21st September 2018, Volume 131 Number 1482

Helena M McAnally, Lesley M Reeves, Judith L Sligo, Robert J Hancox

Physical inactivity is rising in many countries and is now the fourth leading risk factor for global mortality.1 An inactive teenage lifestyle is associated with adult inactivity, which may have serious health implications. Large amounts of sedentary behaviour, such as watching too much television, have been linked to obesity and poor fitness in adult life,2,3 whereas children who achieve and maintain physical fitness are likely to maintain a healthy weight.4 There are other long-term benefits from physical activity observed in both cross-sectional and longitudinal studies: people who are physically active are less likely to develop osteoporosis, stroke, some forms of cancer and type 2 diabetes.5 An active lifestyle is also associated with better mental health and improved academic achievement.5

While the current global epidemic of obesity is well recognised, declines in physical fitness over time have also been recognised,6 but the relationship between these risk factors is not clear. We investigated generational changes in body weight and physical fitness in a cohort of 15 year-olds from 1986/7 and in their children when they were aged 15 (between 2007 and 2015). We anticipated that body weight and BMI values would be higher in the younger generation than their parents at the same age, and hypothesised that weight-adjusted fitness levels will have declined. However, we suspected that absolute aerobic capacity (V’O2max) may not have changed because increased body weight also means a greater workload when moving around and heavier people should have stronger musculature and greater absolute V’O2max for a similar level of physical activity.

Methods

Participants

The Dunedin Multidisciplinary Health and Development Study (hereafter referred to as the “Dunedin Study”) is a longitudinal investigation of health and behaviour in a cohort of individuals born in Dunedin in 1972/73.7 One thousand and thirty-seven children (91% of those eligible) participated in the first follow-up at age three and these individuals have been followed ever since with regular assessments. Data from the assessments at age 15 years in 1986/7 (n=976) and at age 38 years in 2010/12 (n=961) are used in this analysis.

Since 2007, the Next Generation Study has assessed the biological and non-biological children of the Dunedin Study members. The Next Generation Study invites participants and their primary caregiver to a one-day multidisciplinary assessment, which largely replicates the assessment that their parent went through in 1986/7 as part of the Dunedin Study.8 To be eligible for participation, teenagers must be the biological, step, foster or adopted children of a Dunedin Study member and be over 15 years and under 17 years of age. At the time of analysis, 235 of the Dunedin Study members (23% of the original 1,037 participants) had at least one child who consented to participate in the Next Generation Study (87% of those eligible). Of those Dunedin Study members with children in the Next Generation Study, 71 had two children, 17 had three children and one had four children or step-children in the study (343 Next Generation Study participants in total). At the time of their assessment, 277 were aged 15, and 66 were aged 16 years. Exercise test results were missing or excluded from 10 Next Generation participants due to pregnancy, health reasons or equipment malfunction. Hence 333 teenagers (157 girls and 176 boys; mean age 15.59, range 14.91–16.98; one participant was assessed at age 14 years and 11 months) had a valid exercise test while height and weight measures were available for 340 participants (Figure 1).

Figure 1: Participant numbers included in analyses from both the Next Generation and Dunedin Studies.

c 

Note numbers reflect maximum number of participants for an analysis: the number of participants in the analyses of each endpoint vary due to missing data; only 333 NG participants had valid exercise tests. 

When members of the Dunedin Study had a same-sex child in the Next Generation Study, we also used paired-t-tests to conduct mother-daughter and father-son comparisons. If Dunedin Study participants had multiple same-sex children in the Next Generation Study, the child who was closest to 15 years old at the time of assessment was chosen as the member of the dyad; there were 125 mother-daughter dyads and 110 father-son dyads in total. Only dyads that were biologically related were analysed (ie, excluding step- and adopted children).

Ethical approval for both studies was obtained from the Lower South Regional Ethics Committee (LRS/06/10/048/AM01). Participants provided written consent at the time of testing.

Measures

Body mass index (BMI) in kg/m2 was calculated from height and weight measurements using a stadiometer and calibrated scales. Maximal aerobic capacity was measured in both the Dunedin Study and the Next Generation Study via a sub-maximal exercise test on a cycle ergometer to estimate maximal oxygen consumption at peak exercise (V’O2max in L/min) using the modified Åstrand-Rhyming protocol.3,9 As maximum oxygen uptake should be higher in heavier people, V’O2max values are often adjusted for weight. However, this method may over-adjust in very obese people,10 thus V’O2max values are used both with and without adjustment for body weight.

