This Article Abstract Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Holt, G Articles by Reeve, J Search for Related Content PubMed PubMed Citation Articles by Holt, G Articles by Reeve, J

Prevalence of osteoporotic bone mineral density at the hip in Britain differs substantially from the US over 50 years of age: implications for clinical densitometry

¹ Department of Medicine (Box 157), Clinical Gerontology and the Institute of Public Health, Clinical School, Hills Road, University of Cambridge, Cambridge CB2 2QQ, ² Department of Medicine & Therapeutics, University of Aberdeen, Aberdeen AB25 2ZD, ³ Royal National Hospital for Rheumatic Diseases, Bath BA1 1RL, ⁴ Royal Cornwall Hospital, Truro, Cornwall TR1 3LJ and ⁵ Northwick Park Hospital, Harrow HA1 3UJ, UK

Correspondence: Dr J Reeve, Department of Medicine (Box 157), Addenbrooke's Hospital, Hills Road, Cambridge CB2 2QQ, UK

	Abstract

Top Abstract Introduction Materials and methods Results Discussion References

 
The purpose of this study was to compare hip bone mineral density (BMD) recorded in seven population based cohorts in Britain with the third National Health and Nutrition Examination Survey (NHANES III) US population-based reference data, in order to assess geographic variation in the prevalence of osteoporosis. Men and women aged 50–80+ years were randomly recruited from population and health registers. Dual X-ray absorptiometry (DXA) equipment was used to measure BMD at the hip, with the femoral neck and the trochanter regions studied. Prevalences of osteopenia and osteoporosis were estimated in accordance with World Health Organisation diagnostic criteria for women. Young normal data, used to establish cut-off criteria, was from NHANES III. Both male and female British subjects over 50-years-old were found to have significantly higher mean BMD at the femoral neck and trochanter than their US counterparts. Decline in BMD with age in British men appeared slower than in US men. Between British centres there were also statistically significant differences in BMD values in both sexes. British age-adjusted prevalences of osteopenia in women averaged 20% less than those of NHANES III, whereas the prevalence of osteoporosis was substantially lower in British subjects of both sexes (55% in women, 68% in men). Thus, applying the US NHANES III data as the referent, osteoporosis of the proximal femur in Britain appears to be less common than in the US, due primarily to differences in the lower tails of the BMD distributions. Providing that the relationship between fracture rates and BMD is the same in Britain and the US, it would still be appropriate to apply the reference data in fracture risk assessment in the UK.

	Introduction

Top Abstract Introduction Materials and methods Results Discussion References

 
An expert panel from the World Health Organisation (WHO) proposed diagnostic guidelines for osteoporosis in women using bone mineral density (BMD) measurements [1]. They defined cut-off criteria for identifying those with moderately low BMD (osteopenia) and definitely low BMD (osteoporosis), the second group being at high risk of osteoporotic fractures. Values between 1 and 2.5 standard deviations (SDs) below the young healthy mean qualified as osteopenia, and values below 2.5 (SDs) identified cases of osteoporosis. These cut-off levels were not initially applied to measure BMD at particular sites and were defined originally for measurements in women only. Subsequently, the authors of a population-based study in the US, the third National Health and Nutrition Examination Survey (NHANES III), have applied these guidelines to estimate cut-offs for diagnosing osteopenia and osteoporosis in the proximal femur and have provided estimates of prevalences of both osteopenia and osteoporosis in subjects over 50 years of age in the US [2, 3] together with distributions of BMD values by 10-year age bands. The proximal femur is now the preferred site in many circumstances for diagnosis using dual X-ray absorptiometry (DXA) [4].

In this paper we investigate the similarities and differences observed in the distribution of femur BMD values between US subjects from NHANES III and seven population-based cohorts recruited in Britain.

	Materials and methods

Top Abstract Introduction Materials and methods Results Discussion References

 
Subjects
Five centres (Aberdeen, Bath, rural Cambridgeshire, Harrow and Truro) randomly recruited subjects aged 50–80+ years, stratified in 5-year age bands, from general practice registers as part of the European Prospective Osteoporosis Study (EPOS) [5–7]. The mean response rate of the five EPOS centres analysed in this study was 55%. From these, random samples from each centre of between 10–100% were invited for bone densitometry.

