Temporal Trends in Health-Related Components of Physical Fitness of European Children and Adolescents from 1965 to 2025: A Systematic Review of Data on 417,362 Participants from 20 Countries

Abstract

Background

Physical fitness is an important indicator of health during childhood and adolescence. Several components of physical fitness showed declining trends, with periods of acceleration or stabilisation depending on the component. Although Europe is the most studied continent, existing reviews have largely relied on single-test approaches or descriptive summaries, without simultaneously examining all health-related components, accounting for between-study variability analyses, or extending monitoring beyond 2014 to include the COVID-19 pandemic. This systematic review aims to analyse temporal trends of health-related components of physical fitness in European children and adolescents aged 6–16 years, examining differences by sex and age groups.

Methods

A systematic search was conducted in PubMed, Web of Science, SPORTDiscus and Scopus up to December 2025. Eligible studies were conducted in European countries and included children and adolescents aged 6–16 years, using repeated cross-sectional designs and reporting physical fitness components at two or more time points including sample sizes, means and standard deviations. Study quality was assessed with a modified Downs and Black checklist. Data were standardised into z scores based on sample-weighted means and standard deviations, considering the study, physical fitness component, sex and age. Linear mixed models were developed for each physical fitness component to estimate trends (linear, quadratic or cubic) and interactions by sex or age group.

Results

Thirty-eight studies were included, representing 417,362 children and adolescents, from 20 European countries between 1965 and 2023. Cardiorespiratory endurance increased until the 1980s, then declined with a deceleration between 2015 and 2023. Body composition decreased linearly up to 2022, more markedly among males. Upper-body strength declined from the 1970s to 2020s, stabilising between the 1980s and 2000s. Lower-body strength increased until the 1980s and declined until 2023. Flexibility declined continuously, with an accelerated decrease after the 2000s.

Conclusions

Health-related components of physical fitness showed persistent declines among European children and adolescents, with differences in the shape and magnitude of trends. These findings highlight the heterogeneity of the available evidence and the need for a standardised, valid and feasible European fitness test battery. Public initiatives should continue to promote adherence to physical activity guidelines, particularly muscle-strengthening recommendations.

Registration

PROSPERO ID: CRD42024609888.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40798-026-01032-x.

Key Points

This systematic review analyses temporal trends in physical fitness among 417,362 European children and adolescents (aged 6–16) from 38 studies, conducted across 20 countries between 1965 and 2023.

The findings indicate a decline in body composition, upper-body strength, lower-body strength and flexibility. Body composition trend appears to be more pronounced among males. Cardiorespiratory endurance showed a recent deceleration in its decreasing trend.

Test batteries need to be standardised to enhance the monitoring of physical fitness trends across Europe. Public initiatives should continue to promote adherence to WHO’s physical activity recommendations, especially muscle-strengthening guidelines, while also encouraging aerobic physical activity recommendations.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40798-026-01032-x.

Background

The World Health Organization (WHO) recommends that children and adolescents engage in at least 60 min of moderate-to-vigorous physical activity daily, including muscle and bone-strengthening activities at least three times per week [1]. Evidence confirms a dose–response relationship between physical activity and health benefits across several health indicators such as blood pressure, cholesterol, metabolic syndrome, obesity, bone density or depression [2]. In Europe, the HELENA study demonstrated its protective effect against cardiometabolic risk in adolescents [3]. Despite this, over 75% of European adolescents do not meet the above recommendations [4] and may experience the “Exercise Deficit Disorder” described by Faigenbaum et al. [5], as a component of the “Pediatric Inactivity Triad”, along with pediatric dynapenia and physical illiteracy [6, 7]. In this context, beyond monitoring physical activity levels, it is essential to also consider physical fitness, because it has been shown to be a powerful marker of health during childhood and adolescence [8].

