Osteology Unit, Derer’s University Hospital and Policlinic Bratislava, Slovakia, head Ass. Prof. Jaroslava Wendlová, MD, PhD.
Vnitř Lék 2010; 56(7): 764-770
80th Birthday - Jaroslava Blahoše, MD, DrSc.
Patients and methods: We analysed the data in the sample (n = 3,215) of East Slovak women with a primary or secondary osteopenia, osteoporosis and with risk factors for osteoporosis, aged 20–89 years, median 59 years, 95% C. I. (59.31; 60.07) obtained from dual energy X‑ray absorptiometry device (Prodigy-Primo, GE, USA). Measured variables: 1. left proximal femur: T-score total hip, FSI (femur strength index), 2. lumbar vertebrae L1–L4: BMD (bone mineral density).
Objectives: 1. To estimate and to compare an expected frequency of pathological FSI < 1 and T-score total hip ≤ –2.5 SD values in the East Slovak female population. 2. To estimate expected frequency of women with: FSI < 1 and T-score total hip ≤ –2.5 SD (Group A), FSI < 1 and T-score total hip from interval from –1,0 till –2.5 SD (Group B), FSI < 1 and T-score total hip > –1.0 SD (Group C) in the East Slovak female population. 3. To determine, if FSI variable value is a significant predictor of BMD variable values in lumbar vertebrae.
Results: 1. In the East Slovak female population we can expect 14.54% of women with FSI values < 1 and 6.25% of women with osteoporosis in the total hip area according to T-score. 2. For the group A we can expect the mean value (μ) from interval (1.41; 2.36) %, for the group B from interval (4.50; 6.03) % and for the group C from interval (6.76; 8.55) %. 3. Between FSI and BMD L1–L4 variable values there is not a statistically significant dependence, because FSI variable is quantitative and qualitative different variable from BMD variable.
Conclusion: The measurement of FSI variable values may discover a higher percentage of women with a probability of femoral neck fracture by fall than the measurement of BMD variable value in the total hip area. Patient with osteopenia or normal BMD measured in the total hip area may sustain a femoral neck fracture by fall, when she has pathological value of FSI, i.e. she has adverse values of geometric variables of proximal femur (biomechanically unfavourable proximal femur configuration). FSI variable value is not a significant predictor of BMD variable values in lumbar vertebrae L1–L4.
Key words: osteoporosis – femur strength index (FSI) – bone mineral density (BMD) – femoral neck fracture – dual energy X‑ray absorptiometry (DXA) – lumbar vertebrae
With the development of clinical osteology, spheres of interest, objectives and
directions of epidemiological studies have been changing. Rising costs of
complex treatment of fractures and social care of patients with permanent
consequences of fractures (reduced work skills, reduced self-sufficiency,
disability) brought to attention the precise assessment of bone quality,
fracture risk and the prevention strategy of osteoporotic fractures [1,2]. The
treatment costs of fractures caused by falls represent still an economic burden
for the health care system [3–7]. Therefore, the search continues to find more
precise variables in densitometric methods (DXA and QCT – quantitative
computed tomography) to determine patients with a fracture risk by fall.
clinical study we were interested in comparison of two densitomeric variables
measured by DXA and their possibility to discover the female patients at high
risk of proximal femoral fracture by fall.
Objectives of the study
1. To estimate an expected frequency of the occurrence of
pathological FSI < 1 variable values and pathological BMD (T-score
total hip ≤ –2.5 SD) variable values in the East Slovak female
population and to compare them.
2. In the East Slovak female population to estimate
expected 95% C.I. for percentage of women with variable values:
FSI < 1 and T-score total
hip ≤ –2.5 SD (Group A),
FSI < 1 and T-score total
hip from interval from –1.0 till –2.5 SD (Group B),
FSI < 1 and T-score total
hip > –1.0 SD (Group C)
and in the sample to estimate the percentage
of women belonging into groups A, B or C.
3. To determine, if FSI variable value is
a significant predictor of BMD values in lumbar vertebrae L1–L4.
