- Email: [email protected]
The effect of swimming activity on bone architecture in growing rats
J. Bmmrchuntcs
Vol 22. No. 819. pp. 845-851,
cO21-9290189 %3.oo+.W PergamonPress plc
1989.
Prmted1"GreatBr~lain
THE EFFECT OF SWIMMING ACTIVITY ON BONE ARCHITECTURE IN GROWING RATS A.
SIMKIN*$,
I. LEICHTER~~,A.
SWISSA$
and S.
SAMUELOFF*
*Cosell Center for Physical Education, Hebrew University, Givat Ram, Jerusalem; tJerusalem College of Technology, 21 Havaad Haleumi Street, P.O. Box 16031,91160 Jerusalem; $Department of Orthopaedic Surgery and §Jerusalem Osteoporosis Center, Hadassah University Hospital, Jerusalem. Israel Abstract - The effect of non-habitual physical activity on bone architecture in the rat humeral shaft was examined. Two groups of rats were trained to swim for 1 h a day, for 20 weeks, at Iwo training levels. The control group consisted of sedentary rats. Parameters of cross-sectional bone morphology (cross-section areas, principal area moments of inertia and their ratio) were used to evaluate the response of bone architecture to mechanical loading. The strength of bone was assessed by measuring the ultimate compressive force and stress. The cortical cross-section area and principal moments of inertia were found to be significantly higher in the swimming groups than in the controls. Examination of the ratio between the major and minor moments of inertia revealed a pronounced change in the shape of the bone cross-section which became more rounded following swimming training. The ultimate compressive force was significantly higher in the swimming rats while the changes in ultimate stress were not significant. Our results indicate a gain of bone strength due to increased periosteal apposition and modified bone tissue distribution. The marked changes in bone morphology are attributed to the different nature of the forces and moments exerted on the humerus during swimming compared to those prevailing during normal locomotion.
INTRODUCTION
Living bone undergoes an adaptation process in which the structure of the bone is continually modified in response to the stresses imposed on it by mechanical loading. Since this phenomenon was investigated at the end of the 19th century (Roux, 1881; Wolff, 1892), many experimental studies have verified the functional adaptation of bone. In a series of experiments in the last decade, attempts were made to isolate the mechanical factors which affect bone growth and to quantify their relationship to the adaptation process (Lanyon et al., 1982a; Lanyon et al., 1982b; O’Connor et al., 1982; Rubin and Lanyon, 1984). These studies have shown that an increase in bone mass is engendered by intermittent dynamic loading. The amount and architecture of the additional bone tissue depends on the strain level, strain rate, daily number of loading cycles and the variation of strain distribution within the bone. The last factor expresses the dissimilarity between the strain distribution under experimental conditions and that undergone by the bone as a result of habitual activity. The importance of the modification in strain distribution was determined in two sets of in uiuo experiments. In the first (Lanyon et al., 1982a), the radius of a sheep was forced to carry the entire load on the forearm during locomotion by removing a portion of the ulna. In the second (Rubin and Lanyon, 1984), a functionally isolated avian ulna was artificially loaded. In these experiments, the unnatural load distribution was obtained as a result of an external intervention. Received in final form 2 September 1988.
In the search for an activity which would provide a non-habitual but physiological loading regime in animals, we observed that the upper limbs of swimming rats undergo movements which differ from those used in habitual, terrestrial locomotion. In the present study, therefore, the effect of swimming on bone growth was examined in rats which were trained to swim daily for lh. The adaptation of bone architecture in response to mechanical loading was studied by using parameters of cross-sectional bone morphology (Lovejoy et al., 1976; Woo et al., 1981) which enable the investigation of bone tiss’ue redistribution within its cross-section. MATERIALSAND METHODS
Forty female Sabra albino rats were used in this study. At the beginning of the study, they were five weeks old and their mean weight was 150 g, with a standard deviation of 15 g. They were randomly allocated into three groups of different training levels. The first group of 20 rats was trained to swim for 1 h a day, five days a week, for 20 weeks. This training level was considered light swimming training (LST). A second group of 11 animals underwent the same training regime, but was required to swim with additional lead weights tied to the roots of their tails. Each weight was equal to 1% of the rat’s weight in air. This training level was considered moderate training (MST). A third group of nine animals was a non-swimming control group (CONTROLS). The swimming was performed in a cylindrical water container 50cm in diameter and 70cm deep. The temperature of the water was kept at 35°C + 1°C. While not swimming, the animals were kept in cages at a temperature of
845
846
A.