Socio-economic status at age 15 was based on the income and education associated with their parents’ occupation using the Elley-Irving scale for the Dunedin Study and an updated version of the Elley-Irving scale, the NZ Socio-Economic Index, for the Next Generation. In both studies, the highest category for either parent (1=high, 6=low) was recorded.11,12 Ethnicity was self-identified at ages 26, 32 or 38 for the Dunedin Study and at age 15 for the Next Generation Study and recorded as Māori or non-Māori (nearly all non-Māori were New Zealand European, other ethnicities were uncommon). Maternal age at the participant’s birth was recorded for both studies.

Analyses

We compared weight, height, BMI and fitness at age 15 between Dunedin Study members and Next Generation participants in four ways. First we compared the mean values and 95% confidence intervals between the Dunedin and Next Generation studies for each sex. Differences were assessed using independent sample t-tests. Because variances were greater in the Next Generation Study, these t-tests did not assume equal variance. On average, the Next Generation participants were slightly older than the Dunedin Study participants were and therefore analyses were re-run on the sub-sample of the Next Generation cohort who were under the age of 16 at the time of their data collection.

Secondly we compared father-son and mother-daughter dyads, using paired t-tests (for these analyses Dunedin Study members with more than one same-sex child in the Next Generation Study were only paired with the child nearest to 15 years in age).

Because the Next Generation Study comprises children of Dunedin Study members who had children at younger ages (children born to older Study members had not yet reached age 15), the Next Generation participants are less representative of the general population than the Dunedin Study participants. A third analyses used propensity score matching to compare the Next Generation participants to 15 year-old Dunedin Study members.13 Next Generation participants were matched to their same-sex nearest neighbours in the Dunedin Study according to maternal age, childhood socioeconomic status and ethnicity using the pscore command in Stata. This was followed by estimating the effect of being a Next Generation participant (ie, the difference from the matched Dunedin Study member using the attnd command) with bootstrapped standard errors to estimate 95% confidence intervals.

The fourth analysis fitted mixed linear regression models for each of the outcomes, adjusting for socioeconomic status and age (15 or 16 years) with a fixed effect for being a Next Generation participant and a random effect for family, using robust standard errors to calculate confidence intervals. These models allow us to use all available data from both cohorts while adjusting for clustering within family members.

Finally, sex-specific correlations between mothers and daughters and fathers and sons, were done for Dunedin and Next Generation Studies’ weight, BMI and V’O2max values to determine whether levels of body weight and fitness correlated across the generations. Correlations between weight, BMI, and V’O2max from the Dunedin Study participants at ages 15 and 38 (data collected in 1986/7 and 2010/11) were also analysed, to determine whether body weight and fitness at age 15 weight are associated with weight and fitness in adulthood and therefore relevant to adult health. The V’O2max correlations were repeated with and without adjustment for body weight.

Analyses used Stata version 13 (College Station, TX).14 Two-sided p values <0.05 were considered statistically significant. Where non-statistically significant results with p<0.1 were observed, and felt to be important for guiding further research and/or interpreting other results, they are described as ‘non-statistically significant tendencies’. No formal adjustment for multiple comparisons was made in the analyses.

Results

A flow-chart of participants in the Next Generation study is shown in Figure 1. Of the 235 Dunedin study members with one or more Next Generation Study children, there were more mothers (n=134, 57%) than fathers (n=101, 43%). The median socioeconomic status of the Next Generation participants (n=343) was lower than the median socioeconomic status of the Dunedin Study members (n=875) had been at age 15 (3 vs 4, p<0.001).

Mean values for BMI, weight, height and both unadjusted and weight-adjusted V’O2max in each cohort are shown in Table 1. Independent t-tests were conducted to determine if the 16 year-old subsample of Next Generation participants differed from those under 16. The Next Generation 16 year-old boys tended to weigh more than those under 16 (74.6 vs 69.6kg, p=0.075) and had a non-statistically significant tendency towards higher BMI values (24.1 vs 22.4kg/m2 p=0.069), but there were no age-related differences in weight for girls or height or V’O2max in either sex.

Table 1: Mean weight, height and fitness for each Study by sex with 95% confidence intervals.

c
P values from two-tailed independent t-tests. Positive mean differences indicate higher values in the Next Generation Study and negative differences indicate higher values in the Dunedin Study. 

There were no significant differences in age 15 height, weight, BMI or unadjusted or weight-adjusted V’O2max between those Dunedin Study members who went on to become parents of a Next Generation Study member and those from the Dunedin Study who were not parents of a Next Generation Study member at the time of analysis (all p values >0.130).

