By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
The authors state that they have no conflicts of interest.
Physical activity during growth has been associated with altered cortical bone geometry, but it remains uncertain if the physical activity–induced increments in cortical bone size remain when the level of physical activity is diminished or ceased. The aim of this study was to investigate if physical activity during growth is associated with cortical bone geometry in currently inactive young men. In this study, 1068 men (18.9 ± 0.6 [SD] yr) were included. Cortical bone geometry at the tibia and radius were measured using pQCT. A standardized questionnaire was used to collect information about current and previous sport activity. Subjects who continued to be active (n = 678) and who had been previously active (n = 285) in sports had a wider cortical bone (periosteal circumference [PC], 4.5% and 3.2%, respectively) with increased cross-sectional area (CSA; 12.5% and 6.9%) of the tibia than the always inactive subjects (n = 82). In the currently inactive men (n = 367), regression analysis (including covariates age, height, weight, calcium intake, smoking, and duration of inactivity) showed that previous sport activity was independently associated with cortical bone size of the tibia (CSA and PC). Amount of previous sport activity explained 7.3% of the total variation in cortical CSA. Subjects, who ceased their sport activity for up to 6.5 yr previously, still had greater cortical PC and CSA of the tibia than always inactive subjects. The results from this study indicate that sport activity during growth confers positive effects on bone geometry even though sport activity is ceased.
Although variations in bone mass are mostly genetically determined, a lifestyle factor such as weightbearing physical activity makes a significant contribution. Peak bone mass (PBM) is the greatest amount of bone mass achieved during life at a given skeletal site, reaches its maximum during young adulthood, and declines with aging. One strategy to reduce the risk of development of bone fragility later in life is to optimize PBM, which could be done by regular weight-bearing physical activity.
Several cohort studies have indicated that physical training before and during puberty is associated with increased bone acquisition in children and young adults. A limitation with cohort studies reporting higher bone mass in exercising than in nonexercising children and adults is that selection bias could be the reason for the found associations rather than an effect by physical activity on bone mass. More importantly, intervention trials have reported bone mass gains in training children.
Type of exercise is probably also of importance, where maximum effect is believed to be achieved by weight-bearing physical activity including jumping actions, explosive actions like turning and sprinting, and fairly few repetitions rather than endurance or non–weight-bearing activities. The mechanical strength of bone and resistance against fracture is determined by both bone size and volumetric density. The fact that the resistance of bone to bending and torsion forces is related exponentially to its diameter makes the size of the bone an important contributor to bone strength. This means that even small increases in the outer circumference of a bone could make a substantial contribution to its strength and fracture resistance. Physical activity before and during puberty has been associated with an altered cortical bone geometry, especially attributed to periosteal augmentation. A previous randomized controlled trial (RCT) has shown that increased periosteal circumference remains present 12 mo after an exercise intervention in preschool children. It has, however, not been established for how long this cortical bone geometry alteration after physical activity will remain when the level of physical activity is decreased.
We have previously reported, with data from the Gothenburg Osteoporosis and Obesity Determinants (GOOD) study, that physical activity is associated with BMD and cortical bone size in young Swedish men and that the boys who began their physical activity before puberty had higher adult BMD and cortical bone size than boys who started training later. The aim of this study was to investigate if physical activity during growth was associated with cortical bone geometry parameters and volumetric BMD in currently inactive young adult men.
MATERIALS AND METHODS
The population-based GOOD study was initiated with the aim to determine both environmental and genetic factors involved in the regulation of bone and fat mass. Study subjects in the entire GOOD study were randomly identified using national population registers, contacted by telephone, and asked to participate in this study. A total of 1068 men (18.9 ± 0.6 yr of age) from the greater Gothenburg area were included. To be included in the GOOD study, subjects had to be >18 and <20 yr of age and willing to participate in the study. There were no other exclusion criteria; 48.6% of the contacted study subject candidates agreed to participate and were included in this study. The GOOD study was approved by the ethics committee at University of Gothenburg. Written and oral informed consent was obtained from all study participants. The GOOD cohort was found representative of the general young male population in Gothenburg. In this study, the 390 men that were sedentary at the time of inclusion were used for the extended analysis (Tables 2–4; Fig. 1).