Patients and methods
Characteristics of the sample
Using a DXA bone densitometer
Primo, GE, USA) we analysed the data in the sample of East Slovak women
(n = 3,215) aged 20–89 years, median 59 years, 95% C. I.
without case – history of femoral neck fracture,
with risk factors for the development of osteoporosis,
with a primary or secondary osteopenia,
with a primary or secondary osteoporosis.
were examined with the same bone densitometer DXA. The BMD was determined in
the standard region of interest (ROI) – total hip left. BMD values were
given in absolute numbers in g of Ca-hydroxyapathite crystals for cm2 (g/cm2),
as well as in relative numbers as T-score (the number of standard deviations
from the reference group of young healthy women). Osteoporosis or osteopenia
were diagnosed in accordance with the WHO criteria (tab. 1).
the measurement quality (QA), only two operators alternated in measuring with
the DXA device and all women were measured with the same device. The following
variables were measured:
Proximal femur left: BMD total hip (T-score total hip),
Lumbar vertebrae L1–L4:
Characteristics of the population
All women of the East Slovakia with osteopenia, osteoporosis and with
risk factors for osteoporosis (hereafter East Slovak female population).
Definitions of measured variables and variables
describing in the discussion 
FSI (femur strength index) – is a biomechanical variable
determining whether the bone strength in the femoral neck area endures the load
of compressive force impact by fall (normal value: FSI ≥ 1,
pathological value: FSI < 1).
defined as a ratio of estimated elastic limit in compression of the
femoral neck (δE) to the expected compressive stress of a fall on
the greater trochanter adjusted for the patient’s age, height and weight (δC).
α angle (α) is an angle formed by the femoral shaft axis and the perpendicular. The
a angle can acquire both positive and
negative values in the population, depending whether the femur is in
a valgus or varus position.
θ angle (θ) is an angle formed by the femoral neck axis and the
femoral shaft axis.
(hip axis length) –
is a distance (in mm)
from the beginning point of the greater trochanter protuberance to the pelvis
inner rim, measured in the femoral neck axis.
(cross sectional moment of inertia) – is defined as the sum of multiplications of
elementary areas and the squares of their distances from neutral axis, denoted
as IY (given in cm4). The larger the cross sectional
area, the higher the number of small elementary areas (A) and so is the bigger
the second root of the distance (z2) of these elementary areas from
the neutral axis. The enlargement of the cross sectional area is accompanied by
the enlargement of the cross sectional moment of inertia.
Statistical analysis [9–13]
To analyse the data of the sample,
statistical methods were applied using statistical programme systems Statgraphics
We tested the character of the distribution of FSI, T-score
total hip and BMD L1–L4 variable values in the
Using the Goodness-of-Fit Test χ2 and coming from the character of empirical distribution of the
frequencies (probability) of FSI variable values in the sample, we estimated
the expected frequencies of FSI variable values in the East Slovak female
population at the significance level a = 0.01 We
tested the character of the distribution of variable values in the sample and
using the Goodness-of-Fit Test χ2 and coming from the character of
empirical distribution of the frequencies (probability) of T-score total hip variable values in the
sample, we estimated the expected frequencies of T-score total hip variable
values in the East Slovak female population at the significance level a = 0.05.
In the sample we calculated percentage of women, which
belong into group A, B or C.
In the East Slovak female population, we calculated the
95% C.I. (confidence interval) for expected percentage frequencies of women,
which belong into group A, B or C.
Linear regression analysis: To estimate the Pearson’s
correlation coefficients between FSI and BMD L1–L4 variable
values. The coefficients verify the linear association and measure the
intensity of association between variables FSI and BMD L1–L4.
To illustrate the results we used the following statistical
graphs: histograms, circle diagram.
The median (x) for age is 59 years, 95% C. I. (59.31; 60.07).
The table 2 brings an empirical distribution of FSI
variable values frequencies in the sample and the expected distribution of the
FSI values frequencies in the East Slovak female population. In the East Slovak
female population we can expect 0.18% of women with the FSI
values < 0.5 and 14.36% of women with the FSI values from the
interval (0.5; 1.0). It means that we can expect
14.36% + 0.18% = 14.54% of women in the East Slovak female
population with FSI values < 1. At the significance level α = 0,01 it can be assumed that the probability distribution for
FSI variable values in the East Slovak female population is loglogistic with
the parameters derived on the basis of values found in the sample. One patient
with faulty measurement FSI = 0 was excluded from sample
n = 3,214 (fig. 1).