Differences in weight, height and fitness between the cohorts

Mean BMI, weight and height for both sexes in the Next Generation participants were greater than in the Dunedin Study participants (Table 1). Both the unadjusted and weight-adjusted V’O2max values for the Next Generation girls were lower than the Dunedin Study girls. By contrast, the unadjusted V’O2max values for the boys did not differ between studies, but Next Generation boys had lower weight-adjusted V’O2max scores. This pattern of findings was unchanged when the 16 year-old Next Generation Study participants were excluded from analyses (Supplementary Table 1).

The findings from the biological mother-daughter and father-son dyads across the studies, using paired t-tests are shown in Table 2 (note that these analyses are restricted to a single parent-same-sex child dyad: DS members with more than one same-sex child in the Next Generation Study are only paired with the child nearest to 15 years in age). For both father-son and mother-daughter dyads, mean weight, height and BMI were higher in the Next Generation participants. Both absolute and weight adjusted measures of fitness were lower in Next Generation girls than they had been in their mothers at age 15. Absolute levels of fitness (ie, V’O2max not adjusted for weight) were similar across generations within the same family for boys, although sons were less fit than their fathers had been at age 15 after adjusting for body weight (Table 2). Excluding 16 year-old Next Generation Study participants did not change this pattern of findings although the difference in weight-adjusted fitness among boys did not reach statistical significance (Supplementary Table 2).

Table 2: Paired t-tests by sex for weight and fitness in the biological father-son and mother-daughter dyads from the Dunedin Study (DS) and Next Generation (NG) Studies.

 

Mothers (DS)

Daughters (NG)

 

 

n

Mean (SD)

95% CI

Mean (SD)

95% CI

Mean difference

95% CI

p

Weight

kg

71

57.0 (8.1)

55.0–58.5

65.0 (14.7)

61.5–68.5

8.1

4.4 to 11.75

<0.001

Height

m

71

1.62 (0.06)

1.61–1.64

1.66 (0.07)

1.64–1.67

0.03

0.02 to 0.05

<0.001

BMI

kg/m2

71

21.5 (2.6)

20.9–22.1

23.7 (5.0)

22.5–24.9

2.1

0.9 to 3.4

<0.001

VO2max

L/min

62

2.46 (2.37)

2.37–2.55

2.06 (0.53)

1.93–2.20

-0.40

-0.6 to -0.2

<0.001

VO2max/kg

mL/min/kg

62

43.5 (0.5)

42.5–44.5

33.0 (8.9)

30.7–35.3

-10.6

-12.9 to -8.2

<0.001

 

Fathers (DS)

Sons (NG)

 

Weight

kg

45

58.3 (10.7)

55.1–61.5

69.3 (16.0)

64.5–74.2

11.1

5.6 to 16.5

<0.001

Height

m

45

1.69 (0.07)

1.67–1.72

1.74 (0.08)

1.71–1.76

0.04

0.02 to 0.06

<0.001

BMI

kg/m2

45

20.2 (2.7)

19.4–21.0

22.9 (4.5)

21.6–24.3

2.7

1.2 to 4.2

<0.001

VO2max

L/min

37

2.92 (0.54)

2.74–3.10

2.84 (0.83)

2.56–3.11

-0.08

-0.4 to 0.2

0.613

VO2max/kg

mL/min/kg

37

49.2 (5.7)

47.3–51.1

42.2 (13.2)

37.8–46.6

-7.0

-12.2 to -1.9

0.009

Positive mean differences indicate higher values in the Next Generation Study and negative differences indicate higher values in the Dunedin Study. 

Propensity score matching of Next Generation participants to their closest match among the Dunedin Study participants provided propensity scores that were balanced between the studies. The mean differences between the Next Generation Study participants and their matched nearest neighbours in the Dunedin Study Group are shown in Table 3.

Table 3: Propensity-matched analyses of height, weight and fitness between the Dunedin Study and Next Generation Study participants.*

Girls

 

Matched pairs

Mean difference

95% CI

Weight

kg

160

11.3

7.4 to 15.2

Height

m

160

0.03

0.01 to 0.06

BMI

kg/m2

160

3.3

2.3 to 4.3

VO2max

L/min

157

-0.37

-0.47 to -0.26

VO2max/kg

mL/min/kg

157

-11.9

-14.1 to -9.7

Boys

Weight

kg

181

10.4

7.2 to 13.6

Height

m

181

0.04

0.02 to 0.06

BMI

kg/m2

181

2.5

1.6 to 3.7

VO2max

L/min

176

-0.13

-0.34 to 0.08

VO2max/kg

mL/min/kg

176

-8.5

-10.9 to -6.0

*Next Generation participants are matched to their same-sex nearest neighbours in the Dunedin Study according to maternal age, childhood socioeconomic status and ethnicity. Mean differences and 95% confidence intervals represent the mean differences between the measures in the Next Generation participants and the average of the measures in their matched nearest neighbours in the Dunedin Study. Positive mean differences indicate higher values in the Next Generation Study and negative differences indicate higher values in the Dunedin Study. 

