bisphosphonates can irritate the upper gut. However, adequate patient counselling reduces the risk of this side-effect. For example, in FIT there was no difference between alendronate and placebo in the incidence of oesophageal ulcers.

"This is very reassuring," Professor Hosking commented. "Bisphosphonates are the single most important drug in the treatment of osteoporosis. They reduce fracture rate and are well tolerated if used with care." However, he called for more information on the newer bisphosphonates and he expressed concern extrapolating from one study to another, largely because the patient population may differ.

One question shared by many members of the audience was whether or not there are data for the bisphosphonates on numbers needed to treat. Professor Hosking confirmed that the FIT study had not published on this. (See addendum for a numbers needed to treat analysis.)

Strategies for osteoporosis

The final session discussed strategies for developing an osteoporosis service. Dr Mike Davies , Consultant Physician, Manchester Royal Infirmary, described his service, which opened in 1970. The service aims to diagnose and manage osteoporosis, monitor efficacy and outcomes and audit. Diagnosis is important, he noted, as other diseases also manifest as skeletal deformities.

The Manchester team is multi-disciplinary including a physician, radiologist with an interest in the skeleton, biochemists, physiotherapists and rehabilitation physician. This offers co-ordinated care. If patients do not like dairy products it can be difficult to change their diet in a way appropriate to promoting bone health. The physiotherapist may offer hydrotherapy, which Dr Davies believes plays a role in rehabilitation, and a dietician ensures an adequate intake of calcium (1.5g daily compared to an RDA of 700mg). The service uses a DXA scanner offering direct access to fundholders. Dr Davies also noted that education of GPs is important.

Similarly, Dr Mike Stone , Consultant Physician, Llandough Hospital, Cardiff, offers a direct access clinic based at both major trusts in the area. The bone assessment clinic is staffed by the consultant physician, a registrar, a clinical assistant and a specialist nurse serving a population of 700,000 and offering supra-regional referrals.

Dr Stone published guidelines that are distributed to all local GPs informing them about the service and suggesting indications for bone scans. The results are presented with a consultant report offering a diagnosis and advice on treatment and follow-up. "We avoid T and Z scores alone and give practical specific advice," he said. He noted that the advantages of the direct access clinic include saving consultants' time, shorter waiting lists as well as patient and GP satisfaction. However, the disadvantages include inappropriate referrals and an incomplete patient assessment. Dr Stone presented the results of an audit suggesting that 19 per cent of scan requests were inappropriate. As a result all scan requests are now seen by consultants and the team developed a purpose designed form.

Finally, Dr Mike Kirby , a GP in Letchworth, presented results of an audit of the effectiveness of a menopause clinic 10 targeting women likely to be in or around their climacteric (43 to 60 years of age). The practice database revealed a target population of 885 women. Each woman was sent an invitation to attend the clinic and some were re-invited if they did not respond to the first letter. A total of 1,500 invitations were sent out and 780 women attended the clinic, an 88 per cent response rate. Based on risk factors, 72 women were identified as being at high risk of developing osteoporosis and referred for DXA. Of these, 35 women showed low bone mineral density.

Dr Kirby added that the number of women taking HRT or bisphosphonates increased steadily after the introduction of the clinic. Now some 28 per cent of women aged between 45 and 60 take HRT and half of those who underwent a hysterectomy - both figures approximately twice the national average. He concluded that the clinic provides an effective framework for educating women about the benefits of HRT in particular and offering advice about osteoporosis generally.

A survey of patients who attended for bone densitometry investigation showed that these women made more lifestyle changes, asked more questions, followed instructions more carefully, increased dietary calcium intake, increased exercise and were more compliant with treatment than the patients with normal bone density.


  1. Black DM, Cummings SR, Karpf DB, Cauley JA, Thompson DE et al. Randomised trial of effect of alendronate on risk of fracture in women with existing vertebral fractures. Lancet , 1996; 348: 1535-41
  2. Chapuy MC, Arlot ME, Duboeuf F, Brun J, Crouzet B et al. Vitamin D3 and calcium to prevent hip fractures in elderly women. New Engl J Med, 1992; 327: 1637-42
  3. Chapuy MC, Arlot ME, Delmas PD and Meunier PJ. Effect of calcium and cholecalciferol treatment for three years on hip fractures in elderly women. BMJ , 1994; 308: 1081
  4. DoH Advisory Group of Osteoporosis Report. November 1994
  5. Drew SV, Rowe R, Panay N and Studd JWW. Use of hormone replacement therapy following hysterectomy: a general practice audit pilot study. In: Proceedings of the 1996 Bath conference on osteoporosis and bone mineral measurement: 9
  6. Ensrud KE, Palermo L, Black DM et al, Hip and calcaneal bone loss increase with advancing age: longitudinal results from the study osteoporotic fractures. Journal of Bone and Mineral Research, 10: 1778-87
  7. Heikinheimo, RJ et al. Annual injection of vitamin D and fractures of aged bones. Calcif Tissue Int , 1992; 51: 105-10
  8. Hillard TC, Whitcroft SJ, Marsh MS et al. Long-term effects of transdermal and oral hormone replacement therapy on postmenopausal bone loss. Osteoporosis International, 1994; 4: 341-48
  9. Holbrook, TL, Barrett-Coonor, E, Wingard, L. Dietary calcium and risk of hip fracture: 14 year prospective population study. Lancet, 1988; 2: 1046-49
  10. Kirby MC. Menopause and osteoporosis screening in a practice of 9,500 patients, Current research in osteoporosis and bone and mineral measurement, 1996: 9
  11. Lips P et al. Vitamin D supplements and fracture incidence in elderly persons. Ann Intern Med , 1996; 124: 400-6
  12. Marsden J, Sacks NPM. Hormone replacement therapy and breast cancer. Endocrine-Related Cancer; 1996; 3: 81-97
  13. Ring EFJ, Elvins DM and Bhalla AK. British Institute of Radiology, London, pp 128
  14. Royal College of Physicians. Fractured neck of femur, prevention and management. Summary and recommendations of a report of the Royal College of Physicians. Journal of the Royal College of Physicians, London, 1989; 23: 8-12
  15. Wolf SL, Kutner NG, Green RC, McNeely E. The Atlanta FICSIT. Two exercise interventions to reduce frailty in elders. Ann Gerieke Soc, 1993; 41: 329-32


Subsequent to the meeting, Numbers Needed to Treat data has become available. This is as follows:

Number of patients remaining fracture free*

No of patients with fracture / total
No of patients with fracture / total exposed
Cost per fracture saved



(any fracture)

- Harris study based on aggregate data for years 0-3
- FIT study lasted for 2.9 years
- In Harris study, groups one and two considered placebo, Groups three and four active
- Cyclical etidronate therapy costed at 44.7p per day
- Alendronate therapy costed at 91.8p per day
- No other costings included

N.N.T. analysis

V ertebral Fracture*

Hip Fracture*
Etidronate (Harris & Watts)
Previous fracture + low BMD (n=423)


Alendronate (FIT)
Previous fracture + low BMD (n=2027)

Alendronate (FIT)
Previous fracture + low BMD (n=2027)


Calcium/Vit D (Chapuy)
Nursing home residents (n=3270)


*Criteria for review
- randomised double-blind trial versus placebo
- >=100 patients involved
- show statistically significant result

Data analysis: Merck Sharp & Dohme