Assessment of previous sport activity
A standardized questionnaire was used to collect information about type of previous sport activity, frequency of previous sport activity (times per week), age of previous sport activity onset (yr), and duration of previous sport activity (yr). A total of 308 subjects had previously participated in sport activity, whereas 82 subjects had never participated in any sport activity (always inactive). Complete data on previous sport activity were not available for 23 subjects, leaving 285 previously active subjects and 82 always inactive subjects in the extended analysis (Tables 2–4; Fig. 1). A total amount of previous sport activity for each subject was calculated by multiplying the frequency of previous sport activity (times per week) with duration of previous sport activity (yr) for each type of previous sport activity and summarizing all the products for all types of previous sport activity for each subject. Sport activity type (strain) was categorized according to peak strain score based on ground reaction forces of sport activity and classified according to a method previously described. Activities involving jumping actions (e.g., gymnastics, handball, basketball) were given a strain score of 3, activities including explosive actions like turning and sprinting (e.g., soccer, tennis, ice hockey) were given a strain score of 2, and other weight-bearing activities (e.g., jogging, martial arts, weight training) were given a strain score of 1. Nonimpact activities (e.g., swimming, bicycling, and sailing) were given a strain score of 0. The following types of sport activity, with related strain score, were the most common among subjects that previously had been active in sports (each subject could have participated in several types of sports): soccer (n = 201, strain score = 2), martial arts (n = 44, strain score = 1), floor ball (n = 37, strain score = 2), handball (n = 36, strain score = 3), ice hockey (n = 30, strain score = 2), and tennis (n = 29, strain score = 2). To analyze the role of both type and amount of previous sport activity on bone parameters, we calculated an osteogenic index based on a previously described method. The osteogenic index was constructed by multiplying the amount of previous sport activity [(times/week) × years] with the sport activity strain score (strain score 0–3 based on known ground reaction forces) for each type of previous sport activity and summarizing all the products for all types of previous sport activity for each subject. The group of previously physically active subjects was divided into three equal parts according to the duration of inactivity (yr). Tertiles divided the group as follows: group 1, >0 and ≤2.24 yr (n = 95); group 2, >2.24 and ≤4.26 yr (n = 95); group 3, >4.26 yr (n = 95).
To study whether associations between physical activity and cortical bone geometry were site specific, two subgroups of previously physically active subjects were selected. The first group consisted of subjects formerly involved in sports where loading was mainly confined to the lower extremities (n = 71; soccer), excluding subjects that sometimes during life had participated in any sport activity involving upper extremities. The second group contained subjects formerly involved in types of sports including loading of both the upper and lower extremities (n = 101; ice hockey, bandy, weight lifting, martial arts, and American football). In this group, each subject could have participated in several types of sports during different periods in life.
Assessment of covariates
Height and weight were measured using standardized equipment. The CV values were <1% for these measurements. A standardized questionnaire was used to collect information about nutritional intake (dairy products) and smoking. Calcium intake was estimated from dairy product intake. Grip strength was assessed using a Jamar hydraulic hand dynamometer (5030J1; Jamar, Jackson, MI, USA) with adjustable handgrip. The subjects were sitting in a standard chair with both the forearm and dynamometer resting on a table. The subjects were asked to hold the dynamometer firmly and in an upright position and squeeze the handle as hard as they could. Three trials of each hand were performed. The results were recorded in kilograms of force, and the mean value of the three results for the nondominant hand was used in this study.