SIMKIN
22°C + 1 “C on a 12 h light+lark cycle, with water and food ad lib&urn. At the end of the training period, the animals were sacrificed. The forelimbs were removed and the left humeri were disarticulated and all soft tissues were carefully removed. Two specimens for a compression test were prepared from the shaft of each humerus: a proximal one, extending 5 mm distally from the borderline between the shaft and the humeral head and a distal one, extending 5 mm proximally from the proximal edge of the olecranon fossa. Each specimen was prepared using a miniature lathe with a circular saw, 80 x 0.6mm, under continuous irrigation with normal saline. Both faces of each specimen were polished manually with No. 320 silicon carbide paper in order to obtain two parallel surfaces. A photograph of each surface was taken next to a millimetric scale. A magnified print (x 25) of each photograph was then prepared. Each specimen was then compressed in a materialtesting machine (Instron, Model 1125) at a constant deformation rate of 2 mm min-’ (strain rate of approximately 0.007 s- 1). The force-deformation curve for each specimen was observed during compression and after a maximum of the curve was reached, the compression was stopped. The value of the force at the peak of the curve was taken as the ultimate compressive force (UCF). The ultimate compressive stress (UCS) was calculated as the ratio between UCF and the smaller of the two cortical cross-section areas at the two faces of the specimen. The cortical cross-section area (CA) was measured from the magnified prints using the computer integrative-graphics program of Nagurka and Hayes (1980). It was computed as the difference between the total subperiosteal cross-section area (TA) and medullary area (MA) (Fig. 1). Using the same computer program, the minor and major principal area mo-
et al.
ments of inertia were calculated in relation to the two principal axes intersecting at the centroid of the cortical cross-section. The area moment of inertia provides a measure of bone strength and rigidity under bending. The cross-section areas and the area moments of inertia were computed for both faces of each specimen. The value of each parameter was taken as the average of the results obtained for the two faces of the specimen. The ratio between the major and minor area moments of inertia was calculated in order to assess the differences in the shape of the cortical crosssection between the swimming rats and the controls. No significant difference between the ratios would indicate that changes in CA following swimming were evenly distributed around the centroid of the crosssection. An increase or a decrease in the ratio would indicate that the cross-section of the bones in the swimming rats became flatter or rounder, respectively. RESULTS
Figure 2 shows the mean values of the cross-section areas (TA, MA, CA) of the specimens from the proximal (Fig. 2A) and distal (Fig. fZB)humeral shafts for each group. Both TA and MA values in the LST and MST groups were significantly higher than those in the controls, in both the proximal and distal sites. The CA (CA= TA- MA), which represents the net crosssection area of cortical bone, was also higher in the
Controls Light swimming Moderate swimming
’ pa.05 ” p(O.005
Principal Y axis
I
TA
Fig. 1. A cross-section of the rat humeral shaft, 5 mm proximal to the edge of the olecranon fossa. The external perimeter defines the total subperiosteal area (TA). The internal perimeter defines the medullary area (MA) and the shaded area is the cortical area (CA). The principal X and Y axes are defined by the minor and major area moments of inertia.
MA
CA
Fig. 2. The total sub-periosteal area (TA), medullary area (MA), and cortical area (CA) at the proximal (A) and distal (B) humeral shaft of the three groups: sedentary controls, light swimming training (LST), and moderate swimming training (MST). The difference between each of the two swimming groups and the controls is illustrated by the area above the hatching in each bar. The percentage of change is indicated as well, with its level of significance (Mann-Whitney test).
Effect of swimming on bone architecture
two training groups, although the increase at the distal site of LST group was not significant. The increase in CA indicates that more bone was deposited at the subperiosteal surface than was resorbed at the endosteal surface. The percentage increase of the net cortical area in the MST group was nearly twice the increase in the LST group. The increase in
Moderale
swimming ?G
6
2
4
34 z 0. ‘5 2 .E a
2 0
0 Minor
Major
Major/Minor
Fig. 3. The minor and major principal moments of inertia (mm4) and their ratio, at the proximal (A) and distal (B) numerical shaft of the three groups: controls, LST and MST. The difference between each of the swimming groups and the controls is illustrated by the area above the hatching in each bar. The percentage change is indicated as well, with its level of significance (Mann-Whitney test).
