Change in physical activity from adolescence to early adulthood: a systematic review and meta-analysis of longitudinal cohort studies
=====================================================================================================================================

* Kirsten Corder
* Eleanor Winpenny
* Rebecca Love
* Helen Elizabeth Brown
* Martin White
* Esther van Sluijs

## Abstract

**Objective** To systematically review and meta-analyse how physical activity (PA) changes from adolescence to early adulthood (13–30 years).

**Data sources** Seven electronic databases were searched: Medline, Embase, PsycInfo, SCOPUS, ASSIA, SPORTdiscus and Web of Science.

**Eligibility criteria for selecting studies** English-language, longitudinal studies (from 01/1980 to 01/2017) assessing PA ≥twice, with the mean age of ≥1 measurement in adolescence (13–19 years) and ≥1 in young adulthood (16–30 years) were included. Where possible, data were converted to moderate-to-vigorous physical activity (MVPA) min/day, and meta-analyses were conducted between weighted mean differences (WMDs) in adolescence and adulthood. Heterogeneity was explored using meta-regression.

**Results** Of 67 included studies, 49 were eligible for meta-analysis. PA was lower during adulthood than adolescence WMD (95% CI) −5.2 (−7.3 to –3.1) min/day MVPA over mean (SD) 3.4 (2.6) years; heterogeneity was high (I2 >99.0%), and no predictors explained this variation (all p>0.05). When we restricted analysis to studies with data for males (n=29) and females (n=30) separately, there were slightly larger declines in WMD (−6.5 (−10.6 to –2.3) and −5.5 (−8.4 to −2.6) min/day MVPA) (both I2 >99.0%). For studies with accelerometer data (n=9), the decline was −7.4 (−11.6 to –3.1) and longer follow-up indicated more of a decline in WMD (95% CI) (−1.9 (−3.6 to –0.2) min/day MVPA), explaining 27.0% of between-study variation. Of 18 studies not eligible for meta-analysis, nine statistically tested change over time: seven showed a decline and two showed no change.

**Conclusion** PA declines modestly between adolescence and young adulthood. More objective longitudinal PA data (eg, accelerometry) over this transition would be valuable, as would investigating how PA change is associated with contemporaneous social transitions to better inform PA promotion interventions.

**Registration** PROSPERO ref:CRD42015030114.

*   Physical Activity
*   Adolescent

## Introduction

Physical activity in young people has been associated with a reduced risk of obesity,1 metabolic syndrome,2 beneficial effects on mental health,3 school performance,4 sleep duration5 and well-being.6 Most evidence suggests that physical activity declines with age throughout adolescence, although evidence on the magnitude of the decline is equivocal.7 In addition, inactivity may track into adulthood,8 resulting in greater health risks later in life.9 

The transition to adulthood often includes major life changes including relocation, changes in living arrangements, academic and employment status and new social roles.10 Such changes likely influence health-related behaviours.11 Evidence suggests that the transition from adolescence to early adulthood,12 including from high school to post high school,13 might particularly influence physical activity declines. However, the limited evidence is dominated by self-reported behaviour, which is susceptible to various forms of bias and there appears to be little objective data available.

How physical activity changes over the transition from adolescence to adulthood is unclear. Therefore, we asked, ‘How does physical activity change during the period from adolescence to early adulthood (age 13 to 30)?’ To our knowledge, no reviews have investigated physical activity over this transition, but the synthesis of the available evidence via a systematic review is important to better characterise change in physical activity over this important time.

Identifying characteristics of any changes may highlight potential strategies for physical activity promotion, including potentially distinguishing the respective importance of adolescence and young adulthood as targets for physical activity promotion. We aimed to identify the state of the evidence and pinpoint knowledge gaps regarding changes in physical activity occurring during the transition to adulthood. Specifically, we assessed whether physical activity changes between adolescence (13–19 years) and early adulthood (16–30 years) and the extent of this change; we hypothesised that physical activity declines over the transition from adolescence to adulthood.

## Methods

We conducted a literature search of longitudinal observational studies providing data on physical activity in young people (PROSPERO ref: CRD42015030114). We searched seven electronic databases (Medline via PubMed, Embase via Ovid, PsycInfo (EBSCO), SCOPUS, ASSIA via ProQuest, SPORTdiscus and Web of Science Core Collection) from 1980 to January 2017. We only included data collected after 1980 to ensure that results are relevant to contemporary populations while still representing 36 years of research. The search strategy focused on three themes: participants (including adolescents and young adults), outcomes (including moderate-to-vigorous physical activity (MVPA) and physical activity) and study type (including longitudinal and prospective). The full list of terms is available in online supplementary table 1; the search was conducted by EW (research associate). The search was undertaken in parallel with a search on dietary behaviour so the search strategy also contains diet-related terms.

### Supplementary file 1

[[bjsports-2016-097330supp001.docx]](pending:yes) 

### Inclusion criteria

Inclusion was restricted to published longitudinal data with at least two data collection points within the age range; the mean age of at least one measurement was required to be in adolescence (13–19 year) and at least one in young adulthood (16–30 years). The overlapping age definitions enable inclusive definitions of both periods and avoid arbitrary cut-offs limiting our search results. Thirty years old has been suggested as the end of young adulthood,14 and adolescence is defined by the WHO as 10–19 years.15 Transitions from childhood into adolescence or throughout adolescence have been covered by previous reviews,7 16 so we begin our review at age 13 to capture transitions occurring from mid-adolescence to early adulthood. The physical activity measurements included in the papers had to be at least 1 year apart. We were interested in a wide variety of absolute measures of physical activity but excluded studies presenting only sedentary behaviour. Studies reporting no absolute physical activity over at least two time-points (including some tracking studies or those only reporting % participants meeting guidelines) were excluded. We did not approach authors for additional information as we did not want to bias this review in the event of differential author response. Contacting authors may help to ameliorate incomplete reporting of outcomes that may lead to potential bias in systematic reviews of trials17; this is less of an issue in observational data.18 Moreover, previous reviews report limited response (42%) after multiple requests, with the data received biased towards more recent studies.19 All articles published in the English language in a scientific journal regardless of country of origin and containing data collected 1980 onwards were considered for inclusion. Inclusion criteria are presented in table 1.

View this table:
[Table 1](http://bjsm.bmj.com/content/53/8/496/T1)

Table 1 
Inclusion criteria

Three reviewers (KC/EW/EvS) independently reviewed the same 1500 results from the initial title and abstract search (in 3 sets of 500). Iterative comparison and discussion allowed for development of a consistent screening approach; screening of all remaining titles/abstracts was divided between KC, EW and EvS. Subsequently, two reviewers (KC/RL) independently screened full texts for inclusion and discussed discrepancies to reach consensus. Hand searching of the references identified three additional full texts, none of which were included.

Where data from the same study was reported in multiple papers, the following decision hierarchy was applied: (1) the paper with most time-points; (2) additional papers with further time-points; (3) objective data (vs self-report); (4) largest number of baseline participants; (5) earliest publication date; and (6) largest number of follow-up participants.