The Cambridge City Health District centre recruited subjects aged 50–80 years, unstratified by age, from general practitioners' age and sex registers. Approximately 50% of those approached agreed to participate [8]. The seventh centre randomly recruited subjects aged 67–76 years in Norfolk from among those already participating in the European Prospective Investigation into Cancer (EPIC) [9], a multicentre international cohort study designed to look at the association between nutrition, cancers and common chronic diseases. Of the 43% approached who agreed to take part in EPIC, a random sample of the approximately 8000 individuals aged between 65–75 years were invited for bone densitometry measurements and over 90% accepted. Altogether 7426 complete records were included from these EPOS, EPIC and Cambridge study participants (2253 male, 5173 female).

Definition of osteoporosis and osteopenia
Osteopenia and osteoporosis prevalences within the study populations for women were estimated using WHO guidelines [1]. The cut-off values used in this study were derived from results obtained in the NHANES III "non-Hispanic White" young adult reference group [2, 3]. Separate young normal ranges were obtained in the NHANES III study for men and women aged between 20–29 years, because a consensus has arisen that this should be the normal referent age range [4]. Prevalences of osteopenia and osteoporosis, measured at the femoral neck and trochanter regions, within the British cohorts were estimated using the following WHO diagnostic definitions: osteopenia, a BMD value between 1 and 2.5 young NHANES III population SDs below the mean of the young reference group; osteoporosis, a BMD value greater than 2.5 SDs below the young reference mean.

The NHANES III study applied cut-off values to their male data using both male and female young reference populations as alternatives. In the absence of a consensus view on the equivalent diagnostic thresholds in men, we favour the use of the same diagnostic cut-off as for women, an approach endorsed in a recent careful analysis of the existing literature [10], although we also report alternative T-scores for men using NHANES III male reference data as described by Looker et al [2].

Densitometry
The six English centres performed densitometry on the hip using Hologic densitometers, while the Scottish centre in Aberdeen used two Norland densitometers. Five of the six English centres, as well as NHANES III, used Hologic QDR 1000 densitometers, whilst the sixth centre, Bath, used a QDR 4500 (Hologic Inc., Waltham, MA). The BMDs measured in Aberdeen were first converted to standardized units and subsequently converted to Hologic units as described by Pearson et al using the equations described therein [11], so as to make them comparable with the NHANES III data.

To assess whether differences in BMD between British cohorts could be explained by differences in the calibrations of the densitometers used, the British data were cross-calibrated. This was carried out using the European Spine Phantom prototype (ESPp) as described previously [12]. Since it was not possible to cross-calibrate the NHANES III data in the same way, comparisons between the British centres and the reference range were made without cross-calibration.

Statistics
Z-scores were calculated from the age-specific mean values of NHANES III with linear interpolation. This enabled the comparison of BMD in British centres with those of an age matched non-Hispanic White US population. BMD measurements were converted to T-scores using the NHANES III male and female young normal reference data in order to estimate osteopenia and osteoporosis levels. Estimates for males were calculated using both male and female reference population cut-off values. Multiple regression was used to model T-scores and z-scores using age, weight and height as predictors to assess their influence on BMD for men and women. Age, weight and height were fitted as continuous variables. Analysis of variance was used to test for significance of differences between centres in mean values for BMD before and after adjustment for these variables.

Multiple regression was applied to both uncalibrated and calibrated BMD values to test whether differences seen in the uncalibrated results could be explained by the different densitometers used. All statistical modelling was undertaken using SAS (SAS Institute Inc., Cary, NC) or STATA (Stata Corporation, East College Station, TX) software.

	Results

Top Abstract Introduction Materials and methods Results Discussion References

 
Population characteristics
Tables 1 and 2 give a summary of the characteristics of the British population cohorts studied.

Comparison between British centres
There were significant differences between the British centres in mean z-scores (p<0.0001), which remained after adjusting for age, weight and height. These differences did, however, reduce in significance at the trochanter region in women (p<0.1). Only at the femoral neck in women did the rate of change with age differ between centres, from a non-significant reduction in Truro of 0.80 per decade to an increase in z-scores per decade of 0.22 in Norfolk (p<0.001). Significant differences (p<0.0001) between the BMD values of British centres remained when the data were cross-calibrated using the ESPp for both sexes and at both sites. When data had been adjusted for age, weight and height, significant differences between the British centres remained for males (p<0.0001). Among females, centre differences observed at the femoral neck in the uncalibrated BMD values reduced when the data had been cross-calibrated (p<0.1), while at the trochanter region the cross-calibration of data led to larger, and statistically significant, differences between centres.

