Site-Specific Effects of Strength Training on Bone Structure and Geometry of Ultradistal Radius in Postmenopausal Women



Knowledge of the effects of exercise on bone mass in postmenopausal women is limited and controversial. Animal studies have shown that the response of bone to bending strain is an alteration of bone geometry. We studied 250 postmenopausal women, aged 52–72 years, willing to participate in a 6-month exercise program. The first 125 started the program immediately and the remaining 125 served as controls. The training program included exercises designed to maximize the stress on the wrist. One hundred and eighteen of the active group and 116 of the control group completed the study and were reassessed 6 months later. Bone mineral density (BMD) of the femoral neck, lumbar spine, ultradistal and proximal radius was measured by dual-energy X-ray absorptiometry (DXA) both before and at the end of the exercise program. The forearm was also evaluated by peripheral quantitative computed tomography, which measures the area, bone mineral content (BMC), and volumetric density for both the cortical and the trabecular component. The results showed that the DXA measurements at the femoral neck, lumbar spine, ultradistal and proximal radius were similar between the two groups. No significant difference was detected after the exercise program at the proximal radius. At the ultradistal radius, the cross-sectional area of cortical bone rose by 2.8 ± 15.0% (SD, p < 0.05), apparently for both periosteal apposition and corticalization of the trabecular tissue. The volumetric density of cortical bone rose by 2.2 ± 15.8% (p < 0.1), and that of trabecular bone decreased by 2.6 ± 10.7% (p < 0.01). The combined changes in both bone volume and density in the exercise group were associated with marked increase in cortical BMC (3.1 ± 10.7%, p < 0.01) and decrease in trabecular BMC (−3.4 ± 14.2%, p < 0.05), which were statistically different from those observed in the control group (p < 0.05). In conclusion, these results confirm that site-specific moderate physical exercises have very little effect on bone mass. However, it appears that some exercises may reshape the bone segment under stress by increasing both the cross-sectional area and the density of the cortical component. These structural changes are theoretically associated with increases in the bending strength.


The effectiveness of exercise in preventing and treating osteoporosis is mainly based on the assumption that physical activity modifies the risk of falling related to muscle weakness(1-3) and on its effect on the preservation of bone mass. The positive effect of exercise on the skeleton in postmenopausal women are mainly cross-sectional and therefore affected by several biases the most relevant being the duration of training activity and the differences in other lifestyle risk factors generally associated with the level of physical activity.

There are only a few prospective controlled trials that examined the effects of physical training in postmenopausal osteopenic women or the changes in bone density.(4) The results from these trials are not homogeneous. In some studies, no beneficial effects were observed(5-8) in others the increase in bone mineral density (BMD) were very limited.(9-13) The most striking and consistent increases in bone mass in postmenopausal women are observed after prolonged high intensity training(14-18) and in those on hormone replacement therapy (HRT).(11)

The discrepancy in the range of BMD changes are likely due the heterogeneity of the exercise program and of the method of evaluation. A prolonged mechanical stress may affect several aspects of bone modeling and remodeling, depending on the intensity and the direction of the forces applied to a well-defined skeletal segment. In the present study, the exercise program was designed to put the maximum strength on the ultradistal radius, which was then exhaustively studied by both dual-energy X-ray absorptiometry (DXA) and peripheral quantitative computer tomography (pQCT).