For both sexes the weight, height and BMI were higher in the Next Generation participants. Among girls, both unadjusted and weight-adjusted fitness were lower in the Next Generation participants. Boys’ unadjusted V’O2max values were similar for both generations, but boys’ weight-adjusted fitness was lower in the Next Generation participants. Excluding 16 year-old Next Generation Study participants did not change this pattern of findings (Supplementary Table 3).

Finding from the mixed linear regression models were similar to those from the propensity analyses and are shown in Table 4.

Table 4: Mixed linear regression models with random effects for families to account for similarity between members of the same family. (Differences represent mean differences in outcome between Next Generation and Dunedin Study members after adjustment for age (age >16) and socio-economic status).

Girls

 

n

Difference

95% CI

p

Weight

kg

577

9.8

7.2 to 12.4

<0.001

Height

mm

577

0.29

18 to 40

<0.001

BMI

kg/m2

577

2.8

1.9 to 3.7

<0.001

VO2max

L/min

525

-0.39

-0.49 to -0.28

<0.001

VO2max/kg

mL/min/kg

525

-11.5

-13.2 to -9.8

<0.001

Boys

Weight

kg

626

11.8

9.3 to 14.4

<0.001

Height

mm

626

59

45 to 72

<0.001

BMI

kg/m2

626

2.60

1.83 to 3.38

<0.001

VO2max

L/min

573

0.02

-0.15 to 0.20

0.786

VO2max/kg

mL/min/kg

573

-7.23

-9.58 to -4.87

<0.001

The sex-interaction terms for the GLMs were significant for height, and both VO2max and VO2max/kg (all p<=0.001), but were not significant for BMI (p=0.840) or for weight (p=0.137). Positive mean differences indicate higher values in the Next Generation Study and negative differences indicate higher values in the Dunedin Study. 

Correlations between height, weight and fitness

Height at age 15 years was correlated between generations, but there were no significant correlations between weight, BMI or either unadjusted or weight-adjusted V’O2max measures between generations (Table 5).

Table 5: Pairwise correlations for height, weight and fitness between mother-daughter and father-son dyads from the Dunedin Study and Next Generation Study by sex.

Girls

n

r

95% Confidence intervals

P

Weight

71

0.18

-0.06–0.4

0.133

Height

71

0.52

0.32–0.67

<0.001

BMI

71

0.20

-0.04–0.41

0.096

VO2max

62

0.05

-0.21–0.29

0.714

VO2max/kg

62

0.16

-0.09–0.4

0.212

Boys

Weight

45

0.11

-0.19–0.39

0.456

Height

45

0.51

0.25–0.7

<0.001

BMI

45

0.09

-0.21–0.38

0.546

VO2max

37

0.09

-0.25–0.4

0.617

VO2max/kg

37

-0.22

-0.51–0.11

0.188

Positive mean differences indicate higher values in the Next Generation Study and negative differences indicate higher values in the Dunedin Study. 

Among the Dunedin Study participants, absolute V’O2max values at ages 15 were moderately correlated with the values at age 38 in both sexes (females: n=367, r=0.43, p<0.001; males: n=395, r=0.39, p<0.001). Similarly, weight-adjusted V’O2max values were also correlated in both sexes (females: n=367, r=0.24, p<0.001; males: n=395, r=0.35, p<0.001).

Discussion

This comparison of 15 year-olds across two generations has found that there has been a decline in adolescent physical fitness. At the same time we found increases in mean height, body weight and BMI. To our knowledge this is the first study to directly compare the physical fitness of two generations of 15 year-old New Zealanders. The findings support our hypothesis that physical fitness has declined over a generation. This finding is of concern because it indicates that the current generation of adolescents may have a higher burden of non-communicable disease due to poor fitness.

The generational decrease in fitness was most clearly seen among girls for whom there was a decline in both absolute aerobic capacity as well as aerobic capacity adjusted for body weight. Among boys there was little difference in absolute fitness between generations, but because the current generation were heavier, aerobic capacity adjusted for body weight also declined. This pattern of findings was unchanged when we restricted the analyses to compare the Next Generation participants with their same-sex biological parents and was further confirmed using propensity scores to match each Next Generation Study participant to same-sex participants of the Dunedin Study generation based on mother’s age at birth, socio-economic status and ethnicity.