According to Caspersen et al. and the American College of Sports Medicine (ACSM), the health-related components of physical fitness include cardiorespiratory (CR) endurance, muscular strength and endurance (often summarised as muscular fitness or simply strength), body composition, and flexibility [9, 10]. Low levels of CR endurance are even more strongly related to cardiovascular risk factors than physical activity in children and adolescents [11], showing that CR endurance has an important cardioprotective role [12], and is associated with better cardiovascular health in later life [13]. Specifically, maximal oxygen uptake values below 42 in males and 35 mL/kg/min in females increase the risk of cardiovascular disease [14]. There is, also, an association between CR endurance, body fat, and cardiometabolic risk factors such as blood pressure, low-density lipoprotein (LDL), high-density lipoprotein (HDL), total cholesterol, triglycerides or fasting blood glucose [15]. Higher CR endurance and healthy body composition are associated with more favourable cardiometabolic profiles [16], suggesting that improving CR endurance and body composition during childhood and adolescence, could reduce the risk of cardiovascular disease [17]. In addition to CR endurance, healthy body composition during childhood and adolescence is associated with a better cardiovascular profile and reduced mortality risk in later life, while improvements of muscular strength during these periods are related to healthier body composition [13]. Evidence suggests that muscular fitness is also associated with cardiometabolic risk in adolescents [18, 19]. A meta-analysis found that muscular fitness in youth is negatively associated with cardiometabolic risk parameters and positively with bone health [20]. Similar to CR endurance, specific cut-off points have been established to identify children and adolescents at risk of cardiometabolic issues using muscle strength field-based tests [21, 22]. In contrast to the other components, flexibility does not appear to be associated with cardiometabolic risk factors [19] and may lack predictive and concurrent validity for significant health outcomes [23]. Nevertheless, the ACSM continues to include flexibility among the health-related components of physical fitness [10].

This relationship between physical fitness and health, makes it essential to assess this significant indicator of health during childhood and adolescence. A systematic review identified 24 batteries of field-based tests to evaluate physical fitness in children and adolescents worldwide [24]. The Council of Europe in 1988, proposed and recommended the EUROFIT test battery [25, 26]. Since then, several batteries were developed in different European countries such as SLOfit in Slovenia [27], ALPHA-Fitness in Spain [28], INDARES in Czech Republic [29], BOUGE in France [30], ASSO in Italy [31] or FITescola in Portugal [32]. Furthermore, a recent consensus project on the YFIT battery has been published [33], indicating that the tests included in the battery met a consensus target of 80% among 92 European experts, and across the full sample of 169 experts surveyed globally. However, Marques et al. [24] noted that advancements in the physical fitness assessment and the increasing number of existing assessment batteries have complicated the comparison of data from different regions. This challenge persists even in Europe, where countries share geographical, political and cultural similarities. The lack of this standardisation also limits the ability to establish common reference values or inform unified educational and health strategies. This issue is particularly relevant when analysing temporal patterns in physical fitness, which provide valuable insights into the evolving health status of youth populations.

Despite these difficulties, it is still crucial to analyse physical fitness trends over time to better understand the current health status of children and adolescents. Tomkinson and Olds [34] described trends in CR endurance between 1958 and 2003 among 25 million youths aged 6–19 years from 27 countries, including 16 European nations, and reported a global decline of 0.36% per year. Between 1958 and 1970, CR endurance improved by 0.61%, but then declined by 0.54% per year. An updated study by Tomkinson et al. [35], analysed 965,264 children and adolescents from 19 countries (12 European) using the 20-m shuttle run test. Researchers reported a decline of 7.3% (0.22% per year) between 1981 and 2014, with the decline slowing from 0.36 to 0.09% per year after 2000. Systematic reviews by Masanovic et al. [36] and Eberhardt [37], mainly based on European studies, showed a consistent decrease in CR endurance. In contrast, Fühner et al. [38], conducted a z score-based linear mixed model (LMM) which include 12 European studies, and found an increase in CR endurance until 1986, a decline until 2010–2014, and then a stabilisation.

Trends in CR endurance align with the global obesity epidemic, as described by the WHO [39]. Lobstein et al. [40] referred to obesity as a public health crisis in 2004 and observed rising trends in overweight and obesity from 1970 to 2000 across several countries, including Europeans. Between 1985 and 2019, some European countries showed changes in BMI of less than 0.5 kg/m², suggesting more stable BMI trajectories than other countries, such as the USA and New Zealand [41]. A global analysis of 63 million children and adolescents revealed that the prevalence of obesity increased from 1.7% to 6.9% among girls and from 2.1% to 9.3% among boys, between 1990 and 2022 [42]. However, this large-scale surveillance evidence is based on BMI, which has low sensitivity for detecting excess adiposity and fails to identify more than 25% of children with excess of body fat [43]. In contrast, little evidence has analysed secular trends in estimated body composition using more reliable methods. One international study conducted by Olds [44] analysed secular trends in skinfold-estimated relative body fat, reporting an increase of 0.86% body fat per decade between 1954 and 2004 based on 153 cross-sectional studies (47% European). Thus, despite the extensive literature on BMI-based trends, there is a lack of evidence since 2004 regarding the actual body composition trend.