Two hundred and fifty postmenopausal women were recruited from a large group of subjects willing to participate in an exercise program organized by our local health service (ULSS 20, Regione Veneto). The criteria of inclusion were: >5 years postmenopausal, maximum age 72 years, and a body mass index (BMI, kg/m2) ranging from 20–29. None of them had been taking estrogen or had been engaged in a regular exercise program for at least 6 years, smoked more than five cigarettes per day, or took more than seven alcohol units per week. Twenty-five subjects were on angiotensin-converting enzyme (ACE) inhibitors associated with low-dose thiazide for moderate blood pressure elevation. The approximate dietary calcium intake was never lower than 700 mg/day (range 700–1200 mg). Preliminary screening tests included a medical history, physical examination, chest X-ray, and electrocardiogram. The latter was repeated in a few subjects during a standard effort test. None of the subjects referred previous femoral or spine fractures, but an X-ray evidence of that was obtained only in 37 subjects in whom the physical examination was suspicious. The screening process including bone mass measurements overall took 25 days. The selected women were ranked in alphabetical order and only the first 125 started the 6-month exercise training. The remaining 125 women were invited to continue their usual lifestyle for 6 months until their inclusion in the next exercise training course. This group of women served as controls.

The two groups of women were comparable for age, years past menopause, BMI, and baseline bone densitometric values (Tables 1 and 2).

Table Table 1.. Characteristics (Mean ± SD) of Active Subjects and Controls at Baseline
original image
Table Table 2.. Mean (± SD) Absolute Values of Cross-Sectional Areas (mm2), Volumetric bone Density (vBMD, mg/mm3) and Bone Mineral Mass (BMC, mg/cm) as Measured by pQCT at Different Bone Structure of the Forearm and Pertaining to Active Subjects and Controls at Baseline
original image

Subjects provided informed consent to participate in the study, which was approved by the local representatives of the National Health Service and of the consumers.


The training program consisted of warming-up exercises (walking associated with stretching of the arms and respiratory exercises) for a lag of time, which was increased from 15 minutes at the beginning of the program up to 30 minutes by its end. After the warming-up period, the women used to spend 70 minutes on exercises designed to maximize the stress on the wrist (press-up, flexion on the arms in a prone position, playing volleyball either sitting or standing). Overall, 10 minutes (3 minutes every 15) were dedicated to lifting a 500-g weight with the forearm in a partially supinated position, in order to maximize the stress on the muscle (brachioradialis) attached to the distal part of the radius. The number of weight liftings per minute rose progressively from 10 to 25 by the end of the exercise program. Subjects were invited to attend two sessions per week but were encouraged to repeat all exercises at home for at least 30 minutes/day.


The BMD of the femur (neck, trochanter, and Ward's triangle) and the lumbar spine (L2–L4) was measured by DXA using the Hologic 2000 (Waltham, MA, U.S.A.). At the dominant wrist (ultradistal region), the bone mineral content (BMC) was measured by using both a peripheral DXA instrument (Osteograph; NIM, Verona, Italy) and pQCT (XCT 960; Stratec, Unitrem, Rome). The coefficients of variation for DXA measurements were 0.8, 1.9, and 0.9% for spine, hip, and radius measurements, respectively.

The sites of pQCT evaluation were the ultradistal radius, following the procedure recommended by the manufacturer, and the proximal radius. The cross-sectional area of trabecular bone at the ultradistal radius and of the cortical bone at the proximal radius was measured by nonstandard software. The method has been recently described in detail elsewhere.(19) Briefly, the separation of the bone section into trabecular and cortical bone at the ultradistal site was achieved by assuming that more than 45% of the cross-sectional area is occupied by trabecular bone and, then, by a density threshold. This expresses a linear attenuation coefficient arbitrarily ranging from 0 to 100. Any pixel with an attenuation coefficient above 50 is not considered trabecular bone, which is here defined, for the sake of brevity, as cortical bone tissue. At the proximal radius, an attenuation coefficient threshold of 60 was used to define cortical bone tissue. All elaboration processes are automatically computed. The coefficient of variation for ultradistal and proximal measurements is 2% and 1.8%, respectively.(19)

Bone turnover was evaluated by the measurement in spot urine samples (from 10 to 12 a.m.) of the free pyridinium cross-links (Pyrilinks, Metra, Italy) and of the cross-linked N-telopeptide (NTx) of type I collagen (Osteomark, Bouty, Italy), both assessed in nanomoles per millimole of creatinine.

