SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

We used an experimental software measuring the hip axis length (HAL) and bone mineral density (BMD) in specific regions of the lower and upper part of the femoral neck on dual-energy X-ray absorptiometry scans. To determine whether these parameters were significant predictors of the type of hip fracture, we measured 167 healthy women (controls), 24 women with trochanteric, and 42 women with cervical hip fractures within the EPIDOS prospective cohort. EPIDOS is a multicenter prospective study on risk factors for hip fracture performed in 7575 elderly women living at home, aged 75–95 and conducted in five French centers (Amiens, Lyon, Montpellier, Paris, Toulouse). Measurements were performed on data acquired at baseline before the occurrence of fracture. In the cervical fracture group, HAL was significantly longer than in controls (94.2 vs. 92.3, p = 0.03), and the associated odds ratio (OR) adjusted for age, weight, and total femoral neck BMD was significant (OR = 1.64, 95% confidence interval [CI] 1.06–2.55). In contrast, HAL was not significantly different from controls in the trochanteric fracture group. Femoral neck diameter was not a predictor of fracture. The upper and lower femoral neck BMD was lower in the trochanteric fracture group than in controls, and both measurements predicted trochanteric femoral neck fracture. In contrast, the prediction of cervical femoral neck fracture was enhanced by measuring only the upper part of the femoral neck (OR = 2.79 vs. 1.97 for the total femoral neck) while BMD of the lower part was not different from controls. Hip axis length is a predictor of femoral neck fracture. Femoral neck BMD distribution is different between cervical and trochanteric fractures. These results support the hypothesis of a different pathophysiological mechanism between the two types of hip fractures.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

ACCORDING TO BASIC ENGINEERING principles, the mechanical strength of an object is related to physical properties of the material, its geometry, and loading conditions. In vivo, both hip geometry and bone mineral density (BMD) have been shown to relate independently to hip fracture risk. The influence of geometric parameters on the type of hip fracture has been demonstrated first by Glüer et al.1 who performed measurements on conventional radiographs. They demonstrated that the width of the femoral head influences the risk of cervical fracture and that the width of the intertrochanteric region influences the risk of trochanteric fracture. Acetabular bone width was strongly associated with cervical fracture only. The predictive value of hip axis length (HAL) for hip fracture was well documented by Faulkner et al.2 on 64 patients with hip fracture compared with 134 controls in the SOF (Study of Osteoporotic Fracture) cohort. These authors performed manual measurements on dual-energy X-ray absorptiometry (DXA) printouts and reported that an increase of 1 SD in HAL was associated with an increased risk of both cervical (odds ratio [OR] = 1.9, 95% confidence interval [CI] 1.3–3.0) and trochanteric fracture (OR = 1.6, 95% CI 1.0–2.4). In 1995,3 the same authors suggested including automatic HAL measurement within the existing conventional densitometric software. Cummings et al.4 suggested that a shorter HAL might be a major factor accounting for Asian women's lower risk of hip fracture and might contribute to the lower risk in black women. Similar results were observed by Nakamura et al.5 who reported that geometric characteristics of the femoral neck in Japanese women are associated with their lower hip fracture risk. They concluded that femoral geometry, combined with bone mass, may provide further clinical information about the risk of hip fracture. However, Nelson et al.6 reported that the lower hip fracture risk observed in black men, compared with white men, is not due to a difference in hip axis length. Total body and femoral neck BMD measurements have been proven to predict strongly hip fracture in the EPIDOS cohort.7 Another interesting feature concerning BMD was reported by Vega et al.8 who observed that females with trochanteric femoral neck fractures are older, thinner, have lower BMD at all measured sites, and have more frequent previous vertebral fractures than patients with cervical femoral neck fractures. In 1996, a review published by the same group suggested a different pathophysiological mechanism between cervical and trochanteric femoral neck fracture.9

Using a prospective approach and a new experimental densitometric software in a subset of the large EPIDOS cohort, we address the following three issues:

(1) Is the automated evaluation of HAL useful to predict hip fracture and specifically the type of hip fracture?

(2) Is femoral neck BMD distribution similar or different in trochanteric and cervical fractures? Specifically, we looked at the BMD of the lower (basal) and upper (cranial) part of the femoral neck, which are submitted to different stress, respectively compressive and tensile stresses.10

(3) Are measurements of other geometric parameters such as femoral neck diameter able to predict hip fracture?