### Data extraction

KC conducted data extraction with 100% checked for accuracy (RL); discrepancies were resolved by discussion. Data extracted included: baseline date (date of the first wave meeting the inclusion criteria), study name, country, ethnicity, sex, socioeconomic status, and number of participants, age, physical activity measurement method and physical activity data at each relevant time-point. Any analysis of change was also recorded including whether any adjustments were made. Data were extracted for males and females separately where possible; if not available, data for the whole sample were extracted, and data for other subgroups were extracted if an overall measure by sex was not available.20–28 If data were not presented by sex or overall, data summarised by ethnicity were extracted if available22 23 26 28 and otherwise for whichever variable resulted in the least subgroups.

This review originally intended to address a second research question: ‘How do factors such as socio-economic status or particular lifestyle transitions influence these changes in physical activity behaviour?’ Data extraction revealed there was little heterogeneous information available to address this question, and we therefore only focus on the main research question.

### Assessment of risk of bias

Risk of bias was scored using an adapted version of a previously used scale (Effective Public Health Practice Project Quality Assessment Tool), which has been explained in detail previously29 and used in other physical activity reviews.30 Participant representativeness, study size, participant drop-out, quality of physical activity data and quality of change analyses were rated. These criteria were assessed by two independent reviewers (KC and HEB/EvS); in case of disagreement, consensus was derived by discussion with a further reviewer (EW). Each item was scored as ‘Strong’, ‘Moderate’ or ‘Weak’; if a paper provided insufficient information, then it was scored as ‘Weak’. Scores for each item were summed, and quality was defined as ‘Weak’ when at least one item was classed as ‘Weak’. Papers were classed as ‘Strong’ when three out of the five criteria were rated as ‘Strong’, and no items were scored as ‘Weak’; other studies were classed as ‘Moderate’. We scored the size of a study as ‘Strong’ if there were over 1000 participants at the follow-up with the least participants; using 1000 as the threshold was an arbitrary decision, but we deemed this suitable considering 22% of the included studies had ≥1000 participants. Analyses were scored as ‘Strong’ if they used a suitable statistical method to assess change (such as regression) and included appropriate adjustment for potential confounders. Studies were scored as ‘Strong’ if objective methods were used to assess physical activity and were the same at all time-points. If validity and reliability of self-reported methods were stated, and were the same over both time-points, then they were rated as ‘Moderate’. Scoring details are summarised in online supplementary table 2.

### Data preparation

Wherever possible, values from all included studies were converted to a common metric to enable meta-analysis of physical activity change. The metric chosen was the original unit reported for all studies with accelerometer data (MVPA min/day). If a paper presented more than one physical activity outcome, the one most comparable with MVPA was selected and otherwise the value most easily converted to MVPA min/day. The extent of data conversion was recorded for each study so it could be included in meta-regression. Conversion was defined as minimal when reported data were in min/day, min/week or similar and as extensive when any other conversion was necessary. As in a previous review,31 pedometer counts were translated to approximate MVPA using 100 steps/min, and a bout of activity was classed as 30 min. For data reporting MET minutes or MET times/week, the value reported was divided by the appropriate MET value if stated in the paper (or five for MVPA to account for the whole range of MVPA (≥4 METS32)).

Due to the overlapping adolescent and adulthood time periods, decisions were made to class study time-points as ‘adolescence’ or ‘adulthood’. If only two time-points per study were available, the first was classed as adolescence and the second as adulthood. For studies with >2 time-points, at least one time-point was included as adolescence and at least one as adulthood, any further time-points were divided at 17.5 years old (the midpoint between the adolescent and adulthood age ranges). The additional time-points <17.5 years were incorporated with the adolescent estimate and time-points ≥17.5 years were incorporated with the adult estimate so that mean activity values could be used where necessary; each time-point was used once. Sample size at follow-up(s) was assumed to be that of baseline when time-point specific participant numbers were not provided.33–35 Where reported, medians, IQR and variance were converted to mean and SD as suggested previously.36 It was not possible to convert all data to include in the meta-analysis; reasons for this were presenting an activity score/index with insufficient information to convert it into min/day,37–46 presenting an infeasible number of hours of activity (mean of 25 hours/day),47 summing multiple measurements annually48 and presenting energy expenditure per unit of body weight.49 A further seven studies did not report SD at one or more waves so the studies could not be included in the meta-analysis.21 50–55 

### Statistical analysis

Data for each adolescent-adulthood comparison were combined in STATA using a random effects meta-analysis (STATA V.14). As all studies included in the meta-analysis were converted to a common metric, non-standardised weighted mean differences (WMDs) were calculated. Variation attributable to heterogeneity was assessed using the I2 statistic. If heterogeneity (I2 >50%) was seen,18 meta-regression was used to test the impact of potential effect modifiers (baseline age, time between baseline and follow-up (the mean calculated for the adolescent and adult time periods used), region of study, baseline date, accelerometer vs non-accelerometer data, extent of conversion necessary to estimate MVPA). Effect modifiers were selected based on the data extracted and those used in a previous review.18 Risk of bias score was not investigated as a potential effect modifier due to limited variability. Age and time between measurements were calculated using mean ages reported (and/or time between measurements where relevant). Region of study was summarised into four categories: Europe, North America, lower-income and middle-income countries (LMICs; eg, Brazil and Iran) and other high income countries (eg, Australia and Japan). Baseline measurement date was taken from the included papers and referred to that used as baseline in this review where possible (other waves of data may have been available outside our inclusion criteria) but was not reported for nine studies.33 34 38 40 47 49 56 57 Variables that were associated (p<0.05) in single models were included in a multiple model to explore the variance that could be explained.

Due to sex-differences in biology and behaviour during adolescence,58 a post hoc decision was made to stratify meta-analyses by sex where data were available. In addition, due to the high heterogeneity of the data, a further post hoc meta-analysis was conducted only including accelerometer-assessed physical activity. For completeness, a separate meta-analysis including non-accelerometer methods was also conducted.

Lastly, the main meta-analysis was repeated restricting the age of baseline to ≤15 years old and ≥18 years old at follow-up to provide an estimate of change between more discrete age periods.

Funnel plot asymmetry and Eggers test for bias were conducted for all meta-analyses to investigate various forms of potential publication bias.

The results are reported in accordance with relevant guidance (Meta-Analysis of Observational Studies in Epidemiology (MOOSE)59 and Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement (PRISMA)60).

## Results

Of 47 072 articles identified (including physical activity and diet papers), 22 486 duplicates were excluded. The remaining 24 586 had titles and abstracts assessed for inclusion; 23 632 papers were excluded based on title/abstract. Of the remaining 954 papers, 544 included physical activity only and 120 included both physical activity and diet and were taken forward to full-text screening. Of those full texts screened, 541 were excluded with the reasons included in figure 1; this left 123 full-text papers. Of these, 52 full-text papers were excluded as they included duplicate data available in other included papers. This left 69 papers containing data on 73 cohorts (defined as a specific group followed over time; some papers include multiple cohorts) from 67 overarching studies; data from 49 of these studies were eligible for meta-analysis. Study characteristics for all included studies are available in online supplementary table 3. Descriptive characteristics of the included studies, presented as those that were meta-analysed and those that were descriptively summarised are shown in table 2.