847
the CA d both training groups was higher in the distal humeral shaft than in the proximal site. However, the difference between the two sites was not statistically significant. Figure 3 illustrates the mean values of the minor and major principal area moments of inertia at the proximal (Fig. 3A) and distal (Fig. 3B) sites for each group. Both area moments of inertia were significantly higher in the MST group than in the controls. In the LST group the increase was smaller but still significant except for the major area moment of inertia at the distal site. In order to determine whether the shape of the cortical cross-section in the training groups was geometrically similar to that of the control group, the ratio of the major to minor principal area moment of inertia was calculated. Figure 3 also shows that the ratio at both the proximal (Fig. 3A) and distal (Fig. 3B) sites in the training groups was lower than that in the controls. This finding, along with the fact that CA increased in the training groups, indicates that more bone tissue was deposited along the principal Y-axis (Fig. l), and therefore the cross-section of the bone became rounder following the swimming training. The results of the compression tests are presented in Fig. 4. This figure shows the mean value of the UCF and UCS at the proximal and distal humerus for each group. The UCF was higher in both training groups than in the controls; however, the differences were statistically significant in the distal site only. The differences in the UCS between the training groups and the controls were not significant at both sites, although they were more pronounced at the distal site. The fact that only differences in the UCF were statistically significant may indicate that the swimming training caused a greater increase in the overall strength of the bone rather than in the strength of bone tissue. -
6-d‘
Controls Light
! Z
Swimming
Moderate
Swimming
= p
-I
UCF-proximal
UCF-distal
UCS-proxirnal
80 1
UCS-distal
z 0 2
$
Fig. 4. The ultimate compressive force (UCF) and the ultimate compressive stress (UCS) at the proximal and distal humeral shaft of the three swimming groups: controls, LST and MST. The difference between each of the swimming groups and the controls is illustrated by the area above the hatching in each bar. The percentage change is indicated as well. with its level of significance (Mann-Whitney test).
A. SIMKIN
848 DISCUSSION
This study demonstrated that swimming activity in young rats resulted in a higher total subperiosteal and medullary cross-section areas compared to sedentary animals. This indicates that both endosteal resorption and periosteal apposition were greater in the swimming rats than in the sedentary controls. However, the latter exceeded the former, resulting in a significantly greater cortical cross-section area in the swimming rats. Examination of the major and minor principal area moments of inertia revealed that the additional increase of bone mass in the swimming groups was not equally distributed around the centroid of the cortical cross-section. The cortical mass increased along the principal Y axis more than along the X axis. This finding may be related to the fact that the nature of the strain distribution within the bone plays an important role in bone adaptation. It has been shown that when the strain exerted on the bone is substantially different from that which is experienced normally, there is an adaptive remodelling response (Lanyon et al., 1982a, 1982b). This appears also to be the case in the present study since the nature of the forces and moments acting on the humerus during swimming is different from that during locomotion. Electromyographic measurements (amplitude and frequency) have demonstrated that muscular activities in a swimming rat are substantially different from those during treadmill walking (Jasmin and Gardiner, 1987). Our observations on the movements of swimming rats (Fig. 5) have shown that they include not only flexion and extension, as in locomotion, but also rotation and abduction. Moreover, terrestrial locomotion can be characterized by two distinct phases in each limb: a low load swing phase and a high load stance phase, while in swimming there is a continuous resistance force of water during all phases of movement. It should be noted that since all the rats grew during the training period and obviously increased their bone mass, the bone cross-section parameters of the swimming rats had to be compared to those of controls of the same age. The observed differences in the cortical area between the groups could be attributed to either the direct mechanical effect of swimming training on the bone or to an increased food intake which caused an accelerated growth of the swimming rats as reflected by their higher body weight. However, both LST and MST groups reached the same body weight at the end of the training, yet the increase in cortical area of the MST group was almost twice that of the LST group at both the proximal and distal sites. Moreover, the change in the shape of the bone cross-section cannot be attributed to the gain in weight, but, rather, to the strain regime imposed on the bone. Swimming, a non-weightbearing activity, is generally considered inadequate to increase bone mass. Nilsson and Westlin (1977) found that the bone min-
et al.