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Subjects

Women were selected from the EPIDOS cohort. EPIDOS is a multicenter prospective study on the risk factors for hip fracture performed in 7575 elderly women living at home, aged 75–95 years, and conducted in five French centers (Amiens, Lyon, Montpellier, Paris, Toulouse). This study received the approval of local ethical committees, and all the women included in the cohort gave informed consent. Hip BMD was measured once at baseline, and women were checked every 4 months for fracture events by using phone questionnaire. After an average of 2-years of follow-up, 155 hip fractures occurred in the whole cohort. We excluded women from this analysis if they had hip prosthesis, Paget's bone disease, malignant bone disease, renal failure, hyperthyroidism, or treated hypothyroidism.

The data presented in this paper were collected only in the Lyon and Montpellier centers where 80 hip fractures occurred during the 2-year follow-up. These 80 fractured women were compared with a control group of 240 age-matched (three controls for one fracture case) women randomly selected from those who did not suffer a hip fracture during the follow-up period.

DXA measurements and hip axis length

Hip BMD was measured on the nondominant side, with a Lunar DPX Plus osteodensitometer (Lunar Corporation, Madison, WI, U.S.A.) in each of the two centers. All the exams were reanalyzed using an experimental and noncommercialized software (3.7, β version, Lunar Corporation, Madison, WI, U.S.A.) developed to estimate the strength of the femoral neck. Specific parameters were the HAL (from the trochanter to the cotyle) and BMD values of the upper and lower part of the femoral neck region (Fig. 1). This specific software was previously validated in healthy women and men by Yoshikawa et al.11

thumbnail image

Figure FIG. 1.. Specific regions of interest and geometric parameters assessed by DXA with a new experimental software (Lunar Corporation, Madison, WI, U.S.A.).

Download figure to PowerPoint

Because the cotyle was not always included in the scanned area, HAL was consequently not measurable in 27% of the subjects. Among the 320 scans reanalyzed, 233 were measurable, including 167 controls, 24 trochanteric, and 42 cervical hip fractures.

To determine the coefficient of variation of each parameter, two measurements with repositioning were performed on seven women (aged 57.8 ± 11.7 years). The coefficients of variation are shown on Table 1. In vitro accuracy of the femoral neck diameter and HAL measurements have been performed by using a homemade aluminum phantom, having the shape of the upper femur. Two measurements were performed in a 16 cm water filled container. Femoral neck diameter measurements performed with the Lunar software were 4.5% higher than the caliper measurements (30.5 vs. 28.7 cm), and HAL measurements were 3.2% lower than the caliper measurements (10.84 vs. 11.19 cm).

Table Table 1.. ACCURACY AND REPRODUCIBILITY OF DENSITOMETRIC MEASUREMENTS PERFORMED WITH AN EXPERIMENTAL SOFTWARE (LUNAR R)
Thumbnail image of

Because hip densitometric measurements require a correction of the femoral anteversion which could bias HAL measurements, we tested the effect of the rotation on the HAL value on 17 subjects (age 46 ± 15). The mean values were 93.7 ± 6.6 mm with correction and 92.4 ± 6.9 mm without correction of the 15° anteversion of the hip. These values are significantly different when compared by a paired t-test (p = 0.0027), but lower than the theoretical difference (length mutiplied by the cosinus of the angle) estimated to 3.19 mm. Because femoral rotation similarly influences fractures and control group, we neglected this difference in the study.

Statistical analysis

The Gaussian distribution of all measured parameters was verified. To compare the value of each parameter between the two femoral neck fracture groups (cervical and trochanteric) and the controls, an unpaired t-test was performed. The sample was then divided into quartiles with respect to HAL, upper femoral neck BMD, lower femoral neck BMD, total neck BMD, and femoral trochanter BMD. We estimated the OR for the different quartiles derived from 2 × 2 tables. We used a specific Chi square trend test to detect any significant linear trend between risk factors and the number of hip fracture scored the quartiles by 1, 2, 3, and 4 in order to test an association between BMD versus fracture.12 Finally, simple and multiple unconditional logistic regressions were used to evaluate the relative increase of the fracture risk for a variation of 1 standard deviation (SD) of each parameter expressed as an odd ratio standardized for 1 SD (ORSD). This was calculated with the formula (OR = e(β ∗ SD) where β is the logistic regression coefficient and SD the standard deviation of the whole population.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Descriptive analysis

Demographic characteristics of the 233 women and the average values of the measured parameters are given in Tables 2A and 2B. The whole hip fracture group was older and had lower BMD than controls, regardless of the site measured. Trochanteric femoral neck fractures differed from controls by age, weight, height, and BMD, while cervical femoral neck fractures differed from controls for total femoral neck BMD, upper femoral neck BMD, and HAL.