![Figure 1](http://bjsm.bmj.com/https://bjsm.bmj.com/content/bjsports/53/8/496/F1.medium.gif)

[Figure 1](http://bjsm.bmj.com/content/53/8/496/F1)

Figure 1 
Evidence search and exclusion process.

View this table:
[Table 2](http://bjsm.bmj.com/content/53/8/496/T2)

Table 2 
Descriptive characteristics of the included studies, presented as those that were meta-analysed and those that were descriptively summarised (n=67 studies)

Online supplementary table 4 presents the results of individual item and overall risk of bias assessment scores. Initially, 87% agreement was achieved on risk of bias scoring with discrepancies resolved via discussion. Of the 71 papers included, 2 received a ‘Strong’ rating, 3 (4%) were rated as ‘Moderate’ and 66 (93%) as ‘Weak’. The large number of ‘Weak’ scores was mainly due to the strict criteria of any single item scoring as ‘Weak’ meaning that the overall rating for that paper was ‘Weak’.

### Meta-analyses

The pooled WMDs displayed in table 3 and figure 2 indicate that physical activity is lower during adulthood than during adolescence WMD (95% CI) −5.2 (−7.3 to –3.1) min/day MVPA over a mean 3.4 (SD: 2.6) years. There was a statistically significant (χ2=5.8e+06, p<0.001) high amount of heterogeneity present between studies (I2 >99.0%), and none of the included predictors explained this variation (all p>0.05).

![Figure 2](http://bjsm.bmj.com/https://bjsm.bmj.com/content/bjsports/53/8/496/F2.medium.gif)

[Figure 2](http://bjsm.bmj.com/content/53/8/496/F2)

Figure 2 
Change in physical activity (MVPA min/day) from adolescence to adulthood from all eligible studies. MVPA, moderate-to-vigorous physical activity.

View this table:
[Table 3](http://bjsm.bmj.com/content/53/8/496/T3)

Table 3 
Summary of physical activity reported in meta-analysed studies and meta-analysis results

Results for males and females (table 3 and see online supplementary figures 1 and 2 , respectively) both indicated a decline in physical activity from adolescence to adulthood, both with very high heterogeneity between studies.

### Supplementary file 2

[[bjsports-2016-097330supp002.pdf]](pending:yes) 

The results based only on accelerometer data suggested a slightly greater decline in MVPA between adolescence and adulthood −7.4 (−11.6 to –3.1) min/day MVPA with a high amount of heterogeneity present between studies (χ2=272.47, p<0.001, I2=95.0%) (table 3, figure 3). The only variable included in meta-regression that explained any of this variation was the difference in age between time-points with a greater age difference indicating more of a decline: WMD −1.9 (−3.6 to –0.2), p=0.03, with 27.0% of variation between studies explained. This value suggests that for every extra year of follow-up the decline is 1.9 min greater. When restricting to non-accelerometer data, the WMD was −6.3 (−8.1 to –4.4) min/day MVPA and none of the high amount of heterogeneity present between studies (χ2=5541.7, p<0.001, I2=99.2%) was explained by variables tested with meta-regression.

![Figure 3](http://bjsm.bmj.com/https://bjsm.bmj.com/content/bjsports/53/8/496/F3.medium.gif)

[Figure 3](http://bjsm.bmj.com/content/53/8/496/F3)

Figure 3 
Change in physical activity (MVPA min/day) from adolescence to adulthood from eligible studies using accelerometer-assessed physical activity.

When adolescence was restricted to ≤15 years-old and adulthood as ≥18 years-old, three studies were included.28 61 62 When restricted to these distinct age groups, results were non-significant, but the extent of the WMD was comparable with the previous results (WMD −8.2 (−23.8 to 7.5) min/day MVPA); heterogeneity remained high (χ2=38.9, p<0.001, I2=89.7%) (table 3).

Funnel plot asymmetry and Eggers test for bias suggested no evidence of asymmetry in any meta-analyses (all p>0.188), indicating that these tests showed no evidence of publication bias.

Of the 18 studies that were not eligible for meta-analysis, nine statistically tested for change in physical activity over time; of these nine studies, seven showed a decline in physical activity37 41 48 52 and two showed no change between adolescence and adulthood.21 47 All of these studies used questionnaire reported measurement methods with the exception of one using an activity diary49 and one using a recall.55 The median (IQR) age in adolescence was 15.9 (14.9–16.5) years and 19.5 (17.4–22.0) years in adulthood. Median time between baseline and follow-up measurements was 2.0 (1.4–7.0) years and baseline participant numbers were 166 (84–994).

## Discussion

Meta-analysis of published longitudinal data indicated that physical activity declines over the transition from adolescence to adulthood. Results from all 49 eligible studies indicated that daily MVPA declined by a mean of 5.2 min/day, which is equivalent to approximately 13% of the mean baseline value; this decline occurred over a mean duration of 3.4 (SD: 2.6) years. Stratified analyses showed slight sex differences, with a slightly greater decline observed in males. For studies assessing physical activity with accelerometers, a decline of 7.4 min of daily MVPA was observed. This equated to 17% of the mean baseline estimate and occurred over a mean follow-up of 4.3 (SD: 2.9) years. Half of the non-meta-analysed studies statistically tested change in physical activity between adolescence and adulthood; of those, the majority (78%) reported a decline.

### Interpretation

All meta-analyses displayed large heterogeneity. This limits the validity of pooling the results and estimating a single summary result; the results of these meta-analyses should therefore be interpreted with caution. However, random effects meta-analysis should allow for heterogeneity and to some extent have led to relatively conservative results with wider CIs and larger p values when heterogeneity exists.

We investigated potential sources of heterogeneity using meta-regression and by separately reporting results for studies reporting accelerometer-assessed physical activity and self-report assessed activity. It was only in the accelerometer-assessed analysis that any variables investigated in meta-regression explained any variation, with a greater age difference between adolescence and adulthood indicating slightly more of a decline in physical activity as would be expected due to there being more time for change to occur.

The slightly larger decline observed in males than females could be at least partly because boys were more physically active than girls in adolescence. Previous work has observed a larger physical activity decline in girls during early adolescence (<12 years old) but a greater decrease among boys later in adolescence,7 58 suggesting a reduction in sex differences in physical activity later in adolescence. Our results add to this evidence by suggesting that these sex differences may further reduce over the transition to adulthood.

Although the inclusion criteria for early adulthood required at least one measurement of physical activity between the ages of 16 and 30, the majority of these data points were at the younger end of this age range. The adulthood measure in 35% of all studies was over age 21 with only 8% over 25 years old. For studies using accelerometers to measure physical activity, 33% (3/9) had a follow-up after age 21 with 11% (1/9) after 25 years old.

