Table 2 Summary of the articles included in the present review

Bakker, 2003 [44] Netherlands

Amsterdam Growth and Health Longitudinal Study (AGAHLS)

466 Both

27 years

10 years

-

Questionnaire developed for the AGAHLS. Semi-structured. Previous 3 months. Evaluation by metabolic activity score per week (METPA) and mechanical activity score (MECHPA) evaluated by sum of ground reaction force of each PA reported.

BMD

MECHPA was positively associated with BMD in males (β = 0,090, p < 0,001). Subjects of both sex in 2^nd and 3^rd quartile of METPA had greater BMD.

LS

Barnekow-Bergkvist, 2006 [57] Sweden

-

36 Females

15-17 years

20 years

Questions on participation in PA: leisure-time sports activity (yes/no), membership of a sports club (yes/no), and kind of activity/ies.

Questions about leisure-time PA were collected regarding type of activity and frequency of overall PA. Only weight-bearing activities were taken into account.

BMD

Girls who were members of a sports club showed higher adult BMD in all sites. There was no association between current weight-bearing PA and adult BMD.

TB

LS

FN

Baxter-Jones, 2008 [48] Canada

Saskatchewan Pediatric Bone Mineral Accrual Study (PBMAS)

154 Both

8-15 years

15 years

PAQ-C was used on children and PAQ-A on adolescents many times. Nine items scored on a five-point Likert-type scale. Age and sex-specific Z-score was determined. Individuals were ranked into quartiles according to Z-score: highest - active, middle two quartiles- average, lowest – inactive.

PAQ-AD. Used only for controlling the effect of past PA on BMC.

BMC

Active males during childhood/adolescence had higher BMD at TB and FN than inactive ones and active females had higher BMD at FN than inactive.

TB

LS

FN

Cooper, 1995 [58] England

-

153 Females

-

21 years

Participation in sports at school asked in adulthood. Classified in ≤2 h and 2 or more hours per week.

Duration of outdoor walking (none, 1–30, 31–60, 61–120 and >120 min/day), participation in sports and PA at work.

BMC

More than 2 hours/week of participation in sports at school was associated with higher BMD at FN. Duration of walking was positively associated with higher LS and FN BMD.

BMD

BMAD

LS

FN

Delvaux, 2001 [49] Belgium

Leuven Longitudinal Study on Lifestyle, Fitness and Health (LLSLFH)

126 Males

13 years

27 years

A standardized questionnaire was used. Sport activities during the past year were registered. From the reported time and frequency of sport participation, a global average score of hours per week was calculated. No distinction was made between weight-bearing activities and others.

The same questionnaire used on adolescence. Baecke questionnaire. Four indices were calculated: PA at work, sports activities during leisure-time, PA during leisure time excluding sports, and the total PA index as the sum of the three previous indices.

BMC

Sports at 13y were not associated with bone mass. Sports at 18y were positively associated with LS BMC. Occupational and leisure-time PA (excluding sports) were not related to bone mass. Baecke sports index was positively associated with TB BMD and LS BMC and BMD.

BMD

TB

LS

Groothausen, 1997 [43] Netherlands

Amsterdam Growth and Health Longitudinal Study (AGAHLS)

182 Both

13 years

14 years

Questionnaire developed for the AGAHLS. Semi-structured. Previous 3 months. PS determined from 0 to 3 according to ground reaction force of each PA. Two different PS scores: A – sum of all PS, B – the highest PS.

Same procedures used during adolescence period. PA evaluated at 21 and 27 years.

BMD

PS evaluated by sum of all PS in all periods (13–16 years;13–21 years; 21–27 years and whole period) was associated with LS BMD. PS evaluated by the highest PS was associated with LS BMD in 3 periods (13-21 years; 21–27 years and whole period).

PA evaluated from 13 to 16 years.

LS

Kemper, 2000 [46] Netherlands

Amsterdam Growth and Health Longitudinal Study (AGAHLS)

182 Both

13 years

16 years

Questionnaire developed for the AGAHLS. Semi-structured. Evaluation by metabolic activity score per week (METPA) and mechanical activity score (MECHPA) evaluated by sum of ground reaction force of each PA. PA evaluated from 13 to 16 years.