Regarding other fitness components, Tomkinson [45] examined trends in power or explosive strength and speed, with most studies focusing on European youth. From 1958 to 2003, power improved by 0.03% and speed by 0.04% per year. However, between 1958 and 1985 power improved by 0.44% and speed by 0.27% per year, followed by declines of 0.20% and 0.08% per year respectively. Recently, Tomkinson et al. [46] reported that explosive lower body (LB) strength, measured by the standing broad jump, from 16 European and 13 non-European countries, improved from the 1960s to the 1980s, slowed in the 1990s, and declined until 2017. Other systematic reviews found decreases in upper body (UB) strength and explosive LB strength in most studies [36, 37]. Dooley et al. [47] studied over 2 million participants from 19 countries (9 European) and found a global increase in handgrip strength between 1967 and 2017. However, the model of Fühner et al. [38], described a decline until 1982, followed by an increase until 2006 in relative muscle strength. While muscle power showed an increase until 1982 for females and 1994 for males, after which the trend later changed. Finally, although flexibility analyses are insufficient to draw meaningful conclusions, two systematic reviews found a decrease in most of the studies included [36, 37].

Despite the existing evidence on physical fitness trends in children and adolescents, there are some important limitations in the literature. Most recent systematic reviews either rely on descriptive summaries without statistical trend analyses [36, 37] or focus on a single field-based test [35, 46, 47], which exclude studies using other tests. This may limit the ability to accurately represent trends in physical fitness components. In contrast, most studies that have applied statistical analyses while integrating multiple tests within the same component only extended their trend data up to the early 2000s [34, 45]. Furthermore, while BMI-based trends are well documented [41, 42], evidence examining secular changes using more reliable methods to estimate body composition is limited [44]. In addition, excepting Fühner et al. [38], existing reviews have not employed statistical approaches that consider both fixed effects (e.g. time) and random effects, such as variability between studies. Moreover, the evidence mentioned above is from before 2017, and therefore does not reflect recent changes, including the impact of COVID-19 pandemic, which has led to a decrease in physical activity levels among youth [48, 49]. Finally, no systematic review to date has comprehensively analysed the evolution of physical fitness from the perspective of health-related components, over an extended time span and with a focus on European children and adolescents. As Europe is probably the most studied continent, because 47–71% of the studies included in systematic reviews on global trends [34, 36–38, 45–47] are from European countries, it is convenient to update its physical fitness trends. Therefore, this systematic review aims to understand and update the temporal trends of health-related physical fitness in European children and adolescents aged 6–16 years. Secondly, to analyse possible differences between sexes (males vs. females) and between age groups (children vs. adolescents).

Methods

This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement guidelines [50, 51]. The PRISMA checklist is available in Supplementary Material 1—Table S1. Moreover, it was registered in the International Prospective Register of Systematic Reviews (PROSPERO: CRD42024609888). Amendments to the original protocol and their justification can be found in Supplementary Material 2—Table S2.

Eligibility Criteria

Studies published up to December 2025 were included if they met the following criteria:

1. Original published studies.

2. Outcomes of health-related physical fitness components [10] had to be reported as mean, standard deviation or standard error, along with sample size. Alternatively, they must be available upon request.

3. Participants had to be apparently healthy children and adolescents aged 6–16 years, categorized into age groups no more than three years apart, according to Tomkinson [34, 45], or be available upon request.

4. Conducted in European countries.

5. Included at least two different cross-sectional cohorts with similar characteristics assessed at different time periods within the same country (i.e., not longitudinal follow-up of the same participants, who would age between measurements).

6. Written in English or Spanish.

Grey literature was excluded to prioritise peer-reviewed evidence with clearly reported methodologies and outcome data suitable for quantitative modelling. Although this decision may introduce publication bias, this decision was taken to ensure data quality, methodological consistency, and comparability across studies. In addition, studies focusing solely on obesity prevalence or height, weight and somatic development, or those involving children or adolescents with specific conditions (e.g., young athletes or with a specific disease) were also excluded.

Information Sources and Search Strategy

The literature search was conducted using the following databases: PubMed, Web of Science, SPORTDiscus and Scopus. The search utilized the Boolean operator “AND” to link the general topics divided in four blocks: (1) physical fitness terms and related keywords, (2) population (children or adolescents), (3) terms related to temporal trends or changes, and (4) Europe or European countries. The operator “OR” was employed to include related terms and relevant thesauri specific to each database, within each of the four term blocks. The schematic search syntax was as follows: (“cardiorespiratory endurance” or “muscle strength” or “flexibility” or “body composition”) and (“children” or “adolescents”) and (“secular trend” or “temporal trend” or “secular changes” or “temporal changes”) and “Europe”. This syntax represents an illustrative example, and the complete, database-specific search strategies are provided in Supplementary Material 3.