Comparing British with young US data (T-score calculations)
T-scores are commonly used in diagnosis and are frequently based on the NHANES III data. Not surprisingly, applying the NHANES III young reference values to our British female subjects gave mean T-scores significantly below zero at both sites. Among British men the results were dependent on the reference population (male or female) used to define cut-off values (Table 2). There were significant differences in T-scores between British centres for both men and women at both sites (p<0.001). These differences remained after adjusting for age, weight and height, except at the trochanter region in women. Male T-scores declined by 0.15 SD units per decade, though this was only significant at the femoral neck region. Among women, a reduction in T-scores of 0.47 SD units per decade (p<0.001) was found at the trochanter region, while at the femoral neck the rate of change differed between centres, varying between a reduction of 0.25 SD units and a reduction of 1.3 SD units per decade. T-scores at both sites and for both genders increased by approximately 0.03 SD units per kg increase in body weight.

Estimates of the prevalences of osteopenia and osteoporosis in females at both sites for each of the British centres are given in Table 3. These estimates have been adjusted to age 65 years. There were marked variations between the British cohorts, with osteopenia prevalences ranging from 33–60% and osteoporosis prevalences ranging from 2–11%.

	Discussion

Top Abstract Introduction Materials and methods Results Discussion References

 
For men and women alike, BMD levels appeared different in the British cohorts compared with their US peers. This was particularly evident at the lower end of the BMD distributions in the proportions of subjects most vulnerable to osteoporosis and therefore, in all probability [13], consequent fractures. In addition, our cross-sectional analysis suggested British men appeared to lose bone at a slower rate than their US counterparts, equating to a 0.1 rise in z-score with each decade at both the femoral neck and the trochanter region when the z-score was calculated by reference to US data. This could have been the result of a cohort effect and longitudinal studies will be needed to clarify the true cause of this observation.

The overall prevalences of osteoporosis in the British cohorts, adjusted to age 65 years, were markedly lower than estimates for the US, while osteopenia levels were also lower in the British subjects. This suggests that BMD distributions for the UK and US study cohorts were different, and further analysis indicated a significantly positively skewed distribution for British cohorts. The NHANES III BMD data appeared to come from a normal distribution [2]. However, although the NHANES III investigators have reported a marked effect of body weight on BMD values in the hip [14], data were not reported in a form that allowed us to determine to what degree these differences in osteopenia and osteoporosis rates were owing to differences in body composition in the US and UK populations.

The numbers of men estimated to be suffering from both osteopenia and osteoporosis depend on the criteria for defining male osteoporosis. NHANES III compared the numbers affected when using a young adult male reference population with those obtained from applying the female reference population. A similar comparison was carried out on our data and, as with the US data, clearly showed that numbers of men identified as either osteopenic or osteoporotic varied substantially depending on the sex of the reference population applied. Until more research into male osteoporosis is undertaken it remains unclear whether it has more clinical utility to use male or female reference values, although at present we favour the approach of using the female-specific reference range as suggested by Orwoll [10].

There was also marked variation between the British cohorts in their mean femur BMD. This complements previous work showing disparities across Europe in BMD measurements made at the hip [7, 12]. The cross-calibrated data gave more confidence that differences observed between BMD levels of British centres were owing to "real" differences in BMD and not the result of densitometer variation, particularly for males. These differences could not be explained by the age, weight or height of the subjects recruited.

This was a descriptive study examining how well the NHANES III data corresponded to a number of British population cohorts and whether differences were apparent with regards to BMD levels. One potential drawback of using NHANES III as the reference population in the calculation of T-scores and, when looking at an age-matched population, in the calculation of z-scores, was the fact that it was not possible to cross-calibrate the British BMD values to those of the NHANES III data using the ESPp. However, all except one of the British centres, like NHANES III, used the Hologic brand of DXA machines that are claimed by the manufacturer to give values within 1–2% of one another, providing the users follow their instructions with regard to use of the Hologic spine phantom, and this was done. Measurements made in Aberdeen on the Norland XR26 machine, even though converted to "Hologic units", may not have been so closely cross-calibrated to mean Hologic values as the measurements made on Hologic machines.