We observed a smaller generational change in fitness among boys than girls, with no change in absolute levels of V’O2max. One reason for this may be might be that the sons in the Next Generation Study tended to be taller and heavier than their Dunedin Study fathers had been at age 15 whereas the differences observed between mothers and daughters were less marked. It is unclear why this is the case, but it is possible that the higher average age of the Next Generation participants means they have had more pubertal growth than was observed in the Dunedin Study sample. Because absolute aerobic capacity is related to body size, this may have confounded the findings and it is important to note that the weight-adjusted V’O2max values were significantly lower in Next Generation boys.

Another reason for the observed sex-difference may be that boys tend to be more physically active than girls.15–19 Although physical activity declines during adolescence in both sexes,16,20–22 this decline typically occurs between 9–12 years of age in girls, but between 13–16 years in boys.23 Thus, the 15 year-old girls in our research are likely to have experienced several more years of adolescent inactivity than the boys of the same age, which could plausibly explain the greater generational change in fitness in girls.

Our results also show that the 15 year-old Next Generation participants (2007–2015) are substantially heavier than their parents were at the same age in 1987/88. This within-family change is in keeping with established trends: in New Zealand between 1989–1997 mean body weight increased by 3.2kg24 and in many developed countries childhood obesity levels doubled between the 1960s and 1980s and have doubled again since.25 Sedentary behaviours, such as screen use, are increasingly displacing physical activity and contributing to the epidemic of obesity.26,27 Our findings indicate that these changes have been accompanied by a decline in physical fitness. A decline in children and adolescents’ running performance has been observed in several countries including Australia and New Zealand over recent decades.28,29 However, the extent to which these declines are due to an increase in overweight and obesity is difficult to determine from field running tests, whereas cycle ergometer tests are less sensitive to changes in body weight. A secular decline in the V’O2max of nine year-old Danish children was noted between 1985/86 and 1997/98 using a cycle test,30 but no difference was found between three cohorts of Danish adolescents between 1983 and 2003.31 We are not aware of other studies directly comparing objective cycle-ergometer tests of aerobic fitness between parents and offspring at the same age. Variability in how fitness has been measured in previous research has led to some authors to believe that fitness is hard to compare between different cohorts,32 although there seems to be a consensus that children are less active and fit today than in the past.33 Our findings support this consensus.

A surprising observation is that there were no significant correlations in weight, BMI or fitness between mothers and daughters or fathers and sons (Table 5). Although there was a moderately strong correlation for height, there seems to have been very little tracking of adolescent body weight or fitness between generations. Previous studies have mixed findings on the heritability of aerobic fitness. Some studies suggest that this may be 50% or higher, although much of this may be due to shared lifestyle factors rather than genetics.34,35 However, other studies have found only weak correlations between parent and child fitness.36

In contrast to the lack of tracking of fitness between generations, there was a substantial tracking of fitness from age 15 to age 38 years among the Dunedin Study members, indicating that adolescent fitness may have long-term consequences for adult health. By necessity, the equipment used for the exercise test has changed, although both measures were based on the same exercise protocol to obtain a steady state heart rate. The fact that the exercise performance of the Dunedin Study participants at age 38 years was predicted by their age 15 test, suggests that changes in the equipment used do not explain the lack of correlation between generations.

A strength of this research is that the Dunedin Study sample is an unselected population-based cohort with a very high rate of participation. However, because the Dunedin Study members who started their family later in life do not yet have children who are old enough to participate in the Next Generation Study, the current Next Generation Study participants have parents who are younger than average compared to the population. The Next Generation Study participants were also more likely to come from lower socio-economic status families than the Dunedin Study cohort, reflecting the likelihood of having younger parents. We also note that a limitation of our independent t-test analysis of the whole cohort is that it assumes independence between the generations, which may be not the case. However, findings from these analyses are very similar to the three other methods of analyses using matched t-tests between parent-child dyads, propensity score, and linear mixed models that do not make this assumption. The propensity score analyses provided similar findings suggesting that selection bias is unlikely to explain our findings. Although the Next Generation participants tended to be slightly older than their parents at the time of data collection, excluding those over age 16, or adjusting for this in the mixed model, made no material difference to the results.