Same procedures used during adolescence period. PA evaluated at 21 and 27 years.

BMD

Positive association was found between METPA from 13 to 16 y and LS BMD only in males. MECHPA in young adulthood was associated with LS BMD in both sexes. No association was found between METPA and FN BMD. Positive association was found between MECHPA and FN BMD, adjusted for gender

LS

FN

Kemper, 2002 [45] Netherlands

Amsterdam Growth and Health Longitudinal Study (AGAHLS)

302 Both

13 years

19 years

Questionnaire developed for the AGAHLS. Semi-structured. Evaluation by metabolic activity score per week (METPA) and mechanical activity score (MECHPA) evaluated by sum of ground reaction force of each PA. PA evaluated from 13 to 16 years.

Same procedures used during adolescence period. PA evaluated at 21, 27, 29 and 32 years. Only measures performed on 21 and 27 years were analyzed.

BMD

METPA and MECHPA scores from total period (13-27y) were positively associated with LS BMD. METPA in the teenage period (13-16y) and MECHPA in young adulthood period (21-27y) were also positively associated with LS BMD.

LS

Lloyd, 2004 [50] United States

Penn State Young Women’s Health Study

75 Females

12 years

10 years

Questionnaire based on existing instruments. The questionnaire listed 28 activities, including school-based activities; outside-of-school organized activities; and individual activities. The cumulative sports exercise score was an arithmetic sum. Questionnaire applied at least once per year from 12 to 18y.

Same procedures used during adolescence period. Questionnaire applied at least once per year up to 22 years.

BMD

The cumulative sports-exercise score was positively correlated to FN BMD.

FN

McGuigan, 2002 [51] Ireland

Young Hearts Project

460 Both

12-15 years

10 years

PA scores in adolescence were calculated according to a method which assessed normal daily activity patterns based around the typical school day. Activities were scored from 1–100 according to their frequency, intensity and duration.

Modification of the Baecke questionnaire, which records work-related PA, sports-related PA, and non-sports leisure activity. A total activity score was obtained from the sum of scores in these domains to give a total score ranging from 3 to 15.

BMD

Exercise history was the most important predictor of LS BMD in men. PA was also the strongest predictor of FN BMD in men. The results were almost identical when using exercise data collected at the age of 12–15 years.

LS

FN

Mein, 2004 [56] Australia

-

62 Females

18.5 years

9 years

Physical activity questionnaire (PAQ) was used. It measured additional sporting pursuits. To compare the two questionnaires, the units of the PAQ scores were transformed by adding the product of the z-score of the PAQ and the SD of KPAS (sports and exercise index) to this mean of KPAS.

Kaiser Physical Activity Survey (KPAS) was used to evaluate habitual PA and exercise. Four indices could be calculated—Domestic, Occupational, Active Living, and Sports and Exercise. The average of these scores was expressed as a summary score.

BMD

Average PA was positively correlated with LS and FN BMD.

LS

FN

Neville, 2002 [52] Ireland

Young Hearts Project

443 Both

12-15 years

8-10 years

PA scores in adolescence were calculated according to a method which assessed normal daily activity patterns based around the typical school day. Activities were scored from 1–100 according to their frequency, intensity and duration.

Modification of the Baecke questionnaire, which records work-related PA, sports-related PA, and non-sports leisure activity.

BMD

In males, PA during adolescence was associated only with FN BMD. PA on young adulthood was associated with both LS and FN BMD in males. In females, the PA in both periods was not associated with BMD.

LS

FN

A total activity score was obtained from the sum of scores in these domains to give a total score ranging from 3 to 15.

Petit, 2004 [53] United States

Penn State Young Women’s Health Study

76 Females

12 years

10 years

Questionnaire based on existing instruments used from 12 to 18y. The questionnaire listed 28 activities, including school-based activities; outside-of-school organized activities; and individual activities. Cumulative sports exercise score was an arithmetic sum.

-

BMD

Sports exercise score during adolescence was not associated with FN BMD at 22 years and with change on BMD from 17 to 22 years.