Two limitations of this study concern the referent young normal data and choice of hip site(s) for analysis. Concerning the choice of young normal subjects, the US NHANES III young normal data [2] were used because they have become an international referent and no equivalent young normal British data were available. The only possible alternative was the young normal data published by the Quantitative Assessment of Osteoporosis study [12], which were converted into Hologic units by Gibson et al [15] and yielded similar cut-offs to the NHANES III data. However, these data were not population-based. The femoral neck and trochanter regions were chosen as regions of interest before the use of the total hip region became widespread [12]. Since many of the data were submitted on paper records rather than Hologic or Norland database files, access to total hip data from some centres was not available. However, in the large Norfolk centre the prevalence of total hip osteopenia was exactly midway between the prevalences for the femoral neck and the trochanter in women, and the prevalence of total hip osteoporosis was less than 1% different from the prevalence of femoral neck osteoporosis in this elderly female group. We think it unlikely that differences in prevalence of osteoporosis between our study and NHANES III can be attributed to not choosing total hip as our region of primary interest, since in NHANES III total hip osteoporosis and osteopenia prevalences were also similar to those for the other two regions of interest [2].

The possibility that selection or non-response bias might have influenced results needs to be considered. NHANES III had a slightly higher response rate at 61% [16] compared with the combined British cohorts response rate of 55%. It is possible that a higher proportion of those most vulnerable to osteoporosis did not participate from the British cohorts compared to NHANES III. However, O'Neill et al [17] investigated the characteristics of non-responders compared with responders in the European Vertebral Osteoporosis Study, the prevalence phase of EPOS, in relation to their lifestyle and dietary risk factors. O'Neill et al concluded from this analysis that while some differences in the characteristics of responders and non-responders were evident, it was unlikely that these differences would have had a major influence on prevalence estimates for fractures or osteoporosis.

The demonstrated differences in osteoporosis rates between the UK and US study populations has a parallel in the analysis of Petley et al [18]. They found that by applying reference ranges calculated from both locally recruited subjects and from a national population supplied by the manufacturer, there were substantial discrepancies in estimated rates of osteoporosis in a Southampton cohort of women.

In conclusion, it has been shown that population samples over the age of 50 years in Britain show clinically significant differences in the shapes of the distributions of BMD as well as in the mean values of BMD measured at the hip compared with those in the US. These led, in particular, to different estimates of osteoporosis prevalence rates in the two countries. Modest differences between mean BMD values of British populations from different localities were also observed. Fracture rates are known to vary over a two-fold range within Britain [19], so this finding was not unexpected. Using NHANES III data as a referent for the diagnosis of osteoporosis in British subjects will generate different rates of osteoporosis at the population level from those in North America. This lower expected diagnosis rate of osteoporosis in Britain might have implications for referral rates for bone densitometry as well as for calculations of the public health impact of osteoporosis. Nevertheless, in subjects who do have a BMD measurement, it might still be appropriate to apply the US data for fracture risk assessment in Britain, providing the quantitative risk relationship between future fracture and BMD is the same in the two countries.

	Footnotes

 
Current address for Gemma Holt: Institute for the Geography of Health, Department of Geography, University of Portsmouth, Buckingham Building, Lion Terrace, Portsmouth PO1 3AS, UK.

Current address for Nicola Crabtree: Department of Nuclear Medicine, Queen Elizabeth Hospital, Birmingham B15 2TH, UK.

Current address for Mark Lunt: ARC Epidemiology Unit, University of Manchester Stopford Building, Oxford Road, Manchester M13 9PT, UK.

Presented in part at the American Society for Bone & Mineral Research 1998 (GH was the recipient of a Young Investigator award).

Received for publication August 21, 2001. Revision received February 14, 2002. Accepted for publication April 16, 2002.

	References

Top Abstract Introduction Materials and methods Results Discussion References

 