The tool we used to assess risk of bias led to only 2 of 71 (2.8%) papers being scored as strong; 66 (93%) were scored as weak. Our criteria were strict as just one ‘weak’ rating leading to the overall classification as ‘Weak’, Nevertheless, this suggests a paucity of good quality data that is important to further investigate due to the potential importance of early adulthood for developing long lasting health behaviours.

As adolescence and early adulthood are times of increasing autonomy, positive behaviours established during this time could potentially last into later adulthood, making this transitional time a potentially important target for health promotion.63 Furthermore, moving from adolescence into adulthood is characterised by transitions between different social roles14 and major transitions including moving out of the family home, changes in school/work environment and financial circumstances. Many of these transitions are associated with health behaviours including physical activity, but further work would be useful to characterise physical activity over the young adulthood period while taking account of these transitions.11 The period of late adolescence to early adulthood has been suggested as an important, but overlooked, age to establish long-term health behaviour patterns.63 

### Relationship to prior knowledge

Although our results indicate that MVPA appears to decline by as much as 7 min/day between adolescence and early adulthood, the physical activity recommendations for adolescents (60 min/day MVPA) and adults (150 min/week of moderate or 75 min/week of vigorous activity) differ by more than this. As a weekly average, the adult recommendations suggest 21 min of moderate, or 10 min of vigorous activity daily, which is substantially less than the 60 min/day recommended in adolescence. Therefore, the absolute change in activity identified here of up to 7 min/day does not appear to account for the difference between adolescent (60 min/day) and adult physical activity recommendations (~10–21 min/day).

One could therefore argue that this decline is acceptable and should not present a health hazard at a population level. However, when this decline is put in the context of a meta-analysis from a large worldwide study2 indicating that a 10 min differencein MVPA was associated with a smaller waist circumference and lower fasting insulin among young people, this decline may have important health implications at the population level. Furthermore, the mean time between adolescence and adulthood measurements was 3.4 years, and although we did not test the linearity of changes over time, if this change was extrapolated over 10 years (eg, 15–25 years), it would amount to a potential decline of closer to 15 min/day. Further research with objective measures in high-quality studies that continue well into adulthood with multiple time-points are needed to further investigate this. Although we did not assess tracking of activity here, low physical activity in adolescence is likely to progress to adulthood inactivity,35 which is linked to increased risk of many diseases including type 2 diabetes, cancer and also mortality.9 

The evidence on changes in physical activity during other stages of the life course suggests declines of 4% per year throughout childhood64 and 7% per year during adolescence.7 The modest decline observed here is comparable in relative terms, equating to an approximate 13% decline compared with the baseline value over a mean follow-up duration of 3.4 years. Taken together, this evidence suggests that at a population level, physical activity declines steadily throughout the first decades of life and that preventing declines in physical activity throughout this period could be valuable. Studying population-level estimates may mask individual variations in patterns of change and the important influence of various social transitions that may occur at different times for different people (such as starting employment, moving out of the family home or having children). Further research on trajectories of change in physical activity behaviour may help unravel some of these uncertainties.

### Strengths and limitations

To our knowledge, this is the first study to meta-analyse change in physical activity over the transition between adolescence and adulthood. Although heterogeneity was high and results should be interpreted with caution, limiting the results to studies using accelerometer-assessed physical activity, which should reduce measurement error,65 indicated a similar decline. Despite the high heterogeneity, we believe that the review would be less useful without the meta-analyses. Although we investigated potential causes of heterogeneity using meta-regression, few factors were significant.

In the analysis focused on accelerometer-assessed activity, there was some indication that time between measurements may have explained some of the heterogeneity. We acknowledge slight deviations from our published protocol. We did not search in the Cochrane library (we excluded trials), CINAHL (not relevant as is nursing journal database) and Psychlit (not available at our institution). However, as reference checking of included papers did not identify additional papers, we are confident that we have done a comprehensive search. We did not conduct duplicate abstract screening, which is a limitation, although 1500 articles were triple screened to ensure consistency among the three reviewers. Furthermore, although data extraction was 100% checked that it was not extracted in duplicate.

We believe that defining adolescence as 13–19 years old is appropriate. Although menarche may occur for some girls before age 13, socially, especially in developed countries (source of 95% of the data) 13 years old is classed as at least mid-adolescence. Our analysis design distinguishes this review as we focused on the transition out of adolescence into adulthood rather than change in physical activity during adolescence, which has been reviewed previously.7 16 As we used overlapping age groups, an included study may have two time-points that are arguably both in adolescence (such as 13 and 17 years old). The sensitivity analysis restricting to studies reporting data on <15 years old (adolescence) and >18 years old (adulthood) indicated a non-significant decline of −8.2 (−23.8, 7.5) min/day MVPA, but with only three studies included with this strict age criteria, we believe using broader age categories was appropriate to be as inclusive as possible.

A strength of this work is that sufficient homogenous data were present to perform a meta-analysis. However, data conversion was necessary, requiring assumptions that may have led to error in the absolute values used, although this should not have greatly influenced the within-study change estimates. Similarly, we used the data presented in each paper that was most similar to MVPA. Sometimes this was not directly comparable (eg, overall activity) but was converted to min/day. As most studies did not statistically test change in physical activity and only provided descriptive data, this conversion and meta-analysis was useful to gain a tangible overview of results.

The review aimed to characterise change in physical activity between adolescence and adulthood, therefore we did not use all time-points separately as they were condensed into broad adolescent and adulthood age periods; annual change has been examined throughout adolescence7 and was beyond the scope of this review. We were keen to take study size into account so we believe it was most appropriate to use meta-analysis, rather than computing an estimate such as percentage annual change. The included studies were all published in English and mostly from North America, Europe and other high-income countries (95% of studies), so these results are not likely to be generalisable to LMICs. We only included peer-reviewed publications, which may be susceptible to publication bias; however, funnel plots did not indicate evidence of asymmetry.

### Implications for policy, practice and research

Physical activity declines between adolescence and adulthood. Average adult physical activity levels are too low to benefit a range of health outcomes. Continued efforts to maintain or increase adolescent physical activity and to prevent a decline into adulthood could have important public health benefits. More high-quality longitudinal physical activity data over this transition would be valuable, especially in the young adulthood period, to better characterise physical activity patterns over this time to inform how physical activity promotion interventions might most effectively target this important group.

## Conclusion

Published longitudinal data indicate that physical activity declines over the transition from adolescence to adulthood by an average of 5.2 min of MVPA per day. Both males and females had lower physical activity in adulthood compared with adolescence. Objectively assessed physical activity showed a slightly larger decline of 7.4 min of daily MVPA. More objective longitudinal physical activity data over this transition would be valuable, as would investigating how change in physical activity is associated with various contemporaneous social transitions.

### What is already known?

*   Physical activity is lower in adulthood than adolescence.

*   The adolescence–adulthood transition is important to target behaviour change interventions due to increasing autonomy, and positive behaviours established over this time have the potential to last into later adulthood.

### What are the findings?

*   Data from 49 studies indicate that over the transition from adolescence to adulthood moderate-to-vigorous physical activity (MVPA) declines by a mean of 5.2 min/day (or approximately 13% of the baseline value).