Table Table 2A.. CHARACTERISTICS OF THE WHOLE POPULATION
Thumbnail image of

Hip axis length

Regarding the whole hip fracture group, HAL does not appear significantly different from controls. A separate analysis on women who subsequently sustained a cervical fracture shows a longer HAL than controls, while no difference is observed in women with trochanteric fractures (Table 2B). Thus, additional analysis were performed on each group of hip femoral neck fracture. HAL data were divided into quartiles to calculate OR relative to the lowest quartile (quartile 1, 80.1–88.8 mm). In the cervical fracture group, a 2.07 increase of risk was found in the highest quartile (quartile 4, 96.2–106.3 mm), but the trend of OR increase for each quartile did not reach significance (Qui square trend test). In the trochanteric fracture group, none of the calculated OR were significant with this test. Figure 2 shows the change of the OR from the lowest to the highest HAL quartiles for the two types of hip fractures. Simple logistic regression also showed significant results for the HAL in the cervical fracture group but not in the trochanteric fracture group (Table 3). Because age, weight, and BMD are related to hip femoral neck fracture risk, and weight and BMD are related to hip axis length, we adjusted analysis for these parameters. Significance did not change after adjustment. Table 3 summarizes the results of the simple and adjusted logistic regression analysis.

thumbnail image

Figure FIG. 2.. Odds ratio according to quartiles (quartile 1, 80.1–88.8 cm; quartile 2, 88.9–92.7; quartile 3, 92.7–96; quartile 4, 96.2–106.3) of hip axis length (HAL) in trochanteric and cervical fractures.

Download figure to PowerPoint

Table Table 2B.. CHARACTERISTICS OF THE TWO FRACTURED POPULATIONS
Thumbnail image of
Table Table 3.. SIMPLE AND ADJUSTED (AGE, WEIGHT, AND TOTAL NECK BMD) ODDS RATIO FOR AN INCREASE OF 1 SD OF HIP AXIS LENGTH IN THE WHOLE POPULATION WITH HIP FRACTURE AND IN THE TWO TYPES OF HIP FRACTURE SEPARATELY
Thumbnail image of

Regional BMD

In the whole population, BMD measurements appear predictive of risk fracture, regardless of the region of interest. The unconditional adjusted logistic regression analysis shows OR varying from 2.68 (1.75–4.10) for the upper part of femoral neck, to 1.60 (1.13–2.25) for the lower part of the femoral neck.

In the two hip femoral neck fracture groups, total femoral neck BMD, trochanter femoral BMD, and upper femoral neck BMD were significantly different from controls. In the cervical fracture group, the lower femoral neck BMD was not significantly different from controls. A summary of the t-test results is shown in Table 2B. For the upper femoral neck BMD, a regular and significant increase of the OR associated with the decrease of the BMD values was observed in the two hip fracture groups (Chi square trend = 10.53 and 13.58, respectively, for trochanteric and cervical hip fractures, p < 0.05). In contrast, for the lower femoral neck BMD, the Chi square trend test was not significant in the cervical fracture group. Values of the OR for each quartile of upper femoral neck BMD (quartile 1, 0.563–0.935; quartile 4, 0.119–0.446) and lower femoral neck BMD (quartile 1, 0.954–1.247; quartile 4, 0.524–0.784) are plotted in Figs. 3A and 3B for the two types of fracture.

thumbnail image

Figure FIG. 3.. (a) Odds ratio according to quartile of the upper and lower femoral neck BMD in cervical fracture. (b) Odds ratio according to quartile of the upper and lower femoral neck BMD in trochanteric fracture. Range of the upper neck BMD quartiles: quartile 1, 0.563–0.935 g/cm2; quartile 2, 0.495–0.561 g/cm2; quartile 3, 0.447–0.493; quartile 4, 0.119–0.446. Range of the lower neck BMD: quartile 1, 0.954–1.247 g/cm2; quartile 2, 0.869–0.953; quartile 3, 0.784–0.869; quartile 4, 0.524–0.784.

