Article Text

Download PDFPDF

Secular trends in aerobic fitness performance in 13–18-year-old adolescents from 1976 to 2001
  1. P R T Huotari1,
  2. H Nupponen2,
  3. L Laakso1,
  4. U M Kujala3
  1. 1Department of Sport Sciences, University of Jyväskylä, Jyväskylä, Finland
  2. 2Department of Teacher Education in Rauma, University of Turku, Turku, Finland
  3. 3Department of Health Sciences, University of Jyväskylä, Jyväskylä, Finland
  1. Correspondence to P R T Huotari, Department in Sport Sciences, University of Jyväskylä, PL 35, 40014 University of Jyväskylä, Finland; pertti.huotari{at}


Objective To analyse the secular trends in aerobic fitness performance and some of its determinants (body mass index (BMI) and leisure time physical activity (LTPA)) in adolescents.

Design Cross-sectional population-based studies in 1976 and 2001 in Finland. A stratified random sample of Finnish 13–18-year-old adolescents was studied in 1976 (n = 717; 384 boys and 333 girls) and in 2001 (n = 558; 305 boys and 253 girls). The main outcome measure was aerobic fitness, estimated with a 2000 m (for boys) and 1500 m (for girls) running test; the weight and height of participants were also measured. Self-reported weekly frequency of LTPA of at least 30 min duration and regularity of participation in organised sport were obtained by questionnaire. Identical methods were used in 1976 and 2001.

Results Running time was longer in 2001 compared to 1976 in boys (56 s, Cohen d = 0.46, medium effect size; p<0.001) and girls (29 s, d = 0.32, small; p<0.01). BMI and participation in LTPA explained more of the variance in aerobic fitness in 2001 than in 1976 in boys and girls.

Conclusions Aerobic fitness of school-aged children deteriorated between the measurement points. BMI and organised LTPA were better associated with aerobic fitness in 2001 than in 1976. An increase in overweight and obesity was associated with the decrease in aerobic fitness. Although the importance of organised LTPA to fitness increased, it is possible that the decrease in overall physical activity between 1976 and 2001 contributed most to the decrease in the level of aerobic fitness.

Statistics from

Low aerobic fitness has been shown to be a strong risk factor for cardiovascular diseases, type 2 diabetes and mortality in adults.1,,4 Good endurance running capacity in particular seems to predict low risk for metabolic syndrome and mortality.3 5

There is evidence that the level of aerobic fitness has declined among adolescents, although long-term follow-ups are rare.6,,12 Aerobic fitness has declined among Finnish military service conscripts during recent decades,13 but information on long-term changes in aerobic fitness and its determinants among school aged adolescents is lacking.

Overweight and obesity among children and adolescents has increased during the last few decades worldwide.6 7 14 It has also been shown that overweight has an inverse association with aerobic fitness and that poor fitness is already related to impaired health in adolescence.10 15,,17 Although overweight is increasing, some studies have reported an increase in participation in leisure time physical activity (LTPA).18,,22 However, it seems that many adolescents do not participate enough in LTPA and their overall lifestyles may have become more sedentary.

In Finland, in the 1970s the total TV viewing time for adolescents was on average 1.30 h/day while in the past decade screen time has been estimated to be as high as 5–6 h/day.23 Furthermore, the number of cars per capita increased from 0.2 to 0.4 between 1976 and 2001, and over 60% of Finnish homes had a computer and almost all Finns aged 15–39 used a mobile telephone in 2000.23 Technological change over the past 25 years has been rapid, and this is probably the main reason for the more sedentary lifestyle found among adolescents24 25 that, in turn, may have an effect on physical fitness.

The main aims of this study were to analyse secular trends in aerobic fitness performance (later used only aerobic fitness) and study changes in some of the determinants (body mass index (BMI), LTPA, sports club participation) of aerobic fitness in 13–18-year-old adolescents at two measurement points, 1976 and 2001.



In 1976 and in 2001, specially trained teams carried out measurements of fitness among 9–21-year-old adolescents and young adults in Finnish schools. The target group consisted of 13–18-year-old boys and girls who were invited to take part in a long-distance running test in 1976 and in 2001. Of these, 717 adolescents (384 boys and 333 girls) took part in the study in 1976 and 558 pupils (305 boys and 253 girls) in 2001 (table 1). As the anthropometric measurements and questionnaire study were not implemented for all the participants, complete information (ie, the long distance running test, weight, height and the questionnaire data) was only available for a total of 913 participants. Of these, 377 adolescents (211 boys and 166 girls) participated in 1976 and 536 (292 boys and 244 girls) in 2001. Before the running test participants were allowed to refuse to participate in the tests; however, only participants with an illness or injury refused (non-participation due to diseases or injuries <5%). This study was approved by the ethics committee of the Central Finland Health Region.