eral content in the distal femur of professional swimmers did not differ significantly from that of nonathletes. However, recently it has been found that the mineral content of the distal radius and the vertebrae was significantly higher in professional swimmers than in non-exercising controls (Orwoll et al., 1987). To the best of our knowledge, the effect of swimming training on the cross-sectional geometry of bone and on its strength has not been studied previously. Woo et al. (1981) studied cross-sections of the proximal femoral shaft in 1 yr-old miniature swine which underwent a 12 months training program of running, relative to untrained controls. They found that in the trained animals the cortical cross-section area was higher by 23%, the minor area moment of inertia by 27%, and the major moment of inertia by 21%. These results are similar to our findings for the moderate swimming group, except for the much higher increase (57%) in the minor area moment of inertia found in our study. As a result, the ratio of the mean major to the mean minor moments of inertia decreased by 5% in the study of Woo (our calculation) and by 26% in the present study, indicating a more pronounced change in the shape of the bone cross-section. In addition, Woo et al. found that the endosteal diameters were significantly lower in the trained animals, while the periosteal diameters were not significantly different between the two groups, indicating that the additional bone increase due to physical activity took place at the endosteal surface. We found, on the other hand, that both the periosteal and endosteal areas were significantly larger in the swimming rats, resulting in a net increase of bone tissue at the cross-section. The design of the present experiment did not include a separate group performing habitual physical activity, such as walking or running. Therefore, the effect of the modified strain distribution within the bone cannot be distinguished from that of the exercise per se. However, the greater change in the shape of the humeral cross-section in the present study, compared to that found by Woo et al., is attributed to the essential difference in loading regimes between running and swimming as discussed above. Other factors as well, such as the type of animls used in each study, the degree of their maturity, the type of bone examined, and the length of the training period could have contributed to the variance in the results. When the resistance of the bone to compression was examined in the present study, a significant increase was found in the UCF at the distal humerus in both groups of swimming rats. The UCS at this site was also higher in both groups as compared to the controls, but the differences were not significant. These results indicate that swimming training increased the overall compressive strength of bone rather than its intrinsic mechanical properties. This increase in strength may be a result of the higher cortical crosssection area of the bone. Similarly, WOO et al. (1981) did not find any significant changes in the ultimate stress and modulus of elasticity of femoral bone tissue following running training.
849
Effect of swimming on bone architecture The results
of this study indicate
that swimming
training in growing rats increases their bone strength. Swimming activity seems to provide mechanical stimuli of sufficient intensity and diversity to induce an increase in bone tissue and its redistribution within the bone cross-section.
REFERENCES
Jasmin, B. J. and Gardiner, P. F. (1987) Patterns of EMG activity of rat plantaris muscle during swimming and other locomotor activity. J. appl. Physiol. 63, 713-718. Lanyon, L. E., Goodship, A. E., Pye, C. G. and MacFie, H. (1982a) Mechanically adaptive bone remodelling. A quantitative study on functional adaption in the radius following ulna osteotomy in sheep. J. Biomechanics 15, 141-154. Lanyon, L. E., Rubin, C. T., O’Connor, J. A. and Goodship, A. E. (1982b) The stimulus for mechanically adaptive bone remodelling. Osteoporosis (Edited by Menczel, J., Robin, G. C.. Makin. M. and Steinberg,_. R.), DV. 135-147. John Wiley; Chichester. Lovejoy, C. O., Burstein, A. H. and Heiple, K. G. (1976) The biomechanical analysis of bone strength: a method and its 1..
851
application to platycnemia. Am J. phys. Anthropol. 44, 489-506. Nagurka, M. I. and Hayes, W. C. (1980) An interactive graphics package for calculating cross-sectional properties of complex shapes. J. Biomechanics 13, 59-64. Nilsson, B. E. and Westlin, N. E. (1977) Bone density in athletes. Clin. Orthop. 77, 177-182. O’Connor, J. A., Lanyon, L. E. and MacFie, H. (1982) The influence of strain rate on adaptive bone remodelling. J. Biomechanics 15, 767-781. Orwoll, E. S., Ferat, J. L., Oviatt, S. K. and Huntington, K. (1987) The effect of swimming exercise on bone mineral content. Clin. Res. 35, 194A. Roux, W. (1881) Der Zuchtende Kampf der Teile, oder die ‘Teilauslase’ in Organismus (Theorie der ‘jiinktionallen Anpassung’). W. Engelmann, Leipzig.
Rubin, C. T. and Lanyon, L. E. (1984) Regulation of bone formation by applied dynamic loads. J. Bone Jt Surg. 66A, 397402. Wolff, J. (1892) Das Gesetz der Transformation der Knochen. A. Hirschwald, Berlin. Woo, S. L. Y., Kuei, S. C., Amiel, D., Gomez, M. A., Hayes, W. C., White, F. C. and Akeson, W. H. (1981) The effect of prolonged physical training on the properties of long bone: a study of Wolff’s Law. J. Bone Jt Surg. 63A, 78G-786.