Download figure to PowerPoint

All the regional hip BMD measurements (total femoral neck, upper and lower femoral neck BMD, trochanter femoral BMD) were significantly predictive of trochanteric fracture (Table 4). In the cervical fracture group, the lower femoral neck BMD was not. Because age, weight, height, and neck diameter are linked with BMD, OR associated with BMD parameters were corrected for these confounding factors. Significance did not change after adjustment. Unadjusted and adjusted ORs are presented in Tables 4 and 5.

Table Table 4.. SIMPLE AND ADJUSTED (AGE, HEIGHT, WEIGHT, AND NECK DIAMETER) ODDS RATIO FOR A DECREASE OF 1 SD OF BONE MINERAL DENSITY IN THE WHOLE FRACTURED POPULATION
Thumbnail image of
Table Table 5.. SIMPLE AND ADJUSTED (AGE, HEIGHT, WEIGHT, AND FEMORAL NECK DIAMETER) ODDS RATIO FOR A DECREASE OF 1 SD OF BONE MINERAL DENSITY IN THE TWO TYPES OF HIP FRACTURE
Thumbnail image of

Other geometric parameters

A significant difference was observed between the femoral neck diameter of the trochanteric fracture and the control group (p = 0.026), but not between the cervical fracture and the control group (p = 0.95). Femoral neck diameter was a significant predictor of trochanteric fractures in simple regression model (OR 1.47; 95% CI 1.02–2.13) but did not reach significance after adjustment for age, weight, height, and BMD (OR 0.946; 95% CI 1.50–2.40).

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

This nested case-control study has several methodological advantages; all measurements were performed on data acquired before the occurrence of hip fractures, and cases and controls are part of the same cohort, which ensure their comparability. We used an experimental software to estimate different geometric and specific densitometric parameters of the upper extremity of the femur. We report a longer HAL trend and an inhomogeneous femoral neck bone mass distribution in the cervical femoral neck fracture group, compared with controls. In contrast, HAL was not longer than in controls and femoral neck BMD was homogeneously decreased in the trochanteric fracture group.

Regarding the validity of measurements, variability of total femoral neck BMD measurements reported in our study (1.83%) agree with those generally reported (i.e., 1.5–3%).13–15 The higher variability of the upper femoral neck BMD measurements might be explained by a smaller BMD value (0.515 ± 0.01 g/cm2) than that observed in the lower femoral neck region (0.873 ± 0.13 g/cm2) and the total femoral region (0.680 ± 0.11 g/cm2). The number of women was sufficient to determine a minimum significant difference of 0.09 g/cm2 in BMD between fracture and control groups with a statistical power of 80%. The relatively high coefficient of variation reported for the femoral neck diameter (1.95%) could explain the nonsignificant results obtained with this parameter. Given the number of subjects in our study, the minimum difference that we could observe with a 5% of type I and a 20% of type II statistical error is 2.5 mm for the trochanteric hip fracture group and 1.85 mm for the cervical hip fracture group. The coefficient of variation determined for HAL measurements (1.12%) is slightly higher than that reported by Faukner et al.2 probably because of difficulties in identification of the bone edge. In addition, cotyle edge is less clearly determined than inner pelvic brim edge.

t-test analysis showed a clear difference of HAL between the cervical femoral neck fracture group and the controls, while the difference did not reach significance in the trochanteric fracture group. This result was confirmed by the logistic regression analysis which demonstrated a significant predictive value of HAL in the cervical fracture group, but not in the trochanteric fracture group. These data differ from those of Faulkner et al.2 who reported a significant predictive value of HAL for both types of hip fracture. However, these authors reported a higher OR for the cervical (OR = 1.9; 95% CI 1.3–3.0) than for the trochanteric hip fracture group (OR = 1.6; 95% CI 1.0–2.4) in a group of 64 hip fractures.