Table 1

Characteristics of participants in 1976 and 2001

In 1976 data were collected from 56 and in 2001 from 17 comprehensive and high schools. The 1976 sample was a four-phased stratified random sample drawn from schools representing different provinces and types of municipalities (urban and rural). In the first phase a total of 20 towns and communes from 4 geographical areas (west, east, middle and north) of Finland were randomly selected by drawing lots. The random sample of 56 schools, matched for size, was taken from these towns and communes. School classes were selected on the basis of the classes that had a physical education lesson on the assessment day. In 2001 the sample was a three-phased stratified random sample. In the first phase, 9 communes were selected randomly from the 20 communes that participated in 1976; 3 of these were rural communes (<10 000 residents). In the second phase 17 schools were selected from the schools that participated in 1976. In the third phase the school classes were selected as in 1976.

Assessment of aerobic fitness

In spring 1976 and spring 2001, specially trained teams carried out measurements of aerobic fitness in the selected schools using identical methodology.26 Aerobic fitness was estimated with a 2000 m running test for boys and a 1500 m running test for girls. The tests conformed to the international standards for assessments and have been described in detail previously.27 The tests have good reliability and validity, and in previous studies the intratester reliability of long distance running has varied between 0.65 and 0.9428 and validity coefficients between 0.65 and 0.88.29

Assessment of BMI and physical activity

The weight and height of the participants were measured by school health nurses or the measuring teams with the calibrated steelyard and gauge used in school healthcare in Finland. Height and weight were measured to the nearest 1 cm and 0.5 kg and BMI (kg/m2) was calculated. BMI was used as a continuous variable in most of the statistical analyses. In addition, participants were classified into three groups according to the international BMI cut-off recommendation to compare aerobic fitness between normal weight and overweight or obese pupils.30

Physical activity and participation in sports were recorded using a self-report questionnaire that was administered individually in connection with the fitness test. The questionnaire was identical at both measurement points. The questionnaire contained items on frequency per week of physical activity outside school hours (not at all, less than once a month, once a month, 2–3 times a month, at most once a week, 2–6 times a week, every day of at least 30 min duration) and participation in sports club training (not at all, sometimes, regularly).

In the analytical phase the responses were coded from 1 to 3. The LTPA categories were 1 = less than twice a week, 2 = 2–6 times a week and 3 = every day. In previous studies internal consistency coefficients of these items as indicators of reliability have varied from 0.44 to 0.76. The sports club participation variable showed higher interage stability than other variables concerning physical activity.31

Statistical analyses

The running test results are expressed as mean (SD). In addition, an index of aerobic fitness was calculated by transferring the running test results to age-standardised z points separately for boys and girls. The standardised z points were used in the statistical analyses. Mann–Whitney U test, 95% confidence intervals (95% CI) and analysis of covariance (ANCOVA; adjusted to age, BMI and LTPA) were used to compare differences in aerobic fitness between the measurement points in boys and girls. The meaningfulness of differences was examined by calculating the effect size Cohen d using means and standard deviations.32 The association between age, LTPA, BMI and aerobic fitness was examined by linear regression at both measurement points. ANCOVA was applied in testing the differences in aerobic fitness between the physical activity and BMI categories. In addition, ANCOVA was used in comparing aerobic fitness adjusted to BMI in inactive and overweight adolescents between 1976 and 2001. Because of the different aerobic fitness tests used, boys and girls were always analysed separately. All the statistical tests were performed by SPSS statistical software, V.15.0 (SPSS, Chicago, Illinois, USA).


Secular trend in aerobic fitness from 1976 to 2001

The mean 2000 m running test time for boys in 1976 was 559 s (95% confidence interval (CI), 548 to 570 s) and in 2001 615 s (95% CI, 600 to 630 s). The mean 1500 m running test time for girls in 1976 was 494 s (95% CI, 486 to 503 s) and in 2001 523 s (95% CI, 510 to 535 s). The mean difference between 1976 and 2001 among boys was 56 s (increase 10%, p<0.001, Cohen d = 0.46, medium effect size) and among girls 29 s (increase 6%; p<0.01, d = 0.32, small effect size) (fig 1). The secular trend in distributions of running time is shown in figure 2.

Figure 1

Running time in boys and girls in 1976 and in 2001. Age-adjusted p values for difference between 1976 and 2001 are shown.

Figure 2

Distributions of running time in boys and girls in 1976 and 2001.