Vol 22. No. 819. pp. 845-851,
cO21-9290189 %3.oo+.W PergamonPress plc
1989.
Prmted1"GreatBr~lain
THE EFFECT OF SWIMMING ACTIVITY ON BONE ARCHITECTURE IN GROWING RATS A.
SIMKIN*$,
I. LEICHTER~~,A.
SWISSA$
and S.
SAMUELOFF*
*Cosell Center for Physical Education, Hebrew University, Givat Ram, Jerusalem; tJerusalem College of Technology, 21 Havaad Haleumi Street, P.O. Box 16031,91160 Jerusalem; $Department of Orthopaedic Surgery and §Jerusalem Osteoporosis Center, Hadassah University Hospital, Jerusalem. Israel Abstract - The effect of non-habitual physical activity on bone architecture in the rat humeral shaft was examined. Two groups of rats were trained to swim for 1 h a day, for 20 weeks, at Iwo training levels. The control group consisted of sedentary rats. Parameters of cross-sectional bone morphology (cross-section areas, principal area moments of inertia and their ratio) were used to evaluate the response of bone architecture to mechanical loading. The strength of bone was assessed by measuring the ultimate compressive force and stress. The cortical cross-section area and principal moments of inertia were found to be significantly higher in the swimming groups than in the controls. Examination of the ratio between the major and minor moments of inertia revealed a pronounced change in the shape of the bone cross-section which became more rounded following swimming training. The ultimate compressive force was significantly higher in the swimming rats while the changes in ultimate stress were not significant. Our results indicate a gain of bone strength due to increased periosteal apposition and modified bone tissue distribution. The marked changes in bone morphology are attributed to the different nature of the forces and moments exerted on the humerus during swimming compared to those prevailing during normal locomotion.
INTRODUCTION
Living bone undergoes an adaptation process in which the structure of the bone is continually modified in response to the stresses imposed on it by mechanical loading. Since this phenomenon was investigated at the end of the 19th century (Roux, 1881; Wolff, 1892), many experimental studies have verified the functional adaptation of bone. In a series of experiments in the last decade, attempts were made to isolate the mechanical factors which affect bone growth and to quantify their relationship to the adaptation process (Lanyon et al., 1982a; Lanyon et al., 1982b; O’Connor et al., 1982; Rubin and Lanyon, 1984). These studies have shown that an increase in bone mass is engendered by intermittent dynamic loading. The amount and architecture of the additional bone tissue depends on the strain level, strain rate, daily number of loading cycles and the variation of strain distribution within the bone. The last factor expresses the dissimilarity between the strain distribution under experimental conditions and that undergone by the bone as a result of habitual activity. The importance of the modification in strain distribution was determined in two sets of in uiuo experiments. In the first (Lanyon et al., 1982a), the radius of a sheep was forced to carry the entire load on the forearm during locomotion by removing a portion of the ulna. In the second (Rubin and Lanyon, 1984), a functionally isolated avian ulna was artificially loaded. In these experiments, the unnatural load distribution was obtained as a result of an external intervention. Received in final form 2 September 1988.
In the search for an activity which would provide a non-habitual but physiological loading regime in animals, we observed that the upper limbs of swimming rats undergo movements which differ from those used in habitual, terrestrial locomotion. In the present study, therefore, the effect of swimming on bone growth was examined in rats which were trained to swim daily for lh. The adaptation of bone architecture in response to mechanical loading was studied by using parameters of cross-sectional bone morphology (Lovejoy et al., 1976; Woo et al., 1981) which enable the investigation of bone tiss’ue redistribution within its cross-section. MATERIALSAND METHODS
Forty female Sabra albino rats were used in this study. At the beginning of the study, they were five weeks old and their mean weight was 150 g, with a standard deviation of 15 g. They were randomly allocated into three groups of different training levels. The first group of 20 rats was trained to swim for 1 h a day, five days a week, for 20 weeks. This training level was considered light swimming training (LST). A second group of 11 animals underwent the same training regime, but was required to swim with additional lead weights tied to the roots of their tails. Each weight was equal to 1% of the rat’s weight in air. This training level was considered moderate training (MST). A third group of nine animals was a non-swimming control group (CONTROLS). The swimming was performed in a cylindrical water container 50cm in diameter and 70cm deep. The temperature of the water was kept at 35°C + 1°C. While not swimming, the animals were kept in cages at a temperature of
845
846
A.