In summary, we have shown that the trend to increasing body weights among New Zealand adolescents has been accompanied by a decline in aerobic fitness. These trends are likely to result in an increased burden on health and society as a whole. Physical fitness in adolescence tracks into adulthood, and poor fitness is an established health risk. Measures to improve physical fitness among New Zealanders are urgently needed.

Appendix

Supplementary Table 1: Mean weight, height and fitness for boys and girls in each study. Analyses exclude Next Generation participants over 15 years old.

c
P values from two-tailed independent t-tests. Positive mean differences indicate higher values in the Next Generation Study and negative differences indicate higher values in the Dunedin Study. 

Supplementary Table 2: Paired t-tests by sex for weight and fitness in the biological father-son and mother-daughter dyads from the Dunedin Study (DS) and Next Generation (NG) Study. Analyses exclude Next Generation participants over 15 years old.

 

Mothers (DS)

Daughters (NG)

 

 

 

n

Mean (SD)

95% CI

Mean (SD)

95% CI

Mean difference

95% CI

p

Weight

kg

60

57.4 (8.2)

55.3–59.5

65.0 (14.1)

61.3–68.6

7.6

3.8 to 11.4

<0.001

Height

m

60

1.62 (0.06)

1.61–1.64

1.66 (0.07)

1.64–1.67

0.03

0.02 to 0.05

<0.001

BMI

kg/m2

60

21.8 (2.6)

21.1–22.4

23.6 (4.8)

22.4–24.9

1.9

0.6 to 3.2

0.005

VO2max

L/min

54

2.48 (0.35)

2.38–2.57

2.08 (0.53)

1.93–2.22

-0.40

-0.6 to -0.2

<0.001

VO2max/kg

mL/min/kg

54

43.4 (3.9)

42.4–44.5

33.1 (9.2)

30.6–35.6

-10.4

-12.9 to -7.8

<0.001

 

Fathers (DS)

Sons (NG)

 

Weight

kg

34

57.5 (9.8)

54.1–60.9

66.8 (12.8)

62.3–71.3

9.3

4.2 to 14.5

<0.001

Height

m

34

1.69 (0.07)

1.66–1.72

1.72 (0.08)

1.69–1.75

0.04

0.01 to 0.06

0.012

BMI

kg/m2

34

20.1 (2.4)

19.2–20.9

22.4 (4.1)

21.0–23.8

2.4

0.9 to 3.9

0.003

VO2max

L/min

27

2.83 (0.49)

2.64–3.02

2.78 (0.89)

2.43–3.13

-0.05

-0.4 to 0.3

0.763

VO2max/kg

mL/min/kg

27

48.3 (4.0)

46.7–49.9

43.0 (13.8)

37.5–48.4

-5.3

-11.3 to 0.7

0.079

Positive mean differences indicate higher values in the Next Generation Study and negative differences indicate higher values in the Dunedin Study. 

Supplementary Table 3: Propensity-matched analyses of height, weight and fitness between the Dunedin Study and Next Generation Study.*

Girls

 

Matched pairs

Mean difference

95% CI

Weight

kg

134

11.1

8.1 to 14.2

Height

m

134

0.03

0.01 to 0.06

BMI

kg/m2

134

3.3

2.3 to 4.2

VO2max

L/min

131

-0.36

-0.48 to -0.23

VO2max/kg

mL/min/kg

131

-11.6

-13.9 to -9.3

Boys

Weight

kg

143

9.5

6.3 to 12.7

Height

m

143

0.04

0.01 to 0.06

BMI

kg/m2

143

2.4

1.28 to 3.48

VO2max

L/min

142

-0.11

-0.34 to 0.11

VO2max/kg

mL/min/kg

142

-7.9

-10.4 to -5.4

*Next Generation participants are matched to their same-sex nearest neighbours in the Dunedin Study according to maternal age, childhood socioeconomic status and ethnicity. Analyses exclude Next Generation participants 16 years and older. Mean differences and 95% confidence intervals represent the mean differences between the measures in the Next Generation participants and the average of the measures in their matched nearest neighbours in the Dunedin Study. Positive mean differences indicate higher values in the Next Generation Study and negative differences indicate higher values in the Dunedin Study. 

Summary

Fitness was measured at the age of 15 using an exercise cycle in both the Dunedin Study participants and their children at the same age. We found that the current generation were less fit than their parents had been and were more likely to be overweight. The decline in fitness was greater among girls than boys, although both sexes were less fit than parents had been at 15. These intergenerational changes in fitness and weight may have negative implications for the quality of long-term health.

Abstract

Aim

This research examines fitness and body weight in two cohorts of adolescents, to determine continuity and changes in these measures across two generations.