Assessment of cortical bone geometry and trabecular volumetric BMD
A pQCT device (XCT-2000; Stratec Medizintechnik, Pforzheim, Germany) was used to scan the distal leg (tibia) and the distal arm (radius) of the nondominant leg and arm, respectively. A 2-mm-thick single tomographic slice was scanned with a voxel size of 0.50 mm. The cortical volumetric BMD (vBMD; not including the bone marrow; mg/cm3), cortical cross-sectional area (CSA, mm2), endosteal and periosteal circumference (EC and PC, mm), and cortical thickness (mm) were measured using a scan through the diaphysis (at 25% of the bone length in the proximal direction of the distal end of the bone) of the radius and tibia. Trabecular vBMD (mg/cm3) was measured using a scan through the metaphysis (at 4% of the bone length in the proximal direction of the distal end of the bone) of these bones. Tibia length was measured from the medial malleolus to the medial condyle of the tibia, and length of the forearm was defined as the distance from the olecranon to the ulna styloid process. The CVs were <1% for all pQCT measurements.
All data were analyzed using SPSS software, version 15.0 for Windows. Differences in characteristics and bone parameters between subjects always inactive, previously active, and currently active were calculated using an independent samples t-test or by ANOVA followed by least significant difference posthoc test for continuous variables and χ2 test for categorical variables (Table 1). Regression analyses, including age, height, weight, calcium intake, and smoking as covariates, and previous and current physical activity (n = 1068), was used to evaluate the relative importance of ceasing and continuing sport activity on cortical bone geometry. In this regression model, dummy variables were constructed and used for sport activity: dummy variable 1 (subjects continuing physical activity were coded as 1 and all others as 0) and dummy variable 2 (subjects ceasing physical activity were coded as 1 and all others were coded as 0).
Table Table 1.. Anthropometric Characteristics and Bone Parameters of the Total GOOD Cohort Divided Into Those Who Have Always Been Physically Inactive, Those Who Have Ceased to be Physically Active, and Those Who Have Continued to be Physically Active
A total of 285 subjects with all information about previous sport activity available and 82 always inactive subjects were included in the analyses concerning previous sport activity in currently inactive subjects (Tables 2–4; Fig. 1). Bivariate correlations were tested using Pearson's coefficient of correlation (Table 2). The independent predictors of various bone parameters were tested using stepwise multiple linear regression analysis, including age, height, weight, calcium intake, smoking, duration of inactivity, and amount or osteogenic index of previous sport activity (Table 3). In the stepwise multiple linear regression analyses, the group of previously physically active subjects was divided into three equal parts according to total amount [(times/week) × years] and osteogenic index [(times/week) × years × (SA strain score)] of previous sport activity separately. Tertiles divided the group as follows: total amount of previous sport activity, first (coded as 2) >0 and ≤8.5 (n = 93), second (coded as 3) >8.5 and ≤21.5 (n = 93), third (coded as 4) >21.5 (n = 99); total osteogenic index of previous sport activity, first (coded as 2) >0 and ≤12.0 (n = 95), second (coded as 3) >12.0 and ≤38.5 (n = 92), third (coded as 4) >38.5 (n = 98). Always inactive (n = 82) subjects were included and coded as 1 in both groups. The stepwise selection process criterion for entry into the model was p ≤ 0.05, and the criterion for removal from the model was p ≥ 0.10. The percentage of the variation (R2), of each bone parameter, explained by amount of previous sport activity, osteogenic index, and by inactivity, together with all covariates was calculated, using the stepwise linear regression model. Inactivity was included in all regression models, whereas only amount of physical activity was included in the model, presenting the associations between inactivity and bone variables (Table 3, last column). R2 for each variable was calculated as the R2 change of the entire model when adding each variable, until all variables were included in the regression model (Table 3). R2 for each variable explained by amount of physical activity in the different groups of duration of inactivity (group 1–3; Table 4) was performed using the above-described regression model, including only the always inactive subjects and the subjects of each respective inactivity group. Parameters that did not show a normal distribution were logarithmically transformed before entered into the regression model. Differences in characteristics (Table 4) and various bone parameters (Fig. 1) between the subjects divided into always inactive and according to duration of inactivity were studied using ANOVA followed by a least significant difference posthoc test. χ2 tests were used to determine whether the distribution of smokers differed between the previously active and/or always inactive groups (Table 4).