T-score Total Hip Variable
In the East Slovak female population according to values measured in the
total hip area we can expect 6.25% of osteoporotic women with the T-score
values ≤ 2.5 SD and 34.94% of osteopenic women with T-score
values from the interval from –1.0 till –2.5 SD. At the significance level α = 0.05 it can be assumed that the
probability distribution for T-score total hip variable values in the East
Slovak female population is loglogistic with the parameters derived on the
basis of values found in the sample (tab. 3, fig. 2).
expected frequency of the incidence of pathological FSI values of the East
Slovak female population is 2.33 times higher as the expected frequency of
the incidence of pathological T-score for osteoporosis measured in the total
hip area (tab. 4).
FSI and T-score Total Hip Variables
In the sample, from total number of women
with FSI < 1 (n = 471), there were 51.91% women
with FSI < 1 and with T-score total hip > –1.0 SD, 35.59%
women with FSI < 1 and with T-score total hip from interval
from –1.0 till –2.5 SD and 12.50% women with FSI < 1 and
with T-score total hip ≤ –2.5 SD (fig. 3, tab. 5b).
sample from total number of women (n = 3 215) there were 7.62%
women with FSI < 1 and with T-score total
hip > –1.0 SD, 5.23% women with FSI < 1 and
with T-score total hip from interval from –1.0 till –2.5 SD and 1.84%
women with FSI < 1 and with T-score total hip ≤ –2.5 SD
In the East
Slovak female population it can be expected with the probability of
0.95 (95%), that the mean value (μ) of percentage (%) will be from:
(1.41; 2.36) for women with FSI < 1 and simultaneously with
T-score total hip ≤ –2.5 SD,
(4.5; 6.03) for women with FSI < 1 and simultaneously with T-score
total hip from interval from –1.0 till –2.5 SD,
(6.76; 8.55) for women with FSI < 1 and simultaneously with
T-score total hip > –1,0 SD (tab. 5a, 5b).
FSI and BMD L1–L4 Variables
Linear regression analysis: At the significance level α = 0.05 there is a very low direct dependence between FSI
and BMD L1–L4, the Pearson’s correlation coefficients are
near to the zero. Variable FSI explains the variability of BMD L1 variable
values only in 0.250%, of BMD L2 only in 0.230%, of BMD L3 only
in 0.380% and of BMD L4 in less than 0.0% (tab. 6).
At present, FSI is one of a few
available variables, corresponding to biomechanical criteria of loading the
bone by fall, whose values can be determined by DXA in a routine
ambulatory practice. FSI integrates in itself three important bone
bone geometry: HAL, angle a (a), angle θ (θ), CSMI in the minimum cross sectional area of femoral
neck (see definitions),
elasticity and strength.
FSI variable value is calculated from the two‑dimensional bone measurement by DXA, which in comparison
with three-dimensional measurement has some inaccuracies, it is a value
determining the bone quality much more precisely than BMD. Clinical importance
of the FSI variable value lies in the fact that it enables to determine an
individual forecast for femoral neck fractures by fall. The forecast of femoral
neck fractures by fall on the basis of BMD is markedly limited, as BMD
characterises only one physical value-density, which determine only a part
of bone quality.
some clinical retrospective or prospective studies [14–20] agree that the
increase of the values of geometric variables of proximal femur
width of femoral neck) over the mean value in the population is a risk
factor for femoral neck fractures.
Faulkner et al  compared BMD, HAL, and CSMI in femoral neck area and
FSI obtained from DXA measurements in a group of women with and without
hip fracture. Femoral neck BMD and FSI were significantly lower and HAL
significantly higher in fractured group compared with controls. Mean CSMI was
not significantly different between fracture patients and controls. After
adjustment of these variables in regression model, the authors conclude that
BMD, HAL and FSI are significant independent predictors of hip fracture.
Crabtree et al  concluded in their work, that the geometric variation of
proximal femur may contribute to the large variations in hip fracture risk
clinical study we were interested in percentage comparison of pathological
FSI < 1 and T-score total hip ≤ –2,5 SD
variable values in the same sample of East Slovak women and to estimate the
expected frequencies of this variable values in the East Slovak female
population. The expected percentage of FSI pathological variable values
occurrence (14.54%) is 2.33 times higher than the expected percentage of
osteoporosis occurrence in the total hip area (T-score ≤ –2.5 SD)
in the East Slovak female population.
evident from such a simple comparison that BMD and FSI variables are two
different characteristics of bone quality. FSI, which characterizes the bone
quality by means of several integrated biomechanical variables, including in
itself the BMD variable, discovers a higher percentage of women at risk
for femoral neck fracture by fall than BMD. The results show that even
a patient with osteopenia or normal BMD measured in the total hip area may
sustain a femoral neck fracture by fall, in case she has pathological
value of FSI variable, i.e. she has adverse values of geometric variables of
proximal femur HAL, angle a, angle θ
(so-called biomechanically unfavourable proximal femur configuration).
