Summary
The broadband ultrasound attenuation (BUA) of the right calcaneus was measured in 78 women with hip fracture (45 trochanteric and 33 cervical) and in 85 normal women (mean age 78 years and 63 years respectively). All cases were postmenopausal and the hip fractures had resulted from a fall from standing position or less. The women with hip fracture had lower BUA (41 dB/MHz) than normals (67 dB/MHz), but about 10 dB/MHz of this difference was associated with age.
There was no significant difference in BUA between cases with trochanteric (41 dB/MHz) and those with cervical fracture (39 dB/MHz). It is worth noting that 35.5 % of the women with trochanteric fracture were suffering from osteoarthrosis of the knees while only 15% of the women with cervical fractures had the same problem. Over the 90% of the cases with hip fracture had BUA lower than 61 dB/MHz and the 80% lower than 51 dB/MHz.
Keywords: Os-calcis; Postmenopausal; Trochanteric-cervical fractures; Ultrasound.

Introduction
The measurement of the Broadband Ultrasound Attenuation (BUA) of the calcaneus is referred as providing useful informations concerning bone status, because it seems to express bone density and bone architecture [1-5]. The determination of the BUA values in normal and osteoporotic population has been attempted by several authors [2, 6, 7]. In a large prospective study, it was showed that BUA was significantly lower in women who sustained hip fracture during a three-year follow-up than in those who did not [8].
In a previous study, we have reported decreased BUA values in postmenopausal women with fragility fractures [9]. The present study concerns a greater number of low energy hip fractures in postmenopausal women and compares the BUA of the calcaneus in the fracture and normal cases. Also, is compared the BUA of the calcaneus between the cases of trochanteric fractures and those with cervical fracture to see if the bone status - as expressed by the measurement of the BUA - among these groups is responsible for the localization of the fractures of the upper femur.

Materials and methods
The BUA of the calcaneus was measured in 78 concecutive female patients with fracture of the upper femoral end. The mean age was 78 years (51-97) and all women were postmenopausal. Nonambulatory women prior the hip fracture, or women with disease or condition known to affect the bone stutus, were excluded. Forty-five cases concerned trochanteric fractures (59-93 years, mean age 79) and 33 cervical fractures (51-97 years, mean age 78).
The fractures were fragility type, resulting from a low energy fall from standing position or less. The measurements were carried out at the right calcaneus 0-14 days from the time of the accident. Each woman with fracture had taken radiographic examination of the knees and the existence of osteoarthrosis was recorded.
Eighty five postmenopausal women (41-87 years, mean age 63) without fracture consisted the control group. They were derived from a greater group of healthy asymptomatic postmenopausal women, examined during a program for the determination of normal BUA values in the female population in Crete. They were taken from the population registers of the hospital catchment area. They had no history of previous atraumatic fracture and there was no disease or condition known to affect the bone status.
Each woman in the control group had a lateral calcaneus radiograph taken, and most also had lateral spine radiographs. There was no evidence of osteoporotic vertebrae, and the trabecular pattern of the calcaneus was normal according to the method of Jhamaria et al. (1983) [10], which correlates with Singh's index and with bone density [11]. In 6 cases 71-87 years of age, the trabecular pattern was borderline and was considered normal.
The BUA measurement was performed in all our cases by means of a computer-controlled bone analyzer UBA-575 (Walker Sonix, USA). Short term reproducibility was assessed in our previous study by recording 5 independent measurements of BUA in each of 5 women. The coefficient of variation was 3.8 ± 1.4 % [2].
Data are presented as means ± standard deviation. The mean BUA among the cases with trochanteric fracture, the cases with cervical fracture and the normal cases, were compared using Analysis of Variance (ANOVA). Analysis of covariance (ANCOVA) and age-matched control were carried out to compare the differences in BUA between the fracture cases and the normal cases without the dependence on the age. The difference in the BUA among the cases with trochanteric fractures and osteoarthrosis of the knees and the cases with cervical fracture and osteoarthrosis was tested using the t-test (Student's).