Selection Process

The study selection was conducted by two researchers (DDV and AML). All potentially eligible studies were compiled in an Excel spreadsheet. First, duplicate records were removed by cross-referencing titles, authors and abstracts. Then, the remaining studies were screened independently by title and abstract. Subsequently, full-text articles were then retrieved and evaluated in detail based on the eligibility criteria. The criterion for excluding each full-text study was recorded. Any disagreements that occurred at any stage were resolved through discussion with a third researcher (JAC).

Data Collection Process and Data Items

Several members of the review team (DDV, GLB, AMF and AGB) collected data from the included studies using a Microsoft Excel spreadsheet. The data items included the following: country, sex, physical fitness component, test, age, and age group (children under 13 years old or adolescents). Nine studies [52–60] used more than one test to assess the same health-related physical fitness component; consequently, only the most commonly used test was selected, consistent with the rest of the studies. Adiposity-related whole-body outcomes were selected for body composition studies. Sample or subsample sizes for participants aged 6–16 along with means and standard deviations/standard errors, were extracted when available or requested from the corresponding authors. Tutkuviene and Schiefenhövel [61], Kopecký et al. [62], Sedlak et al. [63], Przeweda et al. [64], Spengler et al. [65], Nebiker et al. [66], Unger et al. [59] and Kryst et al. [67] provided samples, means and/or standard deviations upon request.

Three studies [56, 57, 61] reported results for the same test on both the right and the left arm or leg, and these results were averaged. Following the methods described by Wan et al. [68], means and standard deviations for the Bent Arm Hang test in Anselma et al. [69] were estimated from the median and interquartile range. Also, in line with Wan et al. [68], standard deviations for the handgrip and standing broad jump were estimated from the minimum and maximum values in Sandercock and Cohen [60]. In two studies [70, 71], standard deviations were calculated from standard errors. In one study [72], means and standard deviations were presented on a logarithmic scale and were converted to a linear scale prior to inclusion. The data from two studies were presented in plots only [58, 73], so it was extracted using PlotDigitizer, 3.1.6, 2025. The complete database is available in Supplementary Material 4.

Methodological Quality Assessment

The methodological quality of the included studies was assessed using an adapted version of the Downs and Black checklist [74], as adapted by Hinckson et al. [75]. This tool consists of ten criteria divided into four categories: descriptive information, external validity, internal validity, and clinical implications. Each item was scored as “yes”, “no”, or “not available”. Item 9, ‘Valid and reliable outcome measures’, was scored according to the systematic review conducted by Castro-Piñero [76], considering recent evidence on the validity of 20 m shuttle run test [77]. The final score for each study was obtained by summing the positive responses. Based on the total score, studies were classified into five categories of methodological quality: bad (0–20%), poor (21–40%), fair (41–60%); good (61–80%) and excellent (81–100%). The quality assessment was reported in a table.

Synthesis Methods

A table was prepared to summarize the relevant information from each included study, including the reference, country, sample characteristics, time points and duration, physical fitness components, the test used to assess them, and study quality. Outcome measures were not included in this table because the large number of subgroups by each age and sex and by each physical fitness component would make it unwieldy. This information is available in Supplementary Material 4.

To standardize the results, z scores based on sample-weighted means and standard deviations were calculated for each subgroup defined by study, outcome, age and sex. All body composition outcomes were relative to body fat, so z scores values for this component were multiplied by − 1 to facilitate interpretation. Tests for which a lower score indicates better performance were also multiplied by − 1. As a result, in all tests, a higher score indicates better performance. It should be noted that while this standardization facilitates the aggregation of heterogeneous field-based tests, it does not fully eliminate differences in test validity or reliability.

Statistical Analysis

For each health-related physical fitness component, a LMM was specified to estimate the effects of sex, age group (children or adolescents), and linear, quadratic or cubic trends, treating the studies as levels of a random factor (intercept). This approach differs from that of Fühner et al. [38], who specified a LMM nested under other physical fitness components: endurance, relative strength, power and speed. In the present systematic review, each health-related physical fitness component was modelled separately to allow a focused examination of specific temporal patterns, given their distinct physiological and measurement characteristics. Although this approach may reduce statistical power, it facilitates a more interpretable assessment of trends within each component. Thus, a LMM was developed for each health-related physical fitness component: CR endurance, Body Composition, UB Strength, LB Strength, and Flexibility.

The LMMs were estimated using the GAMLj module in Jamovi version 2.6.22 [78–82]. To determine the optimal model structure, two models were developed for each physical fitness component. The first model included the fixed effects of sex, age group, time (linear, quadratic, and cubic), and all possible interactions among them, resulting in 11 fixed effects. The second model included the same primary fixed effects, but only two interactions: time by sex, and time by age group (7 fixed effects). Then, temporal trends were determined using best-fitting cubic (7 fixed effects), quadratic (6 fixed effects) or linear (5 fixed effects) LMM relating the year of testing to z scores. Model selection was based on a comparison of the Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC), to balance model fit and complexity, avoid overparameterization, and select parsimonious LMMs [83–85]. This selection was presented in a table alongside the AIC and BIC values of each tested LMM. The effect estimates for each selected LMM were summarized in a table, and graphical trends along with their residual plots were generated using RStudio version 2024.12.0 + 467, with the syntax for the five LMMs exported from Jamovi.