Table Table 2.. Relationship Between Bone Parameters and Total Amount, Osteogenic Index of Previous Sport Activity, or Duration of Inactivity in Young Men
Table Table 3.. Linear Regression Analysis for the Relationship Between Bone Parameters and Total Amount, Osteogenic Index of Previous Sport Activity, or Duration of Inactivity in Young Men
Table Table 4.. Characteristics of the GOOD Cohort Divided Into Those Who Have Always Been Physically Inactive and According to Duration of Inactivity Among Those Who Have Ceased to be Physically Active
Characteristics and bone parameters of the cohort divided into always inactive, previously active, and currently physically active subjects
Comparisons of anthropometric characteristics and bone variables between the currently inactive subjects, including always inactive and ceased to be active subjects, and the remaining currently physically active subjects of the total GOOD cohort (n = 1068) are shown in Table 1. Anthropometric characteristics and bone variables on the 678 currently physically active subjects have previously been described. The subjects that remained physically active had higher calcium intake and grip strength, were younger, and smoked to a lower extent than the subjects who ceased to be active or subjects who were always inactive (Table 1). There were no differences between the 308 subjects that had ceased to be active and the 82 subjects that had never participated in any sport activity in age, height, length of the radius or tibia, weight, calcium intake, grip strength, or smoking (Table 1).
Subjects who continued to be active and who had been previously active in sports had a wider cortical bone (periosteal circumference, 4.5% and 3.2%, respectively) with increased cross-sectional area (12.5% and 6.9%) of the tibia than the always inactive subjects (Table 1). Subjects who continued to be active had a greater cortical cross-sectional area of the radius than always inactive subjects and those who had previously been active (5.6% and 3.1%, respectively), whereas both the currently and previously active subjects had greater cortical periosteal circumference of the radius than the always inactive subjects (3.7% and 2.1%, respectively; Table 1). Previous and current physical activity was also associated with increased trabecular vBMD and decreased cortical vBMD of both the radius and tibia (Table 1).
A regression analysis, including age, height, weight, calcium intake, smoking, and previous and current physical activity (n = 1068), was used to evaluate the relative importance of ceasing and continuing physical activity on cortical bone geometry. Continuing sport activity could explain 4.3% (β = 0.36, p < 0.001), whereas ceasing sport activity could explain 1.0% (β = 0.19, p < 0.001) of the tibia cortical cross-sectional area. The proportion of the tibia periosteal circumference variation explained by continuing sport activity was 1.4% (β = 0.27, p < 0.001) and by ceasing sport activity 0.9% (β = 0.18, p < 0.001). Both continuing and ceasing sport activity predicted radius periosteal circumference (continuing β = 0.21, p < 0.001, R2 = 1.4%; ceasing β = 0.10, p < 0.05, R2 = 0.3%), whereas only continuing sport activity predicted radius cortical cross-sectional area (continuing β = 0.12, p < 0.001, R2 = 1.4%; ceasing β = 0.06, p = 0.23).
Association of previous sport activity with cortical bone geometry and trabecular vBMD
The role of previous physical activity was studied in the 367 currently inactive subjects (285 previously active and 82 always inactive subjects) in whom all data regarding previous physical activity was available in all further analyses. In the currently inactive subjects (including both always inactive and previously active subjects), parameters reflecting cortical bone size (cross-sectional area, cortical thickness, and periosteal circumference) and trabecular vBMD of the tibia were correlated to both total amount and osteogenic index of previous sport activity, whereas an inverse correlation between these bone parameters and duration of inactivity was seen (Table 2).