Secular trend in determinants of aerobic fitness from 1976 to 2001

Between 1976 and 2001 the average body weight of the boys increased by 4% (p = 0.019) while average body height increased by only 0.5%. In girls average body weight increased by 3% (p = 0.049) but average body height remained the same (table 1). BMI increased by 4% in girls (p = 0.003) and in boys (p = 0.002) over the 25-year period. The proportion of boys who participated in LTPA < twice a week was 30.3% in 1976 and 13.0% in 2001 (table 1). The percentages for girls were 31.9% in 1976 and 15.2% in 2001. Of the boys, 19.7% participated in sports club training in 1976 and 43.1% in 2001. In girls the percentages of regular sports club participation were 10.4% and 26.2%.

Table 2 shows the results of the age-adjusted linear regression analysis for the individual determinants of aerobic fitness and the results of the linear regression analysis in which age, BMI, LTPA and sports club participation were included in the model as explanatory variables. Among boys the cumulative R2 of this model was 0.08 in 1976 and 0.23 in 2001. For girls the R2 of the model was 0.08 in 1976 and 0.34 in 2001.

Table 2

Determinants of aerobic fitness index for boys and girls in 1976 and 2001

Figure 3 shows that aerobic fitness was lower in 2001 than in 1976 in all the LTPA groups among boys. The standardised difference (d) varied from 0.37 to 0.60 (small/medium) over the time. Among girls, differences were found between the two lowest LTPA groups, with a large effect size (d = 0.73) reported for differences between the lowest LTPA groups; however, those reporting daily physical activity had similar fitness levels in 1976 and 2001. According to the BMI international cut-offs 14 (6%) of the boys were overweight or obese in 1976 and 37 (12%) in 2001. Among the girls 7 (4%) were overweight or obese in 1976 and 30 (12%) in 2001. Obese or overweight participants had lower aerobic fitness in 1976 and 2001 among boys (in 1976 p = 0.004 and in 2001 p<0.001) and among girls in 2001 (p<0.001). After adjustment for age and BMI, no statistically significant difference emerged in the aerobic fitness of obese/overweight boys between 1976 and 2001 while overweight/obese girls showed lower aerobic fitness in 2001 than in 1976 (p = 0.036).

Figure 3

Running time by physical activity category in 1976 and 2001 in boys (A) and girls (B). Age-adjusted p values for difference between 1976 and 2001 are shown.


The main findings of this study showed that, compared with the 1976 sample, boys and girls in 2001 had lower aerobic fitness and higher BMI. Secondly, BMI and LTPA, in particular participation in organised sports, explained the higher percentage of aerobic fitness in 2001 than in 1976.

In this study aerobic fitness was expressed as the absolute time taken in a long-distance running test (s) and the corresponding standardised z value. Mean running time increased among boys by 56 s (10%) and among girls by 29 s (6%) from 1976 to 2001. The difference in the mean distance-running speed among boys was 0.33 m/s and 0.17 m/s among girls. This indicates that on average, a boy in 2001 would finish about 180 m behind the average 1976 boy over the 2000 m distance. In girls the corresponding difference in the 1500 m run was 83 m. In addition, as there was a difference in average age between the two samples of 6 months in boys and 4.8 months in girls, we used age, BMI and LTPA as covariates for the examination of secular trends in aerobic fitness. After this adjustment, a significant secular decline was found among boys (p = 0.011) and girls (p = 0.017). The decline in aerobic fitness is in accordance with the findings of a number of shorter-term studies from other countries on secular trends in aerobic fitness.8 6 Tomkinson and colleagues investigated 20-m shuttle run performance in children aged 6–19 years from 11 developed countries during the period 1981–2000 and found a decrease in aerobic performance of 0.43% per year.8 In our study the time taken to run increased on average from 1976 to 2001 by 0.40% per year in boys and 0.24% in girls. However, it is difficult to know the exact time-related patterns of change. The change over the last 10 years could have been more rapid than in the first 10 years.13 Furthermore, distributions showed decline for all percentile levels with the largest change in difference found for the poorest aerobic fitness level in boys and girls. Despite our results, a worldwide analysis of 109 studies from 37 countries showed that in the aerobic performance of children and adolescents the best performing children were from the Northern European countries, including Finland.33

The secular trend in BMI increased by 4% in boys and girls from 1976 to 2001 in our study samples. It is well known that low levels of physical fitness and obesity are associated with a higher prevalence of cardiovascular disease (CVD) risk factors, type 2 diabetes and higher CVD mortality,2 34 and that overweight and obesity among children and adolescents have increased during the last few decades worldwide.14 Overweight is a disadvantage in locomotion and an association between obesity and low aerobic fitness is likely to be seen most clearly in activities involving the carrying of body weight, such as running.15 16