An additional CR endurance LMM was developed that excluded the 20 m shuttle run test. This was done as a sensitivity analysis to observe the impact of this test on CR endurance trend. A further LMM was developed for BMI, to allow a direct comparison between BMI and body composition trends. Model selection and effect estimates were described in the text, and graphical trends for both LMMs were generated alongside CR endurance and body composition in a separate figure for comparison.

Results

There were 1,335 potentially relevant studies, identified in the four databases: PubMed, Web of Science, SPORTDiscus and Scopus. Eleven articles were found through other relevant systematic reviews and studies. After removing duplicate records, and applying the eligibility criteria, 38 studies were finally included in this systematic review. The process was described in Fig. 1.

Fig. 1.

PRISMA Flow diagram of the study selection process [50, 51]

Characteristics of the Included Studies

The characteristics of the included studies are summarised in Table 1. A total of 417,362 (45.7% females; 43.6% adolescents) participants aged 6 to 16 years were analysed. The data were collected from 20 European countries, covering a time span from the first observation year in 1965 to the last in 2023. Fifteen studies conducted repeated cross-sectional analyses involving more than two time points [55, 59, 60, 62, 65–67, 70, 72, 86–91]. Seventeen studies assessed CR endurance [53, 55, 57, 58, 64–66, 71, 88–96], 11 studies assessed body composition [58, 67, 72, 73, 87, 92, 97–101], 11 studies assessed UB strength [53, 54, 56, 61–65, 69, 88, 101], 16 studies assessed LB strength [52–54, 56, 57, 60, 62–64, 66, 69–71, 86, 88, 102], and only seven studies assessed flexibility [54, 56, 57, 69–71, 101].

Table 1.