SIMKIN
22°C + 1 “C on a 12 h light+lark cycle, with water and food ad lib&urn. At the end of the training period, the animals were sacrificed. The forelimbs were removed and the left humeri were disarticulated and all soft tissues were carefully removed. Two specimens for a compression test were prepared from the shaft of each humerus: a proximal one, extending 5 mm distally from the borderline between the shaft and the humeral head and a distal one, extending 5 mm proximally from the proximal edge of the olecranon fossa. Each specimen was prepared using a miniature lathe with a circular saw, 80 x 0.6mm, under continuous irrigation with normal saline. Both faces of each specimen were polished manually with No. 320 silicon carbide paper in order to obtain two parallel surfaces. A photograph of each surface was taken next to a millimetric scale. A magnified print (x 25) of each photograph was then prepared. Each specimen was then compressed in a materialtesting machine (Instron, Model 1125) at a constant deformation rate of 2 mm min-’ (strain rate of approximately 0.007 s- 1). The force-deformation curve for each specimen was observed during compression and after a maximum of the curve was reached, the compression was stopped. The value of the force at the peak of the curve was taken as the ultimate compressive force (UCF). The ultimate compressive stress (UCS) was calculated as the ratio between UCF and the smaller of the two cortical cross-section areas at the two faces of the specimen. The cortical cross-section area (CA) was measured from the magnified prints using the computer integrative-graphics program of Nagurka and Hayes (1980). It was computed as the difference between the total subperiosteal cross-section area (TA) and medullary area (MA) (Fig. 1). Using the same computer program, the minor and major principal area mo-
et al.
ments of inertia were calculated in relation to the two principal axes intersecting at the centroid of the cortical cross-section. The area moment of inertia provides a measure of bone strength and rigidity under bending. The cross-section areas and the area moments of inertia were computed for both faces of each specimen. The value of each parameter was taken as the average of the results obtained for the two faces of the specimen. The ratio between the major and minor area moments of inertia was calculated in order to assess the differences in the shape of the cortical crosssection between the swimming rats and the controls. No significant difference between the ratios would indicate that changes in CA following swimming were evenly distributed around the centroid of the crosssection. An increase or a decrease in the ratio would indicate that the cross-section of the bones in the swimming rats became flatter or rounder, respectively. RESULTS
Figure 2 shows the mean values of the cross-section areas (TA, MA, CA) of the specimens from the proximal (Fig. 2A) and distal (Fig. fZB)humeral shafts for each group. Both TA and MA values in the LST and MST groups were significantly higher than those in the controls, in both the proximal and distal sites. The CA (CA= TA- MA), which represents the net crosssection area of cortical bone, was also higher in the
Controls Light swimming Moderate swimming
’ pa.05 ” p(O.005
Principal Y axis
I
TA
Fig. 1. A cross-section of the rat humeral shaft, 5 mm proximal to the edge of the olecranon fossa. The external perimeter defines the total subperiosteal area (TA). The internal perimeter defines the medullary area (MA) and the shaded area is the cortical area (CA). The principal X and Y axes are defined by the minor and major area moments of inertia.
MA
CA
Fig. 2. The total sub-periosteal area (TA), medullary area (MA), and cortical area (CA) at the proximal (A) and distal (B) humeral shaft of the three groups: sedentary controls, light swimming training (LST), and moderate swimming training (MST). The difference between each of the two swimming groups and the controls is illustrated by the area above the hatching in each bar. The percentage of change is indicated as well, with its level of significance (Mann-Whitney test).
Effect of swimming on bone architecture
two training groups, although the increase at the distal site of LST group was not significant. The increase in CA indicates that more bone was deposited at the subperiosteal surface than was resorbed at the endosteal surface. The percentage increase of the net cortical area in the MST group was nearly twice the increase in the LST group. The increase in
Moderale
swimming ?G
6
2
4
34 z 0. ‘5 2 .E a
2 0
0 Minor
Major
Major/Minor
Fig. 3. The minor and major principal moments of inertia (mm4) and their ratio, at the proximal (A) and distal (B) numerical shaft of the three groups: controls, LST and MST. The difference between each of the swimming groups and the controls is illustrated by the area above the hatching in each bar. The percentage change is indicated as well, with its level of significance (Mann-Whitney test).