In our previous published studies [23,24], we have estimated from
predictive logistic regression model, created from East Slovak female data,
that the rise of HAL, angle a, angle θ variable values by one unit statistically
significant raises the odds for femoral neck fracture by fall. It verifies that
the variation of HAL, angle a, angle θ variable values significant influences the risk of
femoral neck fracture by fall.
measurement of FSI value in clinical practice enables to improve the strategy
of treatment in the prevention of femoral neck fractures.
value < 1 could therefore be used as a criterion for the
initiation of preventive therapeutic interventions in order to avoid femoral
neck fractures. These interventions may include:
wearing a hip joint protector,
to remove the muscular dysbalance in the mm. coxae area ,
techniques how to fall correctly.
variable is quantitative and qualitative different variable from BMD variable,
what shows a little portion of variability of FSI variable values on the
variability of BMD L1–L4 variable values according
to linear regression analysis in our study. Prediction of BMD variable values
by the help of FSI variable values is unreliable and therefore impossible in
measurement and determination of bone quality have been constantly improving.
As demonstrated by the latest clinical studies, the future of bone densitometry
lies in a three-dimensional measurement of FSI an other new progressive
biomechanical variables and also in high resolution QCT [26–30].
New types of DXA and QCT densitometers
provide programs for clinical and ambulatory practice enabling to measure different biomechanical and
geometric variables of proximal femur. Regrettably, these programs are utilized
minimally in practice and osteologists are still rigidly adhering to BMD
values. Offered programs overrun the “guidelines” of international and national
osteological societies, which still have not included the criteria for the
evaluation of new biomechanical variables.
The measurement of FSI variable values may discover
a higher percentage of women in the population with a risk of femoral
neck fracture by fall than the simple measurement of BMD variable values in the
total hip area.
Patient with osteopenia or normal BMD measured in the
total hip area may sustain a femoral neck fracture by fall, when she has
pathological FSI value i.e. she has adverse values of geometric variables of
proximal femur (so-called biomechanically unfavourable proximal femur
Between FSI and BMD L1–L4 variable
values there is not a statistically significant direct dependence, because
FSI variable is quantitative and qualitative different variable from BMD
variable. FSI variable value cannot be a predictor of the BMD variable
values in lumbar vertebrae L1–L4.
The study investigated East Slovak women. Although it
can be supposed that similar results will be found in the whole Slovak female
population and in other female populations at least in central Europe, further
studies will have to verify this.
Declaration – conflict of
The author declares, that she has no competing interests
(financial or non financial).
do redakce: 12. 5. 2010
Prof. Jaroslava Wendlová, MD, PhD.
1. Pietschmann P, Kerschan-Schindl K. Knochenqualität – wissenschaftliche Aspekte versus praktische Relevanz. J Miner Stoffwechs 2004; 11: 16–18.
2. Boonen S, Singer AJ. Osteoporosis management: impact of fracture type on cost and quality of life in patient at risk for fracture I. Curr Med Res Opin 2008; 24: 1781–1788.
3. Viktoria Stein K, Dorner TH, Lawrence K et al. Economic concepts for measuring the costs of illness of osteoporosis: An international comparison. Wien Med Wochenschr 2009; 159: 253–261.
4. Finnern HW, Sykes DP. The hospital cost of vertebral fractures in the EU: estimates using national data sets. Osteoporos Int 2003; 14: 429–436.
5. Lindsay R, Burge RT, Strauss DM. One year outcomes and costs following a vertebral fracture. Osteoporos Int 2005; 16: 78–85.
6. Jahelka B, Dorner T, Terkula R et al. Health‑related quality of life in patients with osteopenia or osteoporosis with and without fractures in a geriatric rehabilitation department. Wien Med Wochenschr 2009; 159: 235–240.
7. Rabenda V, Manette C, Lemmens R et al. The direct and indirect costs of the chronic management of osteoporosis: a prospective follow‑up of 3440 active subjects. Osteoporos Int 2006; 17: 1346–1352.