*   Most studies assessed physical activity by self-report, but the nine studies that used accelerometers showed a mean decline of 7.4 min of daily MVPA (17% of mean baseline value) between adolescence and adulthood.

*   The observed decline in physical activity was slightly larger in males than females (−6.5 vs −5.5 min/day) and, based on the accelerometer data, the decline was larger with each year of follow-up (further decline of −1.9 min/day per year).

### Supplementary file 3

[[bjsports-2016-097330supp003.docx]](pending:yes) 

## Footnotes

*   Contributors KC drafted the manuscript and conducted meta-analyses. EW conducted the searches. KC, EW and EvS screened abstracts for inclusion. KC and RL screened full texts for inclusion and were involved in data extraction. HEB and KC conducted risk of bias screening. All authors contributed to the research question, critically reviewed the manuscript and approved the manuscript prior to submission.

*   Funding Funding for this study and the work of all authors was supported, wholly or in part, by the Centre for Diet and Activity Research (CEDAR), a UKCRC Public Health Research Centre of Excellence (RES-590-28-0002). Funding from the British Heart Foundation, Department of Health, Economic and Social Research Council, Medical Research Council, and the Wellcome Trust, under the auspices of the UK Clinical Research Collaboration, is gratefully acknowledged. The work of KC, HEB, EW and EvS was supported by the Medical Research Council (MC_UU_12015/7). RL is funded by a Gates Cambridge Scholarship.

*   Competing interests KC reports receiving the following grants: Lead Applicant—a cluster randomised controlled trial to evaluate the effectiveness and cost-effectiveness of the GoActive programme to increase physical activity among 13-14 year-old adolescents. Project: 13/90/18 National Institute for Health Research Public Health Research Programme, September 2015—February 2019. Coapplicant—Opportunities within the school environment to shift the distribution of activity intensity in adolescents. Department of Health Policy Research Programme, December 2013—November 2016. KC is a Director of Ridgepoint Consulting Limited, an operational improvement consultancy.

*   Provenance and peer review Not commissioned; externally peer reviewed.

This is an open access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: [http://creativecommons.org/licenses/by/4.0/](http://creativecommons.org/licenses/by/4.0/) 

## References 

1.   Wareham NJ , van Sluijs EM , Ekelund U . Physical activity and obesity prevention: a review of the current evidence. Proc Nutr Soc 2005;64:229–47.[doi:10.1079/PNS2005423](http://dx.doi.org/10.1079/PNS2005423) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1079/PNS2005423&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=15960868&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 
    
    [Web of Science](http://bjsm.bmj.com/lookup/external-ref?access_num=000229743100011&link_type=ISI) 

2.   Ekelund U , Luan J , Sherar LB , et al . International Children’s Accelerometry Database (ICAD) Collaborators. Moderate to vigorous physical activity and sedentary time and cardiometabolic risk factors in children and adolescents. JAMA 2012;307:704–12.[doi:10.1001/jama.2012.156](http://dx.doi.org/10.1001/jama.2012.156) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1001/jama.2012.156&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=22337681&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 
    
    [Web of Science](http://bjsm.bmj.com/lookup/external-ref?access_num=000300278500027&link_type=ISI) 

3.  1.  
    2.   Smith A , 
    3.   Biddle S  
    
    In:  Smith A ,  Biddle S  , eds. Youth physical activity and sedentary behavior. Champaign, IL: Human Kinetics, 2008.
    
    

4.   Singh A , Uijtdewilligen L , Twisk JW , et al . Physical activity and performance at school: a systematic review of the literature including a methodological quality assessment. Arch Pediatr Adolesc Med 2012;166:49–55.[doi:10.1001/archpediatrics.2011.716](http://dx.doi.org/10.1001/archpediatrics.2011.716) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1001/archpediatrics.2011.716&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=22213750&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 
    
    [Web of Science](http://bjsm.bmj.com/lookup/external-ref?access_num=000298736500006&link_type=ISI) 

5.   Collings PJ , Wijndaele K , Corder K , et al . Magnitude and determinants of change in objectively-measured physical activity, sedentary time and sleep duration from ages 15 to 17.5y in UK adolescents: the ROOTS study. Int J Behav Nutr Phys Act 2015;12:61.[doi:10.1186/s12966-015-0222-4](http://dx.doi.org/10.1186/s12966-015-0222-4) 
    
    

6.   Ussher MH , Owen CG , Cook DG , et al . The relationship between physical activity, sedentary behaviour and psychological wellbeing among adolescents. Soc Psychiatry Psychiatr Epidemiol 2007;42:851–6.[doi:10.1007/s00127-007-0232-x](http://dx.doi.org/10.1007/s00127-007-0232-x) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1007/s00127-007-0232-x&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=17639309&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 
    
    [Web of Science](http://bjsm.bmj.com/lookup/external-ref?access_num=000249632500012&link_type=ISI) 

7.   Dumith SC , Gigante DP , Domingues MR , et al . Physical activity change during adolescence: a systematic review and a pooled analysis. Int J Epidemiol 2011;40:685–98.[doi:10.1093/ije/dyq272](http://dx.doi.org/10.1093/ije/dyq272) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1093/ije/dyq272&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=21245072&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 
    
    [Web of Science](http://bjsm.bmj.com/lookup/external-ref?access_num=000293618300018&link_type=ISI) 

8.   Telama R , Yang X , Viikari J , et al . Physical activity from childhood to adulthood: a 21-year tracking study. Am J Prev Med 2005;28:267–73.[doi:10.1016/j.amepre.2004.12.003](http://dx.doi.org/10.1016/j.amepre.2004.12.003) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1016/j.amepre.2004.12.003&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=15766614&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 
    
    [Web of Science](http://bjsm.bmj.com/lookup/external-ref?access_num=000227789900004&link_type=ISI) 

9.   Khaw KT , Wareham N , Bingham S , et al . Combined impact of health behaviours and mortality in men and women: the EPIC-Norfolk prospective population study. PLoS Med 2008;5:e12.[doi:10.1371/journal.pmed.0050012](http://dx.doi.org/10.1371/journal.pmed.0050012) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1371/journal.pmed.0050012&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=18184033&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 

10.  Horn DB , O’Neill JR , Pfeiffer KA , et al . Predictors of physical activity in the transition after high school among young women. J Phys Act Health 2008;5:275–85.[doi:10.1123/jpah.5.2.275](http://dx.doi.org/10.1123/jpah.5.2.275) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=18382036&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 

11.  Cluskey M , Grobe D . College weight gain and behavior transitions: male and female differences. J Am Diet Assoc 2009;109:325–9.[doi:10.1016/j.jada.2008.10.045](http://dx.doi.org/10.1016/j.jada.2008.10.045) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1016/j.jada.2008.10.045&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=19167962&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 

12.  Bell S , Lee C . Emerging adulthood and patterns of physical activity among young Australian women. Int J Behav Med 2005;12:227–35.[doi:10.1207/s15327558ijbm1204\_3](http://dx.doi.org/10.1207/s15327558ijbm1204_3) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1207/s15327558ijbm1204_3&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=16262541&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 