Results
The BUA in the cases with trochanteric fracture was 41 ± 11 dB/MHz and 39 ± 11 in the cases with cervical fracture. These values were significantly different from the BUA found in the normal cases which was 67 ± 10 dB/MHz (p<0.0001). There was no significant difference in the BUA of the two fracture groups (table 1). Using analysis of covariance the above differences between means, remained significant (p<0.0001) without the dependence on the age. Thirty-two age-matched cases (32 cases with hip fracture and 32 normal) showed mean age 66 years and mean BUA 45 and 61 dB/MHz for hip fracture and normal cases respectively.
Osteoarthrosis of the knees was found in 16 out of 45 cases with trochanteric fracture and in 5 out of 33 cervical fracture. The difference in BUA among these groups of cases was not significant.
In 72 out of 78 cases with fracture the BUA values were below 61 dB/MHz and in 63 out of 78 the BUA values were below 51 dB/MHz. In 25 out of 85 normal cases the BUA was lower than 61 dB/MHz and in 6 out of 85 the BUA was below 51 dB/MHz. In the cases age-matched, 87.5% (28/32) of the hip fractures and 34% (11/32) of the normals the BUA was below the 61 dB/MHz . The corresponding percentages for BUA lower than 51 dB/MHz were 75% (24/32) and 12.5% (4/32).

Discussion
Low BUA values have been reported in osteoporotic conditions as well as in cases with fragility fractures [1,2,5,12,13]. In the present study the number of hip fractures was more than twice as much as previous reported by us. It was shown that the BUA in cases with hip fracture (41 and 39 dB/MHz) was significantly lower than the BUA value in the normal group (67 dB/MHz) without the dependence on the age. Variables like the weight and the height did not examined in this study because we have not find the BUA to correlate significantly with them [14]. Lower BUA values (than in our cases) have been reported, in postmenopausal women with hip fracture [12]. However the number of cases was smaller than ours and half of the measurements had been carried out at an interval between 2 weeks to 1 year from the time of fracture. Possibly the late measurement affected the BUA value of the calcaneus.
Stewart et al. [15] however, have reported higher BUA values (49.9± 23.9). These values may be significantly higher than ours. The standard deviation they observed was twice as large as that which we observed. This means a range of measured BUA values from 2 dB/MHz to 98 dB/MHz. We have never seen such extremes values in cases with fragility fracture. However, their reported difference in BUA between hip fractures and controls (18,08 dB/MHz) was quite similar with the respective value we found (16 dB/MHz) using age matched controls.
It has been reported that women who sustain trochanteric fractures have significantly reduced bone mineral density (BMD) compared with women sustain cervical fractures [16, 17]. Hence, significant differences in BUA among these fracture groups should be suspected. The fact that no significant difference in BUA was found between the women with trochanteric and those with cervical fracture, means that there are no different characteristics concerning the bone status - as expressed by BUA - of these groups.
The pathogenic process and the biomechanical characteristics between trochanteric and cervical hip fractures are quite different. A variety of factors are responsible not only for the causation, but also for the localization of the low energy hip fracture [18,19,20]. Among these factors, the neuromuscular condition and the mobility of the joints of the lower limbs, especially of the hip and knee, at the side of the fracture are of great importance. It is interesting to note that 35.5% (16/45) of the patients with trochanteric fracture were suffering from osteoarthrosis of the knee with marked limitation of the knee movements and weakness of the muscles of the thigh. On the other hand only 15% (5/33) of the patients with cervical fracture, presented with osteoarthrosis of the knee joint. In general the patients with cervical hip fractures were more mobile and independent in comparison with those with trochanteric fractures.
These observations emphasise the significance of the condition of the muscles and the joints of the lower limbs not only in the causation but also in the location of the hip fracture.
It is worth noting that over the 90% of the cases with hip fracture had BUA lower than 61 dB/MHz and the 80% lower than 51dB/MHz. Using age-matched control 87,5% of the hip fractures and 34% of the normals had BUA bellow the 61 dB/MHz and 75% and 12.5% respectively had BUA bellow the 51 dB/MHz. For screening purposes such BUA values associated with frequent falls, inactive lifestyles and early menopause, may arouse suspicion of risk for a hip fracture.