Method

Height, weight and fitness were measured in a population-based cohort of 15 year-olds in 1986/7 (Dunedin Study, n=968). The same measures were obtained for their 15–16 year-old children between 2007 and 2015 (Next Generation Study, n=343). Fitness was defined as maximal aerobic capacity (V’O2max). Height and weight were measured in all participants and fitness was adjusted for weight (V’O2max/kg).

Results

The Next Generation participants were, on average, heavier than the Dunedin Study participants had been, and had higher body mass index values (kg/m2). Unadjusted V’O2max values for boys did not differ between generations, but were lower in Next Generation girls compared to Dunedin Study girls. For both sexes, the Next Generation participants had lower weight-adjusted V’O2max values than the Dunedin Study participants. Compared to their parents, weight-adjusted V’O2max values were approximately 25% lower in girls and 15% lower in boys.

Conclusion

Overall adolescents today appear to be less fit and heavier than their parents were at the same age. The decline in fitness over a generation is particularly evident in adolescent girls, although boys also have lower levels of fitness once body weight has been taken into account.

Author Information

Helena M McAnally, Assistant Research Fellow, Department of Preventive and Social Medicine, University of Otago, Dunedin; Lesley M Reeves, Assistant Research Fellow, Department of Surgical Sciences, Dunedin School of Medicine, University of Otago, Dunedin;
Judith L Sligo, Senior Research Fellow, Department of Preventive and Social Medicine, University of Otago, Dunedin; Robert J Hancox, Professor, Department of Preventive and Social Medicine, University of Otago, Dunedin.

Acknowledgements

This research was funded by Health Research Council of New Zealand. HMM was funded by a grant from the Department of Preventive and Social Medicine at the University of Otago. We would like to thank the participants of the Dunedin and the Next Generation Studies. We would also like to thank the research teams for both studies who collected the data, the Study founder Dr Phil Silva and the current director Professor Richie Poulton. We also thank Professor David Fergusson and Mr Andrew Gray for their advice on the analyses and the anonymous reviewers for their thoughtful suggestions on the manuscript. 

Correspondence

Robert J Hancox, Department of Preventive and Social Medicine, University of Otago, Dunedin 9056.

Correspondence Email

bob.hancox@otago.ac.nz

Competing Interests

Dr Hancox reports grants from Health Research Council (NZ) during the conduct of the study.