13.  Nelson MC , Neumark-Stzainer D , Hannan PJ , et al . Longitudinal and secular trends in physical activity and sedentary behavior during adolescence. Pediatrics 2006;118:e1627–e1634.[doi:10.1542/peds.2006-0926](http://dx.doi.org/10.1542/peds.2006-0926) 
    
    [Abstract/FREE Full Text](http://bjsm.bmj.com/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6MTA6InBlZGlhdHJpY3MiO3M6NToicmVzaWQiO3M6MTE6IjExOC82L2UxNjI3IjtzOjQ6ImF0b20iO3M6MjM6Ii9ianNwb3J0cy81My84LzQ5Ni5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

14.  Rindfuss RR . The young adult years: diversity, structural change, and fertility. Demography 1991;28:493–512.[doi:10.2307/2061419](http://dx.doi.org/10.2307/2061419) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.2307/2061419&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=1769399&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 
    
    [Web of Science](http://bjsm.bmj.com/lookup/external-ref?access_num=A1991GP37600001&link_type=ISI) 

15.  Maternal newborn . Child and adolescent health: adolescent development (Internet). [http://www.who.int/maternal\_child_adolescent/topics/adolescence/dev/en/](http://www.who.int/maternal_child_adolescent/topics/adolescence/dev/en/).
    
    

16.  Pearson N , Haycraft E , P Johnston J , et al . Sedentary behaviour across the primary-secondary school transition: A systematic review. Prev Med 2017;94:40–7.[doi:10.1016/j.ypmed.2016.11.010](http://dx.doi.org/10.1016/j.ypmed.2016.11.010) 
    
    

17.  Chan AW , Altman DG . Identifying outcome reporting bias in randomised trials on PubMed: review of publications and survey of authors. BMJ 2005;330:753.[doi:10.1136/bmj.38356.424606.8F](http://dx.doi.org/10.1136/bmj.38356.424606.8F) 
    
    [Abstract/FREE Full Text](http://bjsm.bmj.com/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6MzoiYm1qIjtzOjU6InJlc2lkIjtzOjEyOiIzMzAvNzQ5NC83NTMiO3M6NDoiYXRvbSI7czoyMzoiL2Jqc3BvcnRzLzUzLzgvNDk2LmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ==) 

18.  Brooke HL , Corder K , Atkin AJ , et al . A systematic literature review with meta-analyses of within- and between-day differences in objectively measured physical activity in school-aged children. Sports Med 2014;44:1427–38.[doi:10.1007/s40279-014-0215-5](http://dx.doi.org/10.1007/s40279-014-0215-5) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1007/s40279-014-0215-5&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=24981243&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 
    
    [Web of Science](http://bjsm.bmj.com/lookup/external-ref?access_num=000344974200008&link_type=ISI) 

19.  Selph SS , Ginsburg AD , Chou R . Impact of contacting study authors to obtain additional data for systematic reviews: diagnostic accuracy studies for hepatic fibrosis. Syst Rev 2014;3:107.[doi:10.1186/2046-4053-3-107](http://dx.doi.org/10.1186/2046-4053-3-107) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1186/2046-4053-3-107&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=25239493&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 

20.  Bielemann RM , Ramires VV , Gigante DP , et al . Longitudinal and cross-sectional associations of physical activity with triglyceride and HDLc levels in young male adults. J Phys Act Health 2014;11:784–9.[doi:10.1123/jpah.2012-0175](http://dx.doi.org/10.1123/jpah.2012-0175) 
    
    

21.  Jung ME , Bray SR , Martin Ginis KA . Behavior change and the freshman 15: tracking physical activity and dietary patterns in 1st-year university women. J Am Coll Health 2008;56:523–30.[doi:10.3200/JACH.56.5.523-530](http://dx.doi.org/10.3200/JACH.56.5.523-530) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.3200/JACH.56.5.523-530&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=18400664&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 
    
    [Web of Science](http://bjsm.bmj.com/lookup/external-ref?access_num=000259692200007&link_type=ISI) 

22.  Kimm SY , Glynn NW , Kriska AM , et al . Decline in physical activity in black girls and white girls during adolescence. N Engl J Med 2002;347:709–15.[doi:10.1056/NEJMoa003277](http://dx.doi.org/10.1056/NEJMoa003277) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1056/NEJMoa003277&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=12213941&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 
    
    [Web of Science](http://bjsm.bmj.com/lookup/external-ref?access_num=000177765800002&link_type=ISI) 

23.  Lappe JM , Watson P , Gilsanz V , et al . The longitudinal effects of physical activity and dietary calcium on bone mass accrual across stages of pubertal development. J Bone Miner Res 2015;30:156–64.[doi:10.1002/jbmr.2319](http://dx.doi.org/10.1002/jbmr.2319) 
    
    

24.  Martinez-Gomez D , Mielke GI , Menezes AM , et al . Active commuting throughout adolescence and central fatness before adulthood: prospective birth cohort study. PLoS One 2014;9:e96634.[doi:10.1371/journal.pone.0096634](http://dx.doi.org/10.1371/journal.pone.0096634) 
    
    

25.  Palakshappa D , Virudachalam S , Oreskovic NM , et al . Adolescent physical education class participation as a predictor for adult physical activity. Child Obes 2015;11:616–23.[doi:10.1089/chi.2015.0024](http://dx.doi.org/10.1089/chi.2015.0024) 
    
    

26.  Sagatun A , Kolle E , Anderssen SA , et al . Three-year follow-up of physical activity in Norwegian youth from two ethnic groups: associations with socio-demographic factors. BMC Public Health 2008;8:419.[doi:10.1186/1471-2458-8-419](http://dx.doi.org/10.1186/1471-2458-8-419) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1186/1471-2458-8-419&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=19102770&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 

27.  Zarrett N , Bell BA . The effects of out-of-school time on changes in youth risk of obesity across the adolescent years. J Adolesc 2014;37:85–96.[doi:10.1016/j.adolescence.2013.11.001](http://dx.doi.org/10.1016/j.adolescence.2013.11.001) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1016/j.adolescence.2013.11.001&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=24331308&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 
    
    [Web of Science](http://bjsm.bmj.com/lookup/external-ref?access_num=000330012800010&link_type=ISI) 

28.  Kwon S , Janz KF , Letuchy EM , et al . Developmental trajectories of physical activity, sports, and television viewing during childhood to young adulthood: Iowa bone development study. JAMA Pediatr 2015;169:666–72.[doi:10.1001/jamapediatrics.2015.0327](http://dx.doi.org/10.1001/jamapediatrics.2015.0327) 
    
    

29.  Armijo-Olivo S , Stiles CR , Hagen NA , et al . Assessment of study quality for systematic reviews: a comparison of the Cochrane Collaboration risk of Bias tool and the Effective Public Health Practice Project Quality Assessment Tool: methodological research. J Eval Clin Pract 2012;18:12–18.[doi:10.1111/j.1365-2753.2010.01516.x](http://dx.doi.org/10.1111/j.1365-2753.2010.01516.x) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1111/j.1365-2753.2010.01516.x&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=20698919&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 

30.  Brown HE , Atkin AJ , Panter J , et al . Family-based interventions to increase physical activity in children: a systematic review, meta-analysis and realist synthesis. Obes Rev 2016;17:345–60.[doi:10.1111/obr.12362](http://dx.doi.org/10.1111/obr.12362) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=26756281&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 

31.  Ogilvie D , Foster CE , Rothnie H , et al . Scottish physical activity research C: Interventions to promote walking: systematic review. BMJ 2007;334:1204.
    