Summary of the studies in this systematic review

Study	Country	Sample	Time points (years)	Physical fitness component—test
Tutkuviene and Schiefenhövel [61]	Lithuania	M 7–16 yr (n = 1123; 1116) F 7–16 yr (n = 1113; 1113)	1965; 1985 (20)	UB strength—handgrip (kg)
Kopecký et al. [62]	Czech Republic	F 7–15 yr (n = 9349; 261; 562)	1966; 1968; 2002 (36)	UB strength—medicine ball throw (m) LB strength—standing broad jump (cm)
Chulvi-Medrano et al. [52]	Spain	M 10–11 yr (n = 140; 113)	1969; 2016 (47)	LB strength—standing broad jump (m)
Matton et al. [101]	Belgium	M 12–16 yr (n = 9718; 1217) F 12–16 yr (n = 3406; 1475)	1972; 2005 (33) 1980; 2005 (25)	Body composition—sum of skinfolds (mm) UB strength—bent arm hang (s) Flexibility—sit and reach (cm)
Westerstahl et al. [53]	Sweden	M 16 yr (n = 202; 236) F 16 yr (n = 193; 224)	1974; 1995 (21)	CR endurance—9 min walk-run (m) UB strength—bench press at 25 reps/min (n) LB strength—standing high jump (cm)
Photiou et al. [58]	Hungary	M 7.5–14.5 yr (n = 3672; 3758)	1975; 2000 (25)	CR endurance—1200 m run (s) Body composition—%BF (skinfold-estimated)
Sedlak et al. [63]	Czech Republic	M 6 yr (n = 154; 133) F 6 yr (n = 162; 137)	1977; 2012 (35)	UB strength—ball throw (cm) LB strength—standing broad jump (cm)
Jaakkola et al. [89]	Finland	M 14–16 yr (n = 299; 243; 416; 1178; 701; 354) F 14–16 yr (n = 300; 255; 380; 1205; 682; 365)	1979; 1995; 1998; 2003; 2010; 2020 (41)	LB strength—5-leap (cm)
Moreno et al. [97]	Spain	M 6.5–14.5 yr (n = 1553; 701)	1980; 1995 (15)	Body composition—%BF (skinfold-estimated)
Mészáros et al. [73]	Hungary	M 7–16 yr (n = 10,885; 10,897)	1980; 2005 (20)	Body composition—%BF (skinfold-estimated)
Runhaar et al. [54]	Netherlands	M 9–12 yr (n = 1288; 1010) F 9–12 yr (n = 1315; 1040)	1980; 2006 (26)	UB strength—bent arm hang (s) LB strength—standing high jump (cm) Flexibility—sit and reach (cm)
Kowal et al. [72]	Poland	M 6.5–15.5 yr (n = 1579; 1742; 1524)	1983; 2000; 2010 (27)	Body composition—sum of skinfolds (mm)
Kowal et al. [87]	Poland	F 6.5–15.5 yr (n = 2164; 1526; 1229)	1983; 2000; 2010 (27)	Body composition—sum of skinfolds (mm)
Lovecchio et al. [55]	Italy	M 11–13 yr (n = 439; 454; 247; 242; 285; 261; 285; 318; 349) F 11–13 yr (n = 224; 223; 263; 283; 211; 276; 309; 261; 355)	1985; 1988; 1991; 1994;1997; 2000; 2003; 2006; 2009 (24)	CR endurance—cooper (m/s)
Wedderkop et al. [92]	Denmark	M 9 yr (n = 699; 279) F 9 yr (n = 670; 310)	1986; 1998 (12)	CR endurance—cycle-ergometer watt-max (estimated mL/kg/min) Body composition—%BF (skinfold-estimated)*
Aaberge et al. [57]	Norway	M 15 yr (n = 104; 99) F 15 yr (n = 83; 77)	1988; 2001 (13)	CR endurance—Åstrand-Ryhming (estimated mL/kg/min) LB strength—standing high jump (cm) Flexibility—straight leg raise (cm)
Przeweda et al. [64]	Poland	M 7–16 yr (n = 96,147; 32,344) F 7–16 yr (n = 92,165; 31,467)	1989; 1999 (10)	CR endurance—cooper (m) UB strength—handgrip (kg) LB strength—standing broad jump (cm)
Watkins et al. [98]	Ireland	M 12 and 15 yr (n = 503; 1019) F 12 and 15 yr (n = 512; 998)	1990; 2000 (10)	Body composition—%BF (skinfold-estimated)
Venckunas et al. [88]	Lithuania	M 11–16 yr (n = 2472; 2133; 1914) F 11–16 yr (n = 2544; 2071; 1700)	1992; 2002; 2012 (20)	CR endurance—20 m shuttle run (min) UB strength—bent arm hang (s) LB strength—standing broad jump (cm)