Statistical analysis

The baseline characteristics of the active and control groups were compared by Student's t-test and square-qui as appropriate. Two-tailed analysis of variance was used to compare the differences among the groups in the changes in bone measurement data, and the paired t-test was used to analyze the longitudinal changes within the groups. After the exclusion of the subjects who interrupted the training course within the first month, the changes were adjusted for the mean frequency attendance to the exercise program (Statgraphics plus; Statgraph, Rockville, MD, U.S.A.).


Adherence to the exercise program was calculated as the number of training sessions attended until the second bone mass measurement. Seven of the active subjects interrupted the training course within a month and were then excluded from analysis. The overall frequency attendance ranged from 28 to 54 (40 ± 6 SD), depending also on the timing of the bone measurement. Most women confirmed their adherence to the home exercise program, which, however, could not be quantitated with sufficient precision. Since this could not be quantitated, we did not try any correction of the final results with the compliance to the training course. Nine of the control women did not come back for the second clinical assessment. All women adhering to the study protocol did not change their lifestyle (diet, smoking, exercise outside the training course, drugs) or complain of major health problems during the 6 months of observation.

The percentage changes shown in are adjusted for the mean attendance frequency to the exercise program. This adjustment slightly decreased the variance but had only marginal effects on the statistical significance of the changes. None of the DXA measurements significantly changed at the end of the 6 months of observation in either group (Table 3).

Table Table 3.. Percent Changes (Mean ± SD) in BMD as Measured by DXA at Different Skeletal Sites of Active and Control Group After 6 Months of Observation
original image

The pQCT measurements at the proximal radius included the cross-sectional area of cortical bone, the BMC, and the volumetric BMD (vBMD). None of these parameters showed a relevant changing trend at the 6th month of the study (Table 4).

Table Table 4.. Percent Changes (Mean ± SD) in Volumetric Bone Density (vBMD), Bone Mineral Content (BMC) and Cross-Sectional Area of Cortical Bone, as Measured by pQCT at the Proximal Radius, of Active and Control Group after 6 Months of Observation
original image

At the ultradistal radius, the total BMC did not change significantly (p < 0.2), but the structure of this bone segment was significantly reshaped in a few aspects (Table 5). The cross-sectional area occupied by cortical bone rose by 2.8% (p < 0.05) and this change was at the limit of statistical significance when compared with control women (p = 0.05). Although, neither the total nor the trabecular cross-sectional area changed significantly, the results seem to indicate that the expansion of the cortical area was due to periosteal apposition and occurred also at the expense of trabecular area. The overall density rose significantly (p < 0.05) despite a significant decrease in the density of the trabecular component. The enlargement of the cortical area and of its density led to a very significant (p < 0.01) increase in the cortical bone mass, the change of which also differed also significantly from those of the control group. The changes of the cortical BMC were negatively related (p < 0.01) with those observed for trabecular BMC (data not shown), which also decreased significantly in comparison with the control women (Table 5).

Table Table 5.. Percent Changes (Mean ± SD) in Volumetric Bone Density (vBMD), Bone Mineral Content (BMC) and Cross-Sectional Area of Different Bone Structures, as Measured by pQCT at the Ultradistal Radius, of Active and Control Group after 6 Months of Observation
original image

The urinary excretion of free pyridinoline and NTx did not change significantly in either group, but the decline of NTx in the active group was at the limit of significance (−4.8% ± 28.4 SD, p = 0.07; data not shown).