    [Abstract/FREE Full Text](http://bjsm.bmj.com/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6MzoiYm1qIjtzOjU6InJlc2lkIjtzOjEzOiIzMzQvNzYwNS8xMjA0IjtzOjQ6ImF0b20iO3M6MjM6Ii9ianNwb3J0cy81My84LzQ5Ni5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

32.  Troiano RP , Berrigan D , Dodd KW , et al . Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc 2008;40:181–8.[doi:10.1249/mss.0b013e31815a51b3](http://dx.doi.org/10.1249/mss.0b013e31815a51b3) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1249/mss.0b013e31815a51b3&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=18091006&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 
    
    [Web of Science](http://bjsm.bmj.com/lookup/external-ref?access_num=000251870900025&link_type=ISI) 

33.  Adachi PJC , Willoughby T . From the couch to the sports field: the longitudinal associations between sports video game play, self-esteem, and involvement in sports. Psychol Pop Media Cult 2015;4:329–41.[doi:10.1037/ppm0000042](http://dx.doi.org/10.1037/ppm0000042) 
    
    

34.  Nordström A , Neovius MG , Rössner S , et al . Postpubertal development of total and abdominal percentage body fat: an 8-year longitudinal study. Obesity 2008;16:2342–7.[doi:10.1038/oby.2008.349](http://dx.doi.org/10.1038/oby.2008.349) 
    
    

35.  Raustorp A , Ekroth Y . Tracking of pedometer-determined physical activity: a 10-year follow-up study from adolescence to adulthood in Sweden. J Phys Act Health 2013;10:1186–92.[doi:10.1123/jpah.10.8.1186](http://dx.doi.org/10.1123/jpah.10.8.1186) 
    
    

36.  Wan X , Wang W , Liu J , et al . Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range. BMC Med Res Methodol 2014;14:135.[doi:10.1186/1471-2288-14-135](http://dx.doi.org/10.1186/1471-2288-14-135) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1186/1471-2288-14-135&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=25524443&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 

37.  Boreham C , Robson PJ , Gallagher AM , et al . Tracking of physical activity, fitness, body composition and diet from adolescence to young adulthood: the young Hearts Project, Northern Ireland. Int J Behav Nutr Phys Act 2004;1:14.[doi:10.1186/1479-5868-1-14](http://dx.doi.org/10.1186/1479-5868-1-14) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1186/1479-5868-1-14&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=15462676&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 

38.  de Souza MC , Eisenmann JC , e Santos DV , et al . Modeling the dynamics of BMI changes during adolescence. The Oporto Growth, Health and Performance Study. Int J Obes 2015;39:1063–9.[doi:10.1038/ijo.2015.60](http://dx.doi.org/10.1038/ijo.2015.60) 
    
    

39.  Freitas D , Beunen G , Maia J , et al . Tracking of fatness during childhood, adolescence and young adulthood: a 7-year follow-up study in Madeira Island, Portugal. Ann Hum Biol 2012;39:59–67.[doi:10.3109/03014460.2011.638322](http://dx.doi.org/10.3109/03014460.2011.638322) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.3109/03014460.2011.638322&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=22148930&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 

40.  Magarey AM , Boulton TJ , Chatterton BE , et al . Familial and environmental influences on bone growth from 11-17 years. Acta Paediatr 1999;88:1204–10.[doi:10.1111/j.1651-2227.1999.tb01018.x](http://dx.doi.org/10.1111/j.1651-2227.1999.tb01018.x) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=10591420&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 

41.  Porkka KV , Raitakari OT , Leino A , et al . Trends in serum lipid levels during 1980-1992 in children and young adults. the Cardiovascular risk in Young Finns Study. Am J Epidemiol 1997;146:64–77.[doi:10.1093/oxfordjournals.aje.a009192](http://dx.doi.org/10.1093/oxfordjournals.aje.a009192) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1093/oxfordjournals.aje.a009192&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=9215224&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 
    
    [Web of Science](http://bjsm.bmj.com/lookup/external-ref?access_num=A1997XJ18000007&link_type=ISI) 

42.  Rauner A , Jekauc D , Mess F , et al . Woll A: tracking physical activity in different settings from late childhood to early adulthood in Germany: the MoMo longitudinal study Health behavior, health promotion and society. BMC Public Health 2015;15:391.
    
    

43.  Telama R , Yang X , Leskinen E , et al . Tracking of physical activity from early childhood through youth into adulthood. Med Sci Sports Exerc 2014;46:955–62.[doi:10.1249/MSS.0000000000000181](http://dx.doi.org/10.1249/MSS.0000000000000181) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1249/MSS.0000000000000181&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=24121247&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 

44.  Benítez-Porres J , Alvero-Cruz JR , Carrillo de Albornoz M , et al . The influence of 2-Year changes in physical activity, maturation, and Nutrition on Adiposity in Adolescent Youth. PLoS One 2016;11:e0162395.[doi:10.1371/journal.pone.0162395](http://dx.doi.org/10.1371/journal.pone.0162395) 
    
    

45.  Gunnell KE , Flament MF , Buchholz A , et al . Examining the bidirectional relationship between physical activity, screen time, and symptoms of anxiety and depression over time during adolescence. Prev Med 2016;88:147–52.[doi:10.1016/j.ypmed.2016.04.002](http://dx.doi.org/10.1016/j.ypmed.2016.04.002) 
    
    

46.  Huppertz C , Bartels M , de Geus EJ , et al . The effects of parental education on exercise behavior in childhood and youth: a study in Dutch and Finnish twins. Scand J Med Sci Sports 2016.[doi:10.1111/sms.12727](http://dx.doi.org/10.1111/sms.12727) 
    
    

47.  Bagur-Calafat C , Farrerons-Minguella J , Girabent-Farrés M , et al . The impact of high level basketball competition, calcium intake, menses, and hormone levels in adolescent bone density: a three-year follow-up. J Sports Med Phys Fitness 2015;55:58–67.
    