Influence of HAL on hip fracture risk has already been reported by other researchers, who strongly support influence of femoral geometry.2,3,4,16,17 Recently, a subset of women from the Study of the Osteoporotic Fracture were investigated to determine whether HAL was associated with fractures other than those of femoral neck. HAL values were not significantly different from those of controls for all fracture groups, except for the hip femoral neck fracture group (OR = 1.8). Authors concluded that HAL is specifically associated with hip femoral neck fractures and not with other osteoporotic fractures. Hip femoral neck fracture mechanism seems specifically related to the shape and structure of the hip fractures.15

In our study, HAL evaluation suffers from some limitations. First, it is possible that the nonsignificant results in the trochanteric femoral neck fracture group is related to the small sample size, although significant differences were observed for other parameters such as BMD. Second, the percentage of nonavailable measurement was higher in the control group (30.5%) than in the fracture group (17.5%). Positioning was the main reason for excluding a patient, which should not be related to the type of fracture, although we cannot eliminate a potential bias. We have found that all hip femoral BMD measurements were strongly predictive of trochanteric hip fracture with OR ranging from 2.6 to 4.5 according to the area of measurement. In contrast, femoral neck fracture predictions were higher with the measurements of the lower part of the femoral neck (OR = 2.8) than with measurements of other sites (OR = 1.3–1.9). Women who sustain trochanteric femoral neck fractures have significantly reduced BMD at all sites compared with women who sustain cervical femoral neck fractures.7,18 This difference has recently been confirmed also with ultrasound technique at the level of the calcaneum.19,20 The experimental software used, previously validated by Yoshikawa et al. in 199411 shows a more important age-related decrease of the upper femoral neck BMD versus the lower femoral neck BMD ratio in the female population than in the male population, suggesting a greater loss of bone in the upper half of the femoral neck in women. Our data provide different and additional information on the importance of the difference between upper and lower femoral neck BMD for bone strength.

This difference could be explained, from a biomechanical point of view, by a difference in the compressive (lower femoral neck) and tensile (upper femoral neck) stress on the hip, which may be related to differences in femoral neck length and shape. Consequently, these stress differences could cause an inhomogenous femoral neck bone mass distribution. Faulkner3 suggested that the size and shape of the hip can explain a portion of the observed etiological difference between the different types of osteoporotic fractures. Age and hip BMD seem to play a more important role in the trochanteric than in the cervical hip fracture. Women sustaining trochanteric fractures are older and have lower BMD than women sustaining cervical fractures. Those sustaining cervical fractures appear to have a longer HAL, and/or have a different stress pattern, leading to a different BMD distribution and consequently lower density in the upper part of the femoral neck region.

We conclude that HAL might be a potential predictor of the femoral neck fracture risk, but not of the trochanteric fracture risk. Upper femoral neck BMD measurement is a better predictor of cervical fracture than other areas of the hip. Determination of these two parameters could therefore enhance the prediction of hip fracture. We suggest that hip BMD measurement by DXA should include BMD of both the upper and lower parts of femoral neck.

EPIDOS Study participants

Coordinators: G. Bréart and P. Dargent Molina (epidemiology); P.J. Meunier and A.M. Schott (clinical aspects); D. Hans (DPXA and ultrasound quality control); P.D. Delmas (biochemistry).

Principal investigators (centers): C. Baudoin and J.L. Sebert (Amiens); M.C. Chapuy and A.-M. Schott (Lyon); F. Favier and C. Marcelli (Montpelliers)); C.J. Menkes, C. Cormier, and E. Hausherr (Paris); H. Grandjean and C. Ribot (Toulouse).