A great deal of experimental data have shown the beneficial effects of exercise on the skeleton, but clinically relevant changes on bone mass can be apparently obtained only with exercise programs which introduce stress to the skeleton.(18) Thus, the potential of less strenuous exercises as an agent for the prevention or treatment of osteoporosis remains unclear. It is well accepted that bone mass and bone density are not the sole determinants of ultimate bone strength, and a great contribution is also provided by tissue and geometric properties.(20,21) The latter are generally not considered in the routine clinical assessment of osteoporosis for technical, cultural, and practical problems. In the clinical ground nowadays, the only technology that can provide in vivo geometric insights is computed tomography, yet the correlation between a given shape of cortical bone and its resistance to a clinically relevant bending strain is poorly defined. An additional problem in evaluating the effect of physical exercise on bone geometry is the obvious need for skeletal site specificity.(22)

For the first time in a clinical study, we have been able to look at the structural and geometric changes of a bone segment which was, by design, the most specific target of the training course. The inability of this course to produce increases in BMD at the spine and femur, despite the large size of the study population, is somewhat disappointing. It appears that even in women scarcely active, a reasonable and age-tailored physical exercise program is of little help in preventing bone loss. Anyway, our results, taking into account the duration and the intensity of the program load on the spine and femur, should not be considered discordant from those reported by others.(4-18)

More surprising is the lack of effects on bone mass and density of the ultradistal radius, measured by both traditional DXA and pQCT. From this point of view, our study cannot be easily compared with others. Even though the exercise program was designed to put the maximum strength on the most distal part of the radius, the amplitude and the frequency of the strain are possibly lower than those obtained by high-intensity strength training on weight-bearing bone segments. Ayalon et al.(14) have shown that a 5-month program of dynamic bone-loading exercises for the distal forearm, possibly higher than those adopted in our study, may increase the volumetric bone density of distal radius by 3.8%, as measured by the Compton scattering technique. However, BMC measured by single-photon absorptiometry (SPA) did not demonstrate any significant trend over the whole period of the study.(23) It should also be noticed that the reported bone mass changes with moderate physical exercises over at least 1 year are not impressive.(4) It is therefore conceivable that we were not able to detect any effect on the BMC of the forearm only because the duration and the intensity of the program were insufficient in relation to the precision of the instrument.

Our results indicate that though the overall mass of the bone segment did not change, a remarkable reshape of its structure and geometry took place. The compact bone area rose at apparent expense of the trabecular area. The expansion of cortical area was associated with an increase in its density (and vice versa for the trabecular component), and this excludes the fact that the results are artefactual. Indeed, any consistent bias in the separation of cortical from trabecular bone would have led to opposite changes in area and density. The changes in cortical bone density are likely due to increased compactivity of the tissue at the endosteum with mixed trabecular and cortical structure, but it might also be an artefact as a consequence of the partial volume effects, which become less important the thicker the cortical areas.(19)

The enlargement of the cortical area of the ultradistal radius seems to be due mostly to corticalization of trabecular bone at the endosteal surface, but also to some periosteal apposition (+0.4%; Table 5). These effects of physical exercise on cortical bone have never been studied before in humans, but they are consistent with a number of animal in vivo studies,(24-26) showing that the response to bending strain (rather than compression) is an enlargement of cortical area.

The cortical changes we have observed predict a substantial increase in the bending resistance, which depends upon the material property and the so-called cross-sectional moment of inertia,(22) which is a function of the fourth power of the cortical ring (i.e., the outer diameter [D] to the fourth power minus the inner diameter [d] to the fourth power divided by D). We can reasonably assume that the increased density of cortical bone was associated with an improvement of the material property, whereas the expansion of the cortical area induces an exponential increase in the calculated moment of inertia. The significant diminution in trabecular bone density is difficult to explain. The trabecular area measured was only marginally decreased after 6 months of exercise, and in that site the loss in bone mass was 0.92 mg/cm. It is likely that the trabecular bone loss occurred only at the ultradistal forearm, since no significant changes in BMC were observed by DXA at the axial skeleton, where the trabecular component is also relatively large. Yet we cannot rule out that that the same reshaping phenomenon has occurred also in other skeletal sites. If the trabecular bone loss was a localized phenomenon, we can envisage a communication system between different structures of the same skeletal segment.

In conclusion, our study suggests that site-specific physical exercises are able to reshape both the geometry and the structure of specific bone segments. Whether these changes will translate into increased bending strength of the interested bone is an attractive hypothesis, which requires additional experimentation.