    

48.  Baxter-Jones AD , Eisenmann JC , Mirwald RL , et al . The influence of physical activity on lean mass accrual during adolescence: a longitudinal analysis. J Appl Physiol 2008;105:734–41.[doi:10.1152/japplphysiol.00869.2007](http://dx.doi.org/10.1152/japplphysiol.00869.2007) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1152/japplphysiol.00869.2007&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=18467546&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 

49.  Lantz H , Bratteby LE , Fors H , et al . Body composition in a cohort of Swedish adolescents aged 15, 17 and 20.5 years. Acta Paediatr 2008;97:1691–7.[doi:10.1111/j.1651-2227.2008.01035.x](http://dx.doi.org/10.1111/j.1651-2227.2008.01035.x) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1111/j.1651-2227.2008.01035.x&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=18795912&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 
    
    [Web of Science](http://bjsm.bmj.com/lookup/external-ref?access_num=000260731300022&link_type=ISI) 

50.  Deheeger M , Bellisle F , Rolland-Cachera MF . The French longitudinal study of growth and nutrition: data in adolescent males and females. J Hum Nutr Diet 2002;15:429–38.[doi:10.1046/j.1365-277X.2002.00396.x](http://dx.doi.org/10.1046/j.1365-277X.2002.00396.x) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=12460151&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 

51.  Pinto BM , Cherico NP , Szymanski L , et al . Longitudinal changes in college students' exercise participation. J Am Coll Health 1998;47:23–7.[doi:10.1080/07448489809595615](http://dx.doi.org/10.1080/07448489809595615) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1080/07448489809595615&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=9693476&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 
    
    [Web of Science](http://bjsm.bmj.com/lookup/external-ref?access_num=000075033600015&link_type=ISI) 

52.  Shi HJ , Nakamura K , Kizuki M , et al . Extracurricular sports activity around growth spurt and improved tibial cortical bone properties in late adolescence. Acta Paediatr 2006;95:1608–13.[doi:10.1080/08035250600690609](http://dx.doi.org/10.1080/08035250600690609) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=17129970&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 

53.  Wichstrøm L , von Soest T , Kvalem IL . Predictors of growth and decline in leisure time physical activity from adolescence to adulthood. Health Psychol 2013;32:775–84.[doi:10.1037/a0029465](http://dx.doi.org/10.1037/a0029465) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1037/a0029465&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=22924445&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 

54.  Graham DJ , Sirard JR , Neumark-Sztainer D . Adolescents' attitudes toward sports, exercise, and fitness predict physical activity 5 and 10 years later. Prev Med 2011;52:130–2.[doi:10.1016/j.ypmed.2010.11.013](http://dx.doi.org/10.1016/j.ypmed.2010.11.013) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1016/j.ypmed.2010.11.013&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=21130803&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 

55.  Eime RM , Harvey JT , Sawyer NA , et al . Changes in sport and physical activity participation for adolescent females: a longitudinal study. BMC Public Health 2016;16:533.[doi:10.1186/s12889-016-3203-x](http://dx.doi.org/10.1186/s12889-016-3203-x) 
    
    

56.  Kayihan G . Effect of physical activity on body composition changes in young adults. A four-year longitudinal study. Medicina dello Sport 2014;67:423–35.
    
    

57.  Audrain-McGovern J , Rodriguez D , Rodgers K , et al . Longitudinal variation in adolescent physical activity patterns and the emergence of tobacco use. J Pediatr Psychol 2012;37:622–33.[doi:10.1093/jpepsy/jss043](http://dx.doi.org/10.1093/jpepsy/jss043) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1093/jpepsy/jss043&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=22408052&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 
    
    [Web of Science](http://bjsm.bmj.com/lookup/external-ref?access_num=000305827600004&link_type=ISI) 

58.  Corder K , Sharp SJ , Atkin AJ , et al . Change in objectively measured physical activity during the transition to adolescence. Br J Sports Med 2015;49:730–6.[doi:10.1136/bjsports-2013-093190](http://dx.doi.org/10.1136/bjsports-2013-093190) 
    
    [Abstract/FREE Full Text](http://bjsm.bmj.com/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6ODoiYmpzcG9ydHMiO3M6NToicmVzaWQiO3M6OToiNDkvMTEvNzMwIjtzOjQ6ImF0b20iO3M6MjM6Ii9ianNwb3J0cy81My84LzQ5Ni5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

59.  Stroup DF , Berlin JA , Morton SC , et al . Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis of observational studies in Epidemiology (MOOSE) group. JAMA 2000;283:2008–12.
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1001/jama.283.15.2008&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=10789670&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 
    
    [Web of Science](http://bjsm.bmj.com/lookup/external-ref?access_num=000086436600037&link_type=ISI) 

60.  Moher D , Liberati A , Tetzlaff J , et al . PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. J Clin Epidemiol 2009;62:1006–12.[doi:10.1016/j.jclinepi.2009.06.005](http://dx.doi.org/10.1016/j.jclinepi.2009.06.005) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1016/j.jclinepi.2009.06.005&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=19631508&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 
    
    [Web of Science](http://bjsm.bmj.com/lookup/external-ref?access_num=000270250500003&link_type=ISI) 

61.  Ortega FB , Konstabel K , Pasquali E , et al . Objectively measured physical activity and sedentary time during childhood, adolescence and young adulthood: a cohort study. PLoS One 2013;8:e60871.[doi:10.1371/journal.pone.0060871](http://dx.doi.org/10.1371/journal.pone.0060871) 
    
    

62.  Fortier MD , Katzmarzyk PT , Malina RM , et al . Seven-year stability of physical activity and musculoskeletal fitness in the Canadian population. Med Sci Sports Exerc 2001;33:1905–11.[doi:10.1097/00005768-200111000-00016](http://dx.doi.org/10.1097/00005768-200111000-00016) 
    
    [CrossRef](http://bjsm.bmj.com/lookup/external-ref?access_num=10.1097/00005768-200111000-00016&link_type=DOI) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=11689742&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 
    
    [Web of Science](http://bjsm.bmj.com/lookup/external-ref?access_num=000171960200016&link_type=ISI) 

63.  Nelson MC , Story M , Larson NI , et al . Emerging adulthood and college-aged youth: an overlooked age for weight-related behavior change. Obesity 2008;16:2205–11.[doi:10.1038/oby.2008.365](http://dx.doi.org/10.1038/oby.2008.365) 
    
    

64.  Cooper AR , Goodman A , Page AS , et al . Objectively measured physical activity and sedentary time in youth: the international children’s accelerometry database (ICAD). Int J Behav Nutr Phys Act 2015;12:113.[doi:10.1186/s12966-015-0274-5](http://dx.doi.org/10.1186/s12966-015-0274-5) 
    
    [PubMed](http://bjsm.bmj.com/lookup/external-ref?access_num=26377803&link_type=MED&atom=%2Fbjsports%2F53%2F8%2F496.atom) 

65.  Corder K , van Sluijs EM , Wright A , et al . Is it possible to assess free-living physical activity and energy expenditure in young people by self-report? Am J Clin Nutr 2009;89:862–70.[doi:10.3945/ajcn.2008.26739](http://dx.doi.org/10.3945/ajcn.2008.26739) 
    
    [Abstract/FREE Full Text](http://bjsm.bmj.com/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NDoiYWpjbiI7czo1OiJyZXNpZCI7czo4OiI4OS8zLzg2MiI7czo0OiJhdG9tIjtzOjIzOiIvYmpzcG9ydHMvNTMvOC80OTYuYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9)