References

  1. World Health Organization. Global Strategy on Diet, Physical Activity and Health: Diet and physical activity: a public health priority. 2012 [cited 23 February 2012]; Available from:
  2. http://www.who.int/dietphysicalactivity/en/
  3. Landhuis CE, Poulton R, Welch D, Hancox RJ. Programming obesity and poor fitness: the long-term impact of childhood television. Obesity 2008; 16:1457–9.
  4. Hancox RJ, Milne BJ, Poulton R. Association between child and adolescent television viewing and adult health: a longitudinal birth cohort study. Lancet. 2004; 364:257–62.
  5. Hruby A, Chomitz V, Arsenault LN, et al. Predicting maintenance or achievement of healthy weight in children: the impact of changes in physical fitness. Obesity 2012; 20:1710–7.
  6. Booth FW, Roberts CK, Laye MJ. Lack of exercise is a major cause of chronic diseases. Compr Physiol. 2012; 2:1143–1211. doi:10.1002/cphy.c110025.
  7. Tomkinson GR, Annandale M, Ferrar K. Global changes in cardiovascular endurance of children and youth since 1964: systematic analysis of 25 million fitness test results from 28 countries. Circulation. 2013; 128:A13498.
  8. Poulton R, Moffitt TE, Silva PA. The Dunedin Multidisciplinary Health and Development Study: overview of the first 40 years, with an eye to the future. Soc Psychiatry Psychiatr Epidemiol. 2015; 50:679–93.
  9. Sligo J, Rothstein D, Salter G, Hancox R. How do Researchers do it? A Description of Developing a Multi-disciplinary Research Assessment. Childrenz Issues; 2009:21–9.
  10. Buono MJ, Roby JJ, Micale FG, Sallis JF. Predicting maximal oxygen uptake in children: Modification of the Astrand-Ryhming Test. Pediatric Exercise Science. 1989; 1:278–83.
  11. Ross RM. Interpreting exercise tests. Clinical interpretation of cardiopulmonary excercise (CPX) tests. Houston, TX: CSI Software; 1989.
  12. Poulton R, Caspi A, Milne BJ, et al. Association between children’s experience of socioeconomic disadvantage and adult health: a life-course study. Lancet. 2002; 360:1640–5.
  13. Galbraith C, Jenkin G, Davis P, Coope P. New Zealand Socio-economic Index 1996 Users’ Guide. Wellington, New Zealand: Statistics New Zealand; 2003.
  14. Williamson EJ, Forbes A. Introduction to propensity scores. Respirology (Carlton, Vic). 2014; 19:625–35.
  15. StataCorp. 2013. Stata Statistical Software: Release 13. College Station, TX: StataCorp LP
  16. Adolescent Health Research Group. Youth ‘07: The Health and Wellbeing of Secondary School students in New Zealand. Initial Findings. Auckland: The University of Auckland; 2008.
  17. Hohepa M, Schofield G, Kolt G. Adolescent obesity and physical inactivity. N Z Med J. 2004; 117:U1210.
  18. Reeder AI, Stanton WR, Langley JD, Chalmers DJ. Adolescents’ sporting and leisure time physical activities during their 15th year. Can J Sport Sci. 1991; 16:308–15.
  19. Ruiz JR, Ortega FB, Martinez-Gomez D, et al. Objectively measured physical activity and sedentary time in European adolescents: the HELENA study. Am J Epidemiol. 2011; 174:173–84.
  20. Ridgers ND, Salmon J, Parrish AM, et al. Physical activity during school recess: a systematic review. Am J Prev Med. 2012; 43:320–8.
  21. Aaron DJ, Storti KL, Robertson RJ, et al. Longitudinal study of the number and choice of leisure time physical activities from mid to late adolescence: implications for school curricula and community recreation programs. Arch Pediatr Adolesc Med. 2002; 156:1075–80.
  22. Belanger M, Gray-Donald K, O’Loughlin J, et al. When adolescents drop the ball: sustainability of physical activity in youth. Am J Prev Med. 2009; 37:41–9.
  23. Dovey SM, Reeder AI, Chalmers DJ. Continuity and change in sporting and leisure time physical activities during adolescence. Br J Sports Med. 1998; 32:53–7.
  24. Dumith SC, Gigante DP, Domingues MR, Kohl HW, 3rd. Physical activity change during adolescence: a systematic review and a pooled analysis. Int J Epidemiol. 2011; 40:685–98.
  25. Ministry of Health. NZ Food: NZ People - Key results of the 1997 National Nutrition Survey. Wellington: Ministry of Health; 1999.
  26. OECD. Health at a Glance 2011: OECD Indictors. OECD Publishing; 2011.
  27. French SA, Story M, Jeffery RW. Environmental influences on eating and physical activity. Annu Rev Public Health. 2001; 22:309–35.
  28. Brownson RC, Boehmer TK, Luke DA. Declining rates of physical activity in the United States: what are the contributors? Annu Rev Public Health. 2005; 26:421–43.
  29. Huotari PRT, Nupponen H, Laakso L, Kujala UM. Secular trends in aerobic fitness performance in 13–18-year-old adolescents from 1976 to 2001. Br J Sports Med. 2010; 44:968–72.
  30. Tomkinson GR, Olds TS. Secular changes in aerobic fitness test performance of Australasian children and adolescents. Med Sport Sci. 2007; 50:168–82.
  31. Wedderkopp N, Froberg K, Hansen HS, Andersen LB. Secular trends in physical fitness and obesity in Danish 9-year-old girls and boys: Odense School Child Study and Danish substudy of the European Youth Heart Study. Scandinavian journal of medicine & science in sports. 2004; 14:150–5.
  32. Andersen LB, Froberg K, Kristensen PL, et al. Secular trends in physical fitness in Danish adolescents. Scand J Med Sci Sports. 2010; 20:757–63.
  33. Morrow JR, Jr. 2004 C. H. McCloy Research Lecture: are American children and youth fit? It’s time we learned. Res Q Exerc Sport. 2005; 76:377–88.
  34. Livingstone MB. Energy expenditure and physical activity in relation to fitness in children. Proc Nutr Soc. 1994; 53:207–21.
  35. Foraita R, Brandes M, Günther F, et al. The influence of aerobic fitness on obesity and its parent-offspring correlations in a cross-sectional study among German families. BMC Public Health. 2015; 15:638.
  36. Bouchard C, Daw EW, Rice T, et al. Familial resemblance for VO2max in the sedentary state: the HERITAGE family study. Medicine and science in sports and exercise. 1998; 30:252–8.
  37. Sallis JF, Patterson TL, Morris JA, et al. Familial aggregation of aerobic power: the influence of age, physical activity, and body mass index. Res Q Exerc Sport. 1989; 60:318–24.