In the regression analysis BMI in 2001 explained more of the variation in running performance than was explained by BMI in 1976. In 1976 BMI did not explain any of the boys' aerobic fitness whereas in 2001 BMI explained 12% of the variance of aerobic fitness. In girls the share of the variation explained by BMI had increased from 0.2% to 17%. The international BMI cut-offs were used to compare aerobic fitness between normal weight, overweight and obese pupils.30 The results indicated that the difference in aerobic fitness between overweight/obese and normal weight adolescents was significant in boys and girls. Furthermore, overweight and obese girls had poorer aerobic fitness in 2001 than in 1976, even after adjustment for BMI. In agreement with our study a similar trend in decreased running performance and increased BMI was found in 10–11-year-old children, who required more time to perform a 1.6 km run/walk test in 1997 than in 1985.35 There is also evidence of a similar association between overweight and low aerobic fitness in the 20 m shuttle run test.18 Common factors associated with the increase in body mass in today's society include the increased availability of energy-rich foods and sedentary behaviour (eg, increased screen time), along with decreased daily physical activity among adolescents.20 36

What is already known on this topic

There is some evidence that children's aerobic fitness levels are declining in many countries while, at the same time, the proportion of children classified as overweight or obese continues to increase.

What this study adds

Our long-term study shows a decline in endurance running ability among adolescents between 1976 and 2001. Increased body weight and differences in LTPA, in particular participation in organised sports, are playing an increasing role as determinants of aerobic fitness, whereas participation in other types of physical activity has decreased.

Previous studies suggest that physical activity may prevent obesity, and overall physical activity has been found to correlate negatively with the amount of body fat and positively with aerobic fitness.37 38 However, in general the evidence for the health-promoting effects of physical activity among children and adolescents is not strong.37 39 40 It is also noteworthy that moderate to vigorous physical activity decreased significantly between ages 9–15 years. According to recent longitudinal data 31% of US 15-year-old adolescents were sufficiently physically active (60 min per day) on weekdays and only 17% on weekends.22

Our study indicated that participation in LTPA had a positive association with aerobic fitness in 1976 and 2001. In fact, in 1976 it explained the variance of aerobic fitness best, even if the percentage explained by the whole model was rather low. After adjusting for age, sports club participation in addition to BMI explained the variance of girls' aerobic fitness in 2001. In agreement with our study the secular trend in frequency of LTPA among adolescents in Finland increased during the 30 years from 1977 to 2007 in both sexes, particularly in organised sport.21 This means that intensive LTPA more often than previously means participation in organised sport.21 Although frequency of participation in LTPA increased during the 30 years studied, overall physical activity has perhaps declined during this time. Traditionally, the share of spontaneous physical activity engaged in by adolescents has been larger in Finland than in other countries.41 Notably, aerobic fitness has also decreased among the group of boys who participate daily in LTPA. This may perhaps be explained by the difference in the type of sports as a higher proportion of physically active children participated in endurance type sports in 1976 than in 2001. Despite this, a major reason for the decline in physical fitness, in addition to the increase in obesity, may be the decrease in daily physical activity/increasingly sedentary lifestyle that has accompanied societal changes during the 25-year study period.20 42

This study has some limitations. Aerobic fitness was measured by a long-distance running test. This method may result in errors at the individual level, although it works rather well at the group level, for example in schools. However, in physical education in Finland physical fitness is measured over a long period, including the test of aerobic fitness by long-distance running, and previous research suggests that the running tests have acceptable validity and reliability.28 29 However, a better alternative for investigating aerobic fitness would have been a maximal oxygen uptake test in a laboratory with calibrated equipment.


In 13–18-year-old adolescents aerobic fitness declined during the 25 years from 1976 to 2001. Furthermore, we found a change in the explanatory power of the determinants of aerobic fitness: in 2001 BMI and LTPA explained a higher proportion of aerobic fitness than the same variables did in 1976. Adolescents who do not participate in LTPA and have high BMI are likely to show poorer aerobic fitness. Participation in organised sport, however, increased and it has continuing importance for Finnish adolescents' aerobic fitness.


View Abstract


  • Funding Funding was received from the Finnish Ministry of Education and Academy of Finland. The funding sources had no role in any aspects of the study or this paper.

  • Competing interests None declared.

  • Provenance and Peer review Not commissioned; externally peer reviewed.

  • Ethics approval Ethics approval was obtained from the ethics committee of the Central Finland Health Region.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.