8. Nakamura T, Turner CH, Yoshikawa T et al. Do variations in hip geometry explain differences in hip fracture risk between Japanese and white Americans? J Bone Miner Res 1994; 9: 1071–1076.
9. Rublíková E, Labudová V, Sandtnerová S. Analysis of categorical data. 1st ed. Bratislava: Publishing house Economic University 2009: 41–141.
10. Pacáková V. Statistical methods for economists. 2nd ed. Bratislava: Publishing House: Iura Edition 2009: 174–177.
11. Varga S. Another View on the Fuzzy Regression. Forum Statisticum Slovacum 2009; 3: 1–7.
12. Pacáková V. Aplikovaná poistná štatistika. 1 vyd. Bratislava: Publishing House: IURA Edition 2004: 67–86.
13. Varga Š. Fuzzy predictions in regression models. J Appl Mathem Open Access 2010; 3: 245–251.
14. Kukla C, Gaebler C, Pichl RW et al. Predictive geometric factors in a standardized model of femoral neck fracture. Experimental study of cadaveric human femurs. Injury 2002; 33: 427–433.
15. Gregory JS, Testi D, Stewart A et al. A method for assessment of the shape of the proximal femur and its relationship to osteoporotic hip fracture. Osteoporos Int 2004; 15: 5–11.
16. Alonso CG, Curiel MD, Carranza FH et al. Femoral bone mineral density, neck shaft angle and mean femoral neck with as predictors of hip fracture in men and women. Multicenter Project for Research in Osteoporosis. Osteoporos Int 2000; 11: 714–720.
17. El Kaissi S, Pasco JA, Henry MJ et al. Femoral neck geometry and hip fracture risk: the Geelong osteoporosis study. Osteoporos Int 2005; 16: 1299–1303.
18. Watts NB. Fundamentals and pitfalls of bone densitometry using dual-energy X‑ray absorptiometry (DXA). Osteoporos Int 2004; 15: 847–854.
19. Gnudi S, Malavolta N, Testi D et al. Differences in proximal femur geometry distinguish vertebral from femoral neck fractures in osteoporotic women. Br J Radiol 2004; 77: 219–223.
20. Gnudi S, Ripamonti C, Lisi L et al. Proximal femur geometry to detect and distinguish femoral neck fractures from trochanteric fractures in postmenopausal women. Osteoporos Int 2002; 13: 69–73.
21. Faulkner KG, Wacker WK, Barden HS et al. Femur strength index predicts hip fracture independent of bone density and hip axis length. Osteoporos Int 2006; 17: 593–599.
22. Crabtree N, Lunt M, Holt G et al. Hip geometry, bone mineral distribution, and bone strength in European men and women: The EPOS study. Bone 2000; 27: 151–159.
23. Wendlova J. Logistic regression in estimate of femoral neck fracture by fall. Open Access Emergency Medicine 2010; 2: 29–36.
24. Wendlova J. Expected frequency of femoral neck fractures by fall in the osteoporotic and osteopenic East Slovak female population. (Epidemiological Study). Wien Med Wochenschr 2010; 159. V tisku.
25. Wendlová J. Why is so important to balance the muscular dysbalance in mm. coxae area in osteoporotic patients? Bratisl lek listy 2008; 109: 502–507.
26. Cheng X, Li J, Lu Y, et al. Proximal femoral density and geometry measurements by quantitative computed tomography: association with hip fracture. Bone 2007; 40: 169–174.
27. Manske SL, Liu-Ambrose T, De Bakker PM et al. Femoral neck cortical geometry measured with magnetic resonance imaging is associated with proximal femur strength. Osteoporos Int 2006; 17: 1539–1545.
28. Bousson V, Le Bras A, Roqueplan F et al. Volumetric quantitative computed tomography of the proximal femur: relationships linking geometric and densitometric variables to bone strength. Role for compact bone. Osteoporos Int 2006; 17: 855–864.
29. Engelke K, Adams JE, Armbrecht G et al. Clinical Use of Quantitative Computed Tomography and Peripheral Quantitative Computed Tomography in the Management of Osteoporosis in adults. The 2007 ICSD Official Position. J Clin Densit 2008; 11: 123–162.
30. Sipos W, Pietschmann P, Rauner M et al. Pathophysiology of osteoporosis. Wien Med Wochenschr 2009; 159: 230–234.