847
the CA d both training groups was higher in the distal humeral shaft than in the proximal site. However, the difference between the two sites was not statistically significant. Figure 3 illustrates the mean values of the minor and major principal area moments of inertia at the proximal (Fig. 3A) and distal (Fig. 3B) sites for each group. Both area moments of inertia were significantly higher in the MST group than in the controls. In the LST group the increase was smaller but still significant except for the major area moment of inertia at the distal site. In order to determine whether the shape of the cortical cross-section in the training groups was geometrically similar to that of the control group, the ratio of the major to minor principal area moment of inertia was calculated. Figure 3 also shows that the ratio at both the proximal (Fig. 3A) and distal (Fig. 3B) sites in the training groups was lower than that in the controls. This finding, along with the fact that CA increased in the training groups, indicates that more bone tissue was deposited along the principal Y-axis (Fig. l), and therefore the cross-section of the bone became rounder following the swimming training. The results of the compression tests are presented in Fig. 4. This figure shows the mean value of the UCF and UCS at the proximal and distal humerus for each group. The UCF was higher in both training groups than in the controls; however, the differences were statistically significant in the distal site only. The differences in the UCS between the training groups and the controls were not significant at both sites, although they were more pronounced at the distal site. The fact that only differences in the UCF were statistically significant may indicate that the swimming training caused a greater increase in the overall strength of the bone rather than in the strength of bone tissue. -
6-d‘
Controls Light
! Z
Swimming
Moderate
Swimming
= p
-I
UCF-proximal
UCF-distal
UCS-proxirnal
80 1
UCS-distal
z 0 2
$
Fig. 4. The ultimate compressive force (UCF) and the ultimate compressive stress (UCS) at the proximal and distal humeral shaft of the three swimming groups: controls, LST and MST. The difference between each of the swimming groups and the controls is illustrated by the area above the hatching in each bar. The percentage change is indicated as well. with its level of significance (Mann-Whitney test).
A. SIMKIN
848 DISCUSSION
This study demonstrated that swimming activity in young rats resulted in a higher total subperiosteal and medullary cross-section areas compared to sedentary animals. This indicates that both endosteal resorption and periosteal apposition were greater in the swimming rats than in the sedentary controls. However, the latter exceeded the former, resulting in a significantly greater cortical cross-section area in the swimming rats. Examination of the major and minor principal area moments of inertia revealed that the additional increase of bone mass in the swimming groups was not equally distributed around the centroid of the cortical cross-section. The cortical mass increased along the principal Y axis more than along the X axis. This finding may be related to the fact that the nature of the strain distribution within the bone plays an important role in bone adaptation. It has been shown that when the strain exerted on the bone is substantially different from that which is experienced normally, there is an adaptive remodelling response (Lanyon et al., 1982a, 1982b). This appears also to be the case in the present study since the nature of the forces and moments acting on the humerus during swimming is different from that during locomotion. Electromyographic measurements (amplitude and frequency) have demonstrated that muscular activities in a swimming rat are substantially different from those during treadmill walking (Jasmin and Gardiner, 1987). Our observations on the movements of swimming rats (Fig. 5) have shown that they include not only flexion and extension, as in locomotion, but also rotation and abduction. Moreover, terrestrial locomotion can be characterized by two distinct phases in each limb: a low load swing phase and a high load stance phase, while in swimming there is a continuous resistance force of water during all phases of movement. It should be noted that since all the rats grew during the training period and obviously increased their bone mass, the bone cross-section parameters of the swimming rats had to be compared to those of controls of the same age. The observed differences in the cortical area between the groups could be attributed to either the direct mechanical effect of swimming training on the bone or to an increased food intake which caused an accelerated growth of the swimming rats as reflected by their higher body weight. However, both LST and MST groups reached the same body weight at the end of the training, yet the increase in cortical area of the MST group was almost twice that of the LST group at both the proximal and distal sites. Moreover, the change in the shape of the bone cross-section cannot be attributed to the gain in weight, but, rather, to the strain regime imposed on the bone. Swimming, a non-weightbearing activity, is generally considered inadequate to increase bone mass. Nilsson and Westlin (1977) found that the bone min-
et al.