We next used a linear regression model to distinguish whether amount of physical activity or duration of inactivity was the strongest predictor of bone parameters. In this analysis, both total amount and osteogenic index of previous sport activity were found to be independent predictors of parameters of cortical bone size (cross-sectional area, cortical thickness, and cortical periosteal circumference) of the tibia, whereas duration of inactivity did not predict any of these bone parameters (Table 3). At the radius, duration of inactivity, but not total amount of physical activity or osteogenic index, predicted cortical cross-sectional area and cortical periosteal circumference (Table 3). Total amount and osteogenic index of previous sport activity was inversely associated with cortical vBMD at both the radius and tibia and positively associated with trabecular vBMD at the radius (Table 3). Osteogenic index of previous sport activity was also positively associated, whereas duration of inactivity was inversely associated with trabecular vBMD at the tibia (Table 3). The total amount of previous sport activity explained between 2.7% and 7.3% of the total variation in parameters of cortical bone size and cortical vBMD of tibia, whereas osteogenic index of previous sport activity explained between 1.3% and 7.9% of the variation in the same bone parameters (Table 3). The total amount of previous sport activity and osteogenic index explained between 1.6% and 2.8% of the total variation in cortical endosteal circumference and cortical vBMD at the radius (Table 3). Osteogenic index of previous sport activity explained 2.7% and 2.9% of the variation in trabecular vBMD of the tibia and radius, respectively (Table 3). Inactivity could explain 2.2% of the variation in trabecular vBMD and 1.3% of the variation in radius cortical cross-sectional area (Table 3).
To evaluate if the associations between the parameters reflecting cortical bone size of the tibia and previous sport activity were affected by body composition parameters, we replaced weight with lean and fat mass in the stepwise linear regression analyses. In these analyses, amount and osteogenic index of previous sport activity both predicted cortical cross-sectional area (total amount β = 0.22, p < 0.001, R2 = 4.1%; osteogenic index β = 0.23, p < 0.001, R2 = 4.7%), cortical thickness (total amount β = 0.19, p < 0.001, R2 = 3.3%; osteogenic index β = 0.21, p < 0.001, R2 = 4.4%), and cortical periosteal circumference (total amount β = 0.13, p = 0.001, R2 = 1.4%; osteogenic index β = 0.12, p = 0.001, R2 = 1.4%). In this regression model (performed with total amount of previous sport activity, inactivity duration, age, height, smoking, and calcium intake as covariates), both lean and fat mass predicted tibia cortical cross-sectional area (lean mass β = 0.56, p < 0.001, R2 = 44.3%; fat mass β = 0.13, p < 0.001, R2 = 1.4%), cortical thickness (lean mass β = 0.43, p < 0.001, R2 = 13.0%; fat mass β = 0.11, p = 0.03, R2 = 1.0%), and periosteal circumference (lean mass β = 0.48, p = 0.001, R2 = 42.7%; fat mass β = 0.09, p = 0.03, R2 = 0.7%).
Subgroup analyses comparing always inactive subjects with 71 former soccer (loading mainly confined to the lower extremities) players were performed to study whether associations between sport activity and cortical bone geometry were site specific. Former soccer players had greater adjusted (for age, height, weight, calcium intake, and smoking) cortical cross-sectional area and periosteal circumference of the tibia (cross-sectional area 273.1 ± 27 versus 250.3 ± 23 mm2, p < 0.001; periosteal circumference 75.5 ± 3.4 versus 72.8 ± 3.4 mm, p < 0.001) but not of the radius (cross-sectional area, 95.4 ± 11 versus 93.2 ± 12 mm2, p = 0.25; periosteal circumference, 41.9 ± 2.6 versus 41.1 ± 2.8 mm, p = 0.09) than always inactive subjects. Always inactive subjects were also compared with 101 subjects formerly involved in physical activity types including loading of both the upper and lower extremities (e.g., ice hockey). This group of formerly physically active subjects had greater cortical bone size of the tibia (adjusted cross-sectional area, 261.3 ± 23 versus 250.3 ± 23 mm2, p = 0.002; adjusted periosteal circumference, 74.7 ± 3.8 versus 72.8 ± 3.4 mm, p < 0.001) and increased adjusted periosteal circumference (41.9 ± 2.6 versus 41.1 ± 2.8 mm, p < 0.05) but not adjusted cross-sectional area (95.7 ± 10 versus 93.2 ± 12 mm2, p = 0.13) of the radius than always inactive subjects.