FN

Uusi-Rasi, 2002 [60] Finland

-

92 Females

25-30 years

4.2 years

-

PA was classified into 4 categories according to type and frequency: (1) ‘high’ vigorous PA ≥ 2 times a week, (2) ‘moderate’ vigorous PA ≤ once a week or less demanding PA few times a week, (3) ‘low’ less demanding PA once a week or very light PA several times a week (4) ‘no activity’. Category 1 was the PA + group and categories 3 and 4 were PA– groups.

BMC

There were no statistically significant differences for the FN BMC between the PA + and PA– groups

FN

Uusi-Rasi, 2008 [59] Finland

-

133 Females

25-30 years

10 years

-

PA was classified into 4 categories according to type and frequency: (1) ‘high’ vigorous PA ≥ 2 times a week, (2) ‘moderate’ vigorous PA ≤ once a week or less demanding PA few times a week, (3) ‘low’ less demanding PA once a week or very light PA several times a week (4) ‘no activity’. Category 1 was the PA + group and categories 3 and 4 were PA– groups.

BMC

There was no statistical difference between PA + and PA- group at FN BMC evaluated by three repeated measures (baseline, 5-year, 10-year).

FN

Valimaki, 1994 [61] Finland

Young Finns study

264 Both

9-18 years

11 years

Subjects were asked about weekly frequency of PA exceeding 30 minutes per performance. This same question was used in baseline and 6-year follow-up. Having two or more weekly sessions was called 1 and less than two sessions was called 0. PA on childhood or adolescence was analyzed with PA on adulthood.

Same question was used in 10-year follow-up. The sum of the three years answers ranging from 0 to 3 was calculated.

BMD

LS BMD was greater in males with PA evaluated as score 3. FN BMD was greater in both males and females with PA evaluated as score 3 than subjects with other PA values.

LS

FN

Van Langendonck, 2003 [54] Belgium

Leuven Longitudinal Study of Lifestyle, Fitness and Health (LLSLFH)

154 Males

13 years

27 years

Sports participation inventory was used. Information about the types of sports and the time per week was obtained. The score for 13–18y was calculated. Other two different analyses of PA were performed. 1) PS determined from 0 to 3 according to ground reaction force of each PA. Sum of all PS scores was calculated. 2) Groups were created from the ground reaction force: high, moderate, light or nonimpact.

Same questionnaire used during adolescence asked at 30, 35 and 40y. PS from adulthood was added to PS obtained on adolescence. Groups obtained on second analysis according to engagement on high (H) or nonimpact (N) in each period were: HH, HN, NN. NH group was excluded of analysis, as well as subjects whose sports participation did not meet these criteria.

BMD

PS score during adulthood was a positive predictor of TB and LS BMD. HH group showed greater LS BMD than HN and NN groups.

TB

LS

Wang, 2003 [55] United States

Berkeley Bone Health Study (BBHS)

341 Females

9-10 years

10-15 years

PA level assessed by self reported habitual activities, with scores derived by using MET values and time estimates (years 1, 3, and 5–10). Sedentary activity assessed by self-reports of weekly hours of television-video viewing (years 1, 3, and 5–10).

-

BMC

Physical activity was not associated with bone mass. Only sedentary activity on pre-puberty was negatively associated with FN BMD and BMAD.

BMD

BMAD

TB

LS

FN

Welten, 1994 [47] Netherlands

Amsterdam Growth and Health Longitudinal Study (AGAHLS)

182 Both

13 years

14 years

Questionnaire measuring habitual PA in the last 3 months. Only PA with a minimal of 4 METs were considered. The average weekly time spent in 3 activity level was collected: light (4-7METs), medium heavy (7-10METs) and heavy (>10METs). Total PA/week was the product of the time spent per level of intensity (1, 2 or 3). Only PA with a weight-bearing component was used. Adolescent period was considered from 13 to 17y

Same questionnaire used during adolescence. PA on young adulthood was analyzed on period between 13 and 22y and on total period – between 13 and 28y

BMD

Weight-bearing PA in all periods (13-17y; 13-22y and 13-27y) was positively associated with LS BMD only in males.

LS