Smpokos et al. [71]	Greece	M 6.5 yr (n = 282; 183) F 6.5 yr (n = 253; 153)	1993; 2007 (14)	CR endurance—20 m shuttle run (stages) LB strength—standing broad jump (m) Flexibility—sit and reach (cm)
Huotari et al. [89]	Finland	M 14–15 yr (n = 239; 354; 1183; 683; 337) F 14–15 yr (n = 252; 309; 1207; 663; 334)	1995; 1998; 2003; 2010; 2020 (25)	CR endurance—20 m shuttle run (laps)
Costa et al. [70]	Portugal	M 10–11 yr (n = 229; 198; 151; 262) F 10–11 yr (n = 237; 222; 182; 259)	1996; 2001; 2006; 2011 (15)	LB strength—standing broad jump (cm) Flexibility—Sit and reach (cm)
Møller et al. [93]	Denmark	M 8–10 yr (n = 258; 185) F 8–10 yr (n = 282; 243)	1998; 2004 (6)	CR endurance—cycle-ergometer watt-max (W/kg)
Andersen et al. [94]	Denmark	M 15–16 yr (n = 198; 184) F 15–16 yr (n = 212; 234)	1998; 2004 (6)	CR endurance—cycle-ergometer watt-max (W/kg)
Sandercock et al. [90]	England	M 10–11 yr (n = 158; 145; 157) F 10–11 yr (n = 158; 157; 150)	1998; 2008; 2014 (16)	CR endurance—20 m shuttle run (km/h)
Sandercock and Cohen [60]	England	M 10–11 yr (n = 160; 154; 157) F 10–11 yr (n = 147; 150; 149)	1998; 2008; 2014 (16)	UB strength—handgrip (kg) LB strength—Standing broad jump (cm)
Kryst et al. [67]	Poland	M 8–16 yr (n = 1606; 1139; 949) F 8–16 yr (n = 1460; 1110; 1042)	2000, 2010, 2020 (20)	Body composition—Sum of skinfolds (mm)
Thomas et al. [96]	Wales	M 12.5 yr (n = 33; 33) F 12.5 yr (n = 38; 50)	2002; 2007 (5)	CR endurance—20 m shuttle run (laps)
Arboix-Alió et al. [91]	Spain	M 15–16 yr (n = 256; 233; 212;213) F 15–16 yr (n = 194; 220; 193;179)	2002; 2007; 2012; 2017; (15)	CR endurance—20 m shuttle run (stages)
Huotari et al. [102]	Finland	M 15–16 yr (n = 1167; 656) F 15–16 yr (n = 1181; 634)	2003; 2010 (7)	LB strength—15 s lateral jump (n)
Palomäki et al. [95]	Finland	M 15–16 yr (n = 1116; 661) F 15–16 yr (n = 1142; 640)	2003; 2010 (7)	CR endurance—20 m shuttle run (m)
Anselma et al. [69]	Netherlands	M 10–12 yr (n = 1372; 1010) F 10–12 yr (n = 1315; 1040)	2006; 2016 (10)	UB strength—bent arm hang (s) LB strength—standing high jump (cm) Flexibility—sit and reach (cm)
Spengler et al. [65]	Germany	M 6–7 yr (n = 274; 270; 282; 260; 251; 251; 231; 241; 258; 221) F 6–7 yr (n = 264; 259; 270; 238; 241; 241; 263; 250; 216; 220)	2006; 2007; 2008; 2009; 2010; 2011; 2012; 2013; 2014; 2015 (9)	CR endurance—6 min run (m) UB strength—push-ups in 40 s (n)
Unger et al. [59]	Austria	M 9–11 yr (n = 217; 209; 198; 123; 157; 144; 146; 134; 141; 174; 143; 101; 137; 142; 123; 122; 148; 151) F 9–10 yr (n = 31; 55; 58; 35; 47; 34; 51; 41; 48; 43; 58; 45; 51; 55; 44; 36; 50; 31)	2006; 2007; 2008; 2009; 2010; 2011; 2012; 2013; 2014; 2015; 2016; 2017; 2018; 2019; 2020; 2021; 2022; 2023	CR endurance—8 min run (m) UB strength—medicine ball throw (cm) LB strength—standing broad jump (cm)
Kryst et al. [56]	Poland	M 8–16 yr (n = 911; 927) F 8–16 yr (n = 853; 978)	2010; 2020 (10)	UB strength—handgrip (kg) LB strength—standing broad jump (cm) Flexibility—sit and reach (cm)
Nebiker et al. [66]	Switzerland	M 6–8 yr (n = 673; 674; 779; 753; 826; 716; 338; 824) F 6–8 yr (n = 628; 628; 762; 790; 817; 705; 315; 725)	2014; 2015; 2016; 2017; 2018; 2019; 2020; 2021 (7)	CR endurance—20 m shuttle run (laps) LB strength—15 s lateral jump (n)
Kutac et al. [99]	Czech Republic	M 11-16 yr (n = 693; 461)	2019, 2021 (2)	Body composition—%BF (bioimpedance)
Artymiak et al. [100]	Poland	M 11–15 yr (n = 512; 288) F 11–15 yr (n = 557; 305)	2020, 2022 (2)	Body composition—Sum of skinfolds (mm)