Duration of inactivity and cortical bone size
The differences in cortical bone size between always inactive and/or subjects previously active in sports, divided into three equal parts according to duration of inactivity (yr), are presented in Fig. 1. There were no differences between the subjects in the different subgroups in age, height, weight, calcium intake, or smoking (Table 4). Both the adjusted (for age, height, weight, calcium intake, and smoking) cross-sectional area and adjusted periosteal circumference of the tibia were significantly higher in group 3 (longest inactivity duration; 2.9% and 2.5%), group 2 (7.0% and 2.3%), and group 1 (shortest inactivity duration, 7.3% and 3.0%) than in the always inactive group (Fig. 1). Cross-sectional area of the tibia was also higher in both group 2 (4.0%) and group 1 (4.3%) than in group 3 (Fig. 1). In group 1, amount of previous physical activity could explain 12.4% (β = 0.35, p < 0.001) and 4.4% (β = 0.22, p < 0.001) of the variance in tibia cross-sectional area and periosteal circumference, respectively, whereas the corresponding proportion of the variance explained in group 2 was 9.3% (cross-sectional area, β = 0.31, p < 0.001) and 4.8% (periosteal circumference, β = 0.23, p < 0.001). In group 3, amount of physical activity could explain 2.0% (β = 0.14, p = 0.01) and 3.2% (β = 0.18, p = 0.001) of the variation in tibia cross-sectional area and periosteal circumference, respectively.
No association between cross-sectional area, periosteal circumference, and duration of inactivity was found for the radius (Fig. 1).
In this study, we reported that previous sport activity was associated with increased cortical bone size in young adult currently physically inactive Swedish men. These associations were mainly found at the weight-bearing tibia and were in our regression analysis independent of inactivity duration.
In a recently published review, Hind and Burrows concluded that several RCTs have shown that weight-bearing exercise has positive effects on bone mass in boys but that the long-term effects are still unknown. Furthermore, they highlighted that these studies have used DXA techniques to assess the bone phenotype as an important limitation. Because of the limitation of the DXA technique, being 2D, a larger bone will have a falsely higher BMD and it cannot determine whether changes in BMD are caused by bone mineral or in bone geometrical parameters. Some reports have shown that athletes maintain the benefits, in terms of higher bone mass, of previous training when the level of physical activity is decreased or ceased, whereas other studies have reported that the benefits of physical activity are lost after its cessation. The limitation of using DXA to assess the bone phenotype can be attributed these previous studies.
By using pQCT, we could discriminate between which of the bone geometrical parameters that into adulthood still were associated with sport activity during growth. Some, but not all, previous studies using pQCT have reported that physical activity during childhood has a positive effect on cortical bone size. However, it is not known whether any positive effects of sport activity during childhood and adolescence on cortical bone geometry persist until adulthood in men. In our study, we showed that currently physically inactive men who had been active in sports during growth had greater cortical cross-sectional area, cortical thickness, and cortical periosteal circumference of the tibia than subjects who had never trained. Interestingly, previous amount of sport activity explained as much as 7.3% of the total variation in cortical cross-sectional area of the tibia, suggesting an important contribution of sport activity in the development of the cortical envelope of weight-loaded bones. The proportion of the variance in tibia cross-sectional area explained by previous sport activity was greater in subjects who had the least inactivity time (12.4% in group 1 with a mean inactivity duration of 1.2 yr) than in subjects with the longest inactivity time (2.0% in group 3 with a mean inactivity time of 6.5 yr). It is difficult to discern whether the above discrepancies were caused by differences in amount of physical activity or in inactivity duration between these groups, but our regression analysis indicated that amount of physical activity was the more important determinant of cortical bone size of the tibia.
Although we can show an independent association between previous sport activity (in currently inactive men) and cortical bone size, it should be pointed out that continuing sport activity is likely to further enhance cortical bone acquisition, as shown by comparing the previously and currently physically active subjects. Even though subjects continuing their sport activity had the largest cortical bone size of the tibia, we found that subjects who had previously trained had 6.9% larger cross-sectional area and 3.2% greater periosteal circumference than always inactive subjects, indicating remaining benefits of physical activity during growth. Each SD decrease in cortical cross-sectional area has in postmenopausal women been associated with 3.6 times increased radius fracture risk. If these results could be translated to young men in this study, where the differences in tibia cross-sectional area between previously active and always inactive subjects equal to ∼0.5 SD, previous physical activity could result in nearly halved risk of future fracture. The much smaller differences in structural parameters, as seen when comparing tibia cross-sectional area and periosteal circumference in always inactive subjects and the subgroup of previously active subjects with the longest duration of inactivity (6.5 yr), is likely of little clinical significance.