References
1. Agren M, Karellas A, Leahey D, Marks S, Baran D (1991) Ultrasound attenuation of the calcaneus: a sensitive and specific discriminator of osteopenia in postmenopausal women. Calcif Tissue Int, 48 (4):240-44
2. Damilakis J, Dretakis E, Gourtsoyiannis N (1992) Ultrasound attenuation of the calcaneus in female population: normative data. Calcif Tissue Int, 51 (3): 180-83
3. Langton CM, Palmer SB, Porter RW (1984) The measurement of broadband ultrasonic attenuation in cancellous bone. Eng Med, 13 (2): 89-91
4. McCloskey EV, Murray SA, Miller et al (1990) Broadband ultrasound attenuation of the os-calcis: relationship to bone mineral at other skeletal sites. Clin Sci, 78 (2): 227-33
5. Brandenburger GH (1993) Clinical determination of bone quality:Is ultrasound an answer? Calcif Tissue Int, 53(Suppl 1): S151-S156
6. Schott AM, Hans D, Sornay-Rendu E, Delmas PD, Meunier PJ (1993) Ultrasound measurements on os-calcis: Precision and age related changes in a normal female population. Osteoporosis Int, 3: 249-254
7. Yamazaki K, Kushida K, Ohmura A, Sano M, Inoue T (1994) Ultrasound bone densitometry of the os-calcis in Japanese women. Osteoporosis Int, 4: 220-225
8. Porter RW, Miller CG, Grainger D, Palmer SB (1990) Prediction of hip fracture in elderly women: a prospective study. Br Med J, 301: 638-641.
9. Dretakis E, Kontakis G, Steriopoulos C, Dretakis C (1994) Decreased broadband ultrasound attenuation of the calcaneus in women with fragility fracture. Acta Orthop Scand, 65 (3): 305-8
10. Jhamaria NL, Lal KB, Udawat M, et al (1983) The trabecular pattern of the calcaneum as an index of osteoporosis. J Bone Joint Surg (Br), 65 (2): 195-8.
11. Zhu H (1990) Survey and analysis of incidence and relevant factors of osteoporosis in the elderly (with a report of 2041 cases) Chung Hua I Hsueh Tsa Chih, 70 (5): 248-51
12. Baran D, Kelly A, Karellas A et al (1988) Ultrasound attenuation of the os calcis in women with osteoporosis and hip fractures. Calcif Tissue Int, 43 (3):138-42
13. Waud CE, Lew R, Baran DT (1992) The relation between ultrasound and densitometric measurements of bone mass at the calcaneus in women. Calcif Tissue Int, 51:415-418
14. Kontakis GM (1995) The significance of the measurement of the broadband ultrasound attenuation of the calcaneus for the prevention of the hip and Colles' fractures in postmenopausal women. Thesis, University of Crete, Greece
15. Stewart A, Reid DM, Porter RW (1994) Broadband Ultrasound attenuation and dual energy x-ray absorptiometry in patients with hip fractures: which technique discriminates fracture risk. Calcif Tissue Int, 54:466-69
16. Erikson Sav, Widhe TL (1988) Bone mass in women with hip fracture. Acta Orthop Scand, 59: 19-23
17. 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. Osteoporosis Int, 1: 81-86
18. Mautalen C, Vega E (1993) Different characteristics of cervical and trochanteric hip fractures. Osteoporosis Int, Suppl. 1: S102-106
19. Christodoulou NA , Dretakis EK (1984) Significance of muscular disturbances in the localization of fractures of the proximal femur. Clin Orthop, 187: 215-217
20. Dretakis EK, Christodoulou NA (1983) Significance of endogenic factors in the location of fractures of the proximal femur. Acta Orthop Scand, 54 (2): 198