eral content in the distal femur of professional swimmers did not differ significantly from that of nonathletes. However, recently it has been found that the mineral content of the distal radius and the vertebrae was significantly higher in professional swimmers than in non-exercising controls (Orwoll et al., 1987). To the best of our knowledge, the effect of swimming training on the cross-sectional geometry of bone and on its strength has not been studied previously. Woo et al. (1981) studied cross-sections of the proximal femoral shaft in 1 yr-old miniature swine which underwent a 12 months training program of running, relative to untrained controls. They found that in the trained animals the cortical cross-section area was higher by 23%, the minor area moment of inertia by 27%, and the major moment of inertia by 21%. These results are similar to our findings for the moderate swimming group, except for the much higher increase (57%) in the minor area moment of inertia found in our study. As a result, the ratio of the mean major to the mean minor moments of inertia decreased by 5% in the study of Woo (our calculation) and by 26% in the present study, indicating a more pronounced change in the shape of the bone cross-section. In addition, Woo et al. found that the endosteal diameters were significantly lower in the trained animals, while the periosteal diameters were not significantly different between the two groups, indicating that the additional bone increase due to physical activity took place at the endosteal surface. We found, on the other hand, that both the periosteal and endosteal areas were significantly larger in the swimming rats, resulting in a net increase of bone tissue at the cross-section. The design of the present experiment did not include a separate group performing habitual physical activity, such as walking or running. Therefore, the effect of the modified strain distribution within the bone cannot be distinguished from that of the exercise per se. However, the greater change in the shape of the humeral cross-section in the present study, compared to that found by Woo et al., is attributed to the essential difference in loading regimes between running and swimming as discussed above. Other factors as well, such as the type of animls used in each study, the degree of their maturity, the type of bone examined, and the length of the training period could have contributed to the variance in the results. When the resistance of the bone to compression was examined in the present study, a significant increase was found in the UCF at the distal humerus in both groups of swimming rats. The UCS at this site was also higher in both groups as compared to the controls, but the differences were not significant. These results indicate that swimming training increased the overall compressive strength of bone rather than its intrinsic mechanical properties. This increase in strength may be a result of the higher cortical crosssection area of the bone. Similarly, WOO et al. (1981) did not find any significant changes in the ultimate stress and modulus of elasticity of femoral bone tissue following running training.
849
Effect of swimming on bone architecture The results
of this study indicate
that swimming
training in growing rats increases their bone strength. Swimming activity seems to provide mechanical stimuli of sufficient intensity and diversity to induce an increase in bone tissue and its redistribution within the bone cross-section.
REFERENCES
Jasmin, B. J. and Gardiner, P. F. (1987) Patterns of EMG activity of rat plantaris muscle during swimming and other locomotor activity. J. appl. Physiol. 63, 713-718. Lanyon, L. E., Goodship, A. E., Pye, C. G. and MacFie, H. (1982a) Mechanically adaptive bone remodelling. A quantitative study on functional adaption in the radius following ulna osteotomy in sheep. J. Biomechanics 15, 141-154. Lanyon, L. E., Rubin, C. T., O’Connor, J. A. and Goodship, A. E. (1982b) The stimulus for mechanically adaptive bone remodelling. Osteoporosis (Edited by Menczel, J., Robin, G. C.. Makin. M. and Steinberg,_. R.), DV. 135-147. John Wiley; Chichester. Lovejoy, C. O., Burstein, A. H. and Heiple, K. G. (1976) The biomechanical analysis of bone strength: a method and its 1..
851
application to platycnemia. Am J. phys. Anthropol. 44, 489-506. Nagurka, M. I. and Hayes, W. C. (1980) An interactive graphics package for calculating cross-sectional properties of complex shapes. J. Biomechanics 13, 59-64. Nilsson, B. E. and Westlin, N. E. (1977) Bone density in athletes. Clin. Orthop. 77, 177-182. O’Connor, J. A., Lanyon, L. E. and MacFie, H. (1982) The influence of strain rate on adaptive bone remodelling. J. Biomechanics 15, 767-781. Orwoll, E. S., Ferat, J. L., Oviatt, S. K. and Huntington, K. (1987) The effect of swimming exercise on bone mineral content. Clin. Res. 35, 194A. Roux, W. (1881) Der Zuchtende Kampf der Teile, oder die ‘Teilauslase’ in Organismus (Theorie der ‘jiinktionallen Anpassung’). W. Engelmann, Leipzig.
Rubin, C. T. and Lanyon, L. E. (1984) Regulation of bone formation by applied dynamic loads. J. Bone Jt Surg. 66A, 397402. Wolff, J. (1892) Das Gesetz der Transformation der Knochen. A. Hirschwald, Berlin. Woo, S. L. Y., Kuei, S. C., Amiel, D., Gomez, M. A., Hayes, W. C., White, F. C. and Akeson, W. H. (1981) The effect of prolonged physical training on the properties of long bone: a study of Wolff’s Law. J. Bone Jt Surg. 63A, 78G-786.