Previous sport activity was not associated with cortical cross-sectional area and thickness of the radius. At this bone site, previous sport activity was associated with increased endosteal and periosteal circumferences, reflecting both increased periosteal apposition and endosteal expansion, probably explaining the lack of association with cortical cross-sectional area and thickness.
Trabecular vBMD has previously been reported to be increased in the dominant arm versus the nondominant arm in female tennis players. In agreement with this study, our results show a positive association between previous sport activity and trabecular vBMD. In contrast, we found a negative association between cortical vBMD (measured without including the marrow) and previous sport activity, an association that has not previously been reported in other cohorts. We speculate that exercise-induced periosteal apposition and augmentation of the cortical bone size during growth is accompanied by a reduced cortical mineralization.
Because it is well known that sport activity types with high strain give the most favorable bone acquisition, we also calculated osteogenic index for sport activity, based both on the amount of previous sport activity and sport activity type. Lee et al. showed in a previous study on mice that engendered physiological strain resulted in an osteogenic response with periosteal apposition. It has also been reported that bone loading exercises in growing boys resulted in increased periosteal circumference. In this study, we showed that sport activity osteogenic index was associated with increased periosteal circumference, supporting these previous studies.
We also performed subgroup analyses to study whether the associations between amount of sport activity and cortical bone geometry were site specific. As could be expected, we found that subjects who previously participated in sport activity with mainly loading of the lower extremities (soccer) had greater cortical bone size of the tibia, whereas no differences were found at the radius. Another subgroup, with subjects previously involved in sport activity with loading of both the upper and lower extremities (e.g., ice hockey), was found to have a wider cortical bone (increased periosteal circumference) of both the radius and tibia, indicating that the greater cortical bone size was caused by loading associated sport activity.
There are some limitations associated with this study. The cross-sectional design does not allow direct cause–effect relationship between the studied parameters. It is possible that men with genetically larger and stronger bones could be more likely to be more successful in and participate to a higher extent in physical activity during growth. However, we could not find any difference in body size parameters (height, tibia or radius length, or weight) or in grip strength between subjects who had previously been active in sport activity and sedentary men. Although this argues against a problem with selection bias, it can not be ruled out to be the cause of the found associations. The inclusion rate in the population based GOOD study was 48.6%. Hence, selection bias could have been present in the original inclusion of the whole cohort. Furthermore, a large proportion of performed analyses in this study were performed on the 367 currently inactive subjects, who may possible not reflect the general male population at this age. This selection process could also have influenced the found associations.
By adjusting our regression model for body constitution parameters (lean and fat mass) we could show that the association between previous sport activity and cortical bone size was attenuated, suggesting that the effect of previous sport activity on cortical bone is mediated partly by current body constitution.
Lifetime sports activity participation was assessed using a retrospective self-reporting questionnaire, which could have limited the ability of the subjects to recall past sport activity and cause bias and misclassification. However, several studies have reported that people can recall activity patterns of up to 10 yr ago with high reliability and that recall of more vigorous activity is more accurate than recall of less intensive activities.
Our results indicate that the previous findings concerning maintained benefits in terms of greater cortical bone size of previous training when the physical activity is ceased probably derive from periosteal apposition. This periosteal augmentation confers considerable benefits in terms of greater bone strength and announces the possibility that physical activity during growth could be an important contributor to reduce the risk of fracture later in life.
In conclusion, we found that sport activity during childhood and adolescence was associated with increased cortical bone size in currently physically inactive Swedish young adult men, suggesting that sport activity during growth confers positive effects on bone geometry even though sport activity is ceased.