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

We are grateful to Mrs. G. Genolet and P. Roy for excellent technical assistance. We are grateful to Mr. Ph. Garrido for providing us electronic copies of the exams. The patients were recruited in the EPIDOS Study which was supported by a contract INSERM/MSD-Chibret.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES
  • 1
    Glüer CC, Cummings ST, Pressman A, Li J, Glüer K, Faulkner KG, Grampp S, Genant HK, and the study of the osteoporotic fractures research group 1993 Prediction of hip fractures from pelvic radiograph: The study of the osteoporotic fractures J Bone Miner Res 9:671677.
  • 2
    Faulkner KG, Cummings SR, Black D, Palermo L, Glüer CC, Genant HK 1993 Simple measurement of femoral geometry predicts hip fracture: The study of osteoporotic fractures J Bone Miner Res 8:12111217.
  • 3
    Faulkner KG, McClung M, Cummings SR 1993 Automated evaluation of hip axis length for predicting hip fracture J Bone Miner Res 9:10651070.
  • 4
    Cummings SR, Cauley JA, Palermo L, Ross PD, Wasnich RD, Black D, Faulkner KG, for the study of the osteoporotic research group 1994 Racial differences in hip axis length might explain racial differences in rates of hip fracture Osteoporos Int 4:226229.
  • 5
    Nakamura T, Turner CH, Yoshikawa T, Slemenda CW, Peacock M, Burr DB, Mizuno Y, Orimo H, Ouchi Y, Johnston CCJR 1994 Do variation in hip geometry explain differences in hip fracture risk between japanese and white americans? J Bone Miner Res 9:10711076.
  • 6
    Nelson DA, Jacobsen G, Barondess DA, Parfitt AM 1995 Ethnic differences in regional bone density, hip axis length, and life style variables among healthy black and white men J Bone Miner Res 10:782787.
  • 7
    Schott AM, Cormier C, Hans D, Favier F, Hausherr E, Dargent P, Delmas PD, Ribot C, Sebert JL, Bréart G, Meunier PJ 1995 Total body and femoral neck BMD predict hip fracture in 7598 population-based very elderly women: The EPIDOS prospective study. Seventeenth Annual Meeting of the American Society for Bone and Mineral Research, Baltimore Convention center, Baltimore, Maryland, September 9–13, 1995 J Bone Miner Res 10 (Suppl 1):abstract 29.
  • 8
    Vega E, Mautalen C, Gomez H, Garrido A, Melo L, Sahores AO 1991 Bone mineral density in patients with cervical and trochanteric fractures of the proximal femur Osteoporos Int 1:8186.
  • 9
    Mautalen CA, Vega EM, Einhorn TA 1996 Are the etiologies of cervical and trochanteric hip fractures different? Bone 18:133S137S.
  • 10
    Pauwels F, 1977 Biomécanique de la hanche saine et pathologique, un atlas. Principes, technique et résultats d'une thérapeutique causale. Springer-Verlag, Berlin, Germany.
  • 11
    Yoshikawa T, Turner CH, Peacock M, Slemenda CW, Weaver CM, Teegarden D, Markwardt P, Burr DB 1994 Geometric structure of the femoral neck measured using dual-energy X Ray absorptiometry J Bone Miner Res 9:10531064.
  • 12
    BIOSTATISTICS: A Methodology for the Health Science. Fisher LD, Van Belle G, (eds.) 1993 A Wiley-Interscience Publication, John Wiley & Sons, Inc., New York, NY, U.S.A.
  • 13
    Johnson J, Dawson-Hughes B 1991 Precision and stability of dual energy X ray measurements Calcif Tissue Int 49:174178.
  • 14
    Laskey MA, Flaxman ME, Barber RW, Trafford S, Hayball MP, Lyttle KD, Crisp AJ, Compston JE 1991 Comparative performance in vitro and in vivo of Lunar DPX and Hologic QDR-1000 dual energy X ray absorptiometers Br J Radiol 64:10231029.
  • 15
    Pye DW, Bassey EJ, Amstrong A, 1992 Precision in practice: Experience with a lunar DPX-L densitometer. Bath Conference on Osteoporosis and Bone Mineral Measurement, Bath, U.K., British Institute of Radiology, Vol P21.
  • 16
    Faulkner KG 1995 Hip axis length and osteoporotic fractures J Bone Miner Res 10:506508.
  • 17
    Boonen S, Koutri R, Dequeker J, Aerssens J, Lowet G, Nijs J, Verbeke G, Lesaffre E, Geusens P 1995 Measurement of femoral geometry in type I and type II osteoporosis: Difference in hip axis length consistent with heterogeneity in the pathogenesis of osteoporotic fractures J Bone Miner Res 10:19081912.
  • 18
    Erikson SAV, Widhe TL 1988 Bone mass in women with hip fracture Acta Ortho Scand 59:1923.
  • 19
    Mautalen C, Vega E, Gonzalez D, Carrilero P, Otano A, Silbermann F 1995 Ultrasound and dual X ray absorptiometric densitometry in women with hip fracture Calcif Tissue Int 57:165168.
  • 20
    Schott AM, Weil Engerer S, Hans D, Duboeuf F, Delmas PD, Meunier PJ 1995 Ultrasound discriminates patients with hip fracture equally well as dual-energy X-ray absorptiometry and independently of bone mineral density J Bone Miner Res 10:243249.