1. Kanis J, Melton III L, Christiansen C, Johnston C, Khaltaev N. Perspective: the diagnosis of osteoporosis. J Bone Miner Res 1994;9:1137–41.[Medline]
2. Looker AC, Orwoll ES, Johnston CC Jr, Lindsay RL, Wahner HW, Dunn WL, et al. Prevalence of low femoral bone density in older US adults from NHANES III. J Bone Miner Res 1997;12:1761–8.[Medline]
3. Looker AC, Wahner HW, Dunn WL, Calvo MS, Harris TB, Heyse SP, et al. Updated data on proximal femur bone mineral levels of US adults. Osteoporos Int 1998;8:468–89.[Medline]
4. National Osteoporosis Foundation. Osteoporosis: review of the evidence for prevention, diagnosis and treatment and cost-effectiveness analysis. Osteoporos Int 1998;8(Suppl. 4):S7–S80.
5. O'Neill TW, Felsenberg D, Varlow J, Cooper C, Kanis JA, Silman AJ, et al. The prevalence of vertebral deformity in European men and women: the European Vertebral Osteoporosis Study. J Bone Miner Res 1996;11:1010–7.[Medline]
6. Lunt M, Felsenberg D, Reeve J, Benevolenskaya L, Cannata J, Dequeker J, et al. Bone density variation and its effects on risk of vertebral deformity in men and women studied in 13 European centres: the EVOS Study. J Bone Miner Res 1997;12:1883–94.[Medline]
7. Lunt M, Felsenberg D, Adams J, Benevolenskaya L, Cannata J, Dequeker J, et al. Population-based geographic variations in DXA bone density in Europe: the EVOS study. Osteoporos Int 1997;7:175–89.[Medline]
8. Murphy S, Khaw K-T, May H, Compston JE. Milk consumption and bone mineral density in middle-aged and elderly women. BMJ 1994;308:939–40.[Abstract/Free Full Text]
9. Jakes RW, Khaw KT, Day NE, Bingham S, Welch A, Oakes S, et al. Patterns of physical activity behaviour and heel bone density: the EPIC-Norfolk study. BMJ 2001;322:140–3.[Abstract/Free Full Text]
10. Orwoll E. Assessing bone density in men. J Bone Miner Res 2000;15:1867–70.[Medline]
11. Pearson J, Dequeker J, Henley M, Bright J, Reeve J, Kalender W, et al. European semi-anthropomorphic spine phantom for the calibration of bone densitometers: assessment of precision, stability and accuracy. Osteoporos Int 1995;5:174–84.[Medline]
12. Pearson J, Dequeker J, Reeve J, Felsenberg D, Henley M, Bright J, et al. Dual X-Ray absorptiometry of the proximal femur: normal European values standardised with the European spine phantom. J Bone Miner Res 1995;10:315–24.[Medline]
13. Marshall D, Johnell O, Wedel H. Meta-analysis of how well measures of bone mineral density predict occurrence of osteoporotic fractures. BMJ 1996;312:1254–9.[Abstract/Free Full Text]
14. Beck TJ, Looker AC, Ruff CB, Sievanen H, Wahner HW. Structural trends in the ageing femoral neck and proximal shaft: analysis of the third national health and nutrition examination survey dual-energy X-ray absorptiometry data. J Bone Miner Res 2000;15:2297–308.[Medline]
15. Gibson JH, Harries M, Mitchell A, Godfrey R, Lunt M, Reeve J. Determinants of bone density and prevalence of osteopenia among female runners in their second to seventh decades of age. Bone 2000;26:591–8.[Medline]
16. Looker AC, Johnston CCJ, Wahner HW, Dunn WL, Calvo MS, Harris TB, et al. Prevalence of low femoral bone density in older US women from NHANES III. J Bone Miner Res 1995;10:796–802.[Medline]
17. O'Neill TW, Marsden D, Silman AJ, on behalf of the European Vertebral Osteoporosis Study Group. Differences in the characteristics of responders and nonresponders in a prevalence survey of vertebral osteoporosis. Osteoporos Int 1995;5:327–34.[Medline]
18. Petley GW, Cotton AM, Murrills AJ, Taylor PA, Cooper C, Cawley MID, et al. Reference ranges of bone mineral density for women in southern England: the impact of local data on the diagnosis of osteoporosis. Br J Radiol 1996;69:655–60.[Abstract]
19. Cooper C. Epidemiology and public health impact of osteoporosis. In: Reid DM, editor. Baillière's Clinical Rheumaology. London: Baillière Tindall; 1993;459–77.

This article has been cited by other articles:

				J.L. Newton, D.E.J. Jones, K. Wilton, J. Pairman, S.W. Parry, and R.M. Francis Calcaneal bone mineral density in older patients who have fallen. QJM, April 1, 2006; 99(4): 231 - 236. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Holt, G Articles by Reeve, J Search for Related Content PubMed PubMed Citation Articles by Holt, G Articles by Reeve, J