*Reported as median and its 95% confidence interval. Not included in the linear mixed model

%BF % body fat, CR cardiorespiratory, F female, LB lower body, M male, m metres, min minute, reps repetitions, s seconds, UB upper body, yr year

A total of 23 different physical fitness tests were identified in the selected studies. From most to least frequent, the following tests were utilized to evaluate CR endurance: the 20-m shuttle run test was the most commonly test used in eight studies [66, 71, 88–91, 95, 96], the cycle-ergometer watt-max test was used in three studies [92–94], the 12-min Cooper test in two studies [55, 64], and the 9-min walk-run test [53], the 1200-m run test [58], the 6-min run test [65], the 8-min run test [59], and the Åstrand-Ryhming test [57] were all used in one study. In summary, the 20-m shuttle run test was the most commonly used assessment tool for evaluating CR endurance. It should be noted that none of the studies that used this test estimated peak oxygen uptake using an equation. This parameter was estimated solely in two studies, with cycle-ergometer watt max test [92] and Åstrand-Ryhming test [57]. In terms of body composition, five studies estimated body fat percentage using skinfold measurements [58, 73, 92, 97, 98], five studies used the sum of skinfolds [67, 72, 87, 100, 101] to measure adiposity and one study employed bioimpedance [99]. UB strength was evaluated using the handgrip test in four studies [56, 60, 61, 64] and using the bent arm hang test in another four studies [54, 69, 88, 101]. Anselma et al. [69] noted a high occurrence of zeros, which may compromise the Bent arm hang test’s discriminative power. Other studies assessed this component through medicine ball throw [59, 62], ball throw [63], the maximum number of repetitions of bench press at a rate of 25 lifts per minute [53] and push ups in 40 s [65]. In contrast to UB strength, where several aspects of strength are represented (including maximal strength, explosive strength, or muscular endurance), all the tests used to assess LB strength focused on jumps. This indicates that explosive strength or power is the primary aspect of LB strength examined in this review, with jumps being the most commonly used tests to assess this physical fitness component. The tests included the standing broad jump, which was used in ten studies [52, 56, 59, 60, 62–64, 70, 71, 88], the standing high jump was used in four studies [53, 54, 57, 69], lateral jumps in 15 s used in two studies [66, 102] and the 5-leap test, used in one study [86]. The sit and reach was the most frequently utilized method to evaluate the flexibility and was used in six studies [54, 56, 69–71, 101] while only one study employed the straight leg raise test [57]. The quality ratings for the studies indicated that 20 were rated as excellent, 15 as good, and only three were rated below good. It should be noted that only nine studies collected data from representative samples at all time points [54, 58, 64, 72, 73, 87, 88, 95, 102], and a further nine studies collected data from at least one representative sample [62, 67, 69, 86, 89, 92, 97, 98, 101]. The following studies collected data from representative samples in the following cities or regions: Budapest [58] and Kraków [56, 72, 87], Olomuc [62], Zaragoza [97], Odense [92], Hertogenbosch [69] and Northern Ireland [98]. Only six studies collected data from national samples at all time points: Hungary [58, 73], the Netherlands [54], Poland [64], Lithuania [88] and Finland [95, 102]. Additionally, five studies gathered data from at least one national sample from their respective countries: the Czech Republic [62], Belgium [101], Finland [86, 89], and the Netherlands [69]. The complete quality assessment is shown in Table 2.

Table 2.

Quality assessment for included studies

	Reporting						External Validity	Internal Validity		Power	Score	Descriptor
	1	2	3	4	5	6	7	8	9	10 †	%
Tutkuviene and Schiefenhövel [61]	1	1	0	1	0	1	0	1	1	–	67	Good
Kopecký et al. [62]	1	1	0	1	0	0	0*	0	0	–	33	Poor
Chulvi-Medrano et al. [52]	1	1	0	1	1	1	0	1	1	–	78	Good
Matton et al. [101]	1	1	1	1	1	0	0*	1	0a	–	67	Good
Westerstahl et al. [53]	1	1	1	1	1	0	0	1	0	–	67	Good
Photiou et al. [58]	1	1	1	1	1	0	1	1	0a	–	78	Good
Sedlak et al. [63]	1	1	0	1	0	0	0	1	0a	–	44	Fair
Jaakkola et al. [86]	1	1	1	1	1	1	0*	1	0	–	78	Good
Moreno et al. [97]	1	1	1	1	1	1	0*	1	1	–	89	Excellent
Mészáros et al. [73]	1	1	1	1	1	0	1	1	1	–	89	Excellent
Runhaar et al. [54]	1	1	1	1	1	0	1	1	0a	–	78	Good
Kowal et al. [72]	1	1	1	1	1	0	1	1	1	–	89	Excellent
Kowal et al. [87]	1	1	1	1	1	0	1	1	1	–	89	Excellent
Lovecchio et al. [55]	1	1	1	1	1	0	0	1	0	–	67	Good
Wedderkop et al. [92]	1	1	1	1	1	0	0*	1	1	–	78	Good
Aaberge et al. [57]	1	1	1	1	1	1	0	1	0a	–	78	Good
Przeweda et al. [64]	1	1	0	1	0	0	1	1	0a	–	56	Fair
Watkins et al. [98]	1	1	1	1	1	1	0*	1	1	–	89	Excellent
Venckunas et al. [88]	1	1	1	1	1	0	1	1	0a	–	78	Good
Smpokos et al. [71]	1	1	1	1	1	1	0	1	0a	–	78	Good
Huotari et al. [89]	1	1	1	1	1	0	0*	1	1	–	78	Good
Costa et al. [70]	1	1	1	1	1	1	0	1	0a	–	78	Good
Møller et al. [93]	1	1	1	1	1	0	0	1	1	–	78	Good
Andersen et al. [94]	1	1	1	1	1	1	0	1	1	–	89	Excellent
Sandercock et al. [90]	1	1	1	1	1	1	0	1	1	1	89	Excellent
Sandercock and Cohen [60]	1	1	1	1	1	1	0	1	1	–	89	Excellent
Kryst et al. [67]	1	1	1	1	1	0	0*	1	1	–	78	Good
Thomas et al. [96]	1	1	1	1	1	1	0	1	1	–	89	Excellent
Arboix-Alió et al. [91]	1	1	1	1	1	1	0	1	1	–	89	Excellent
Huotari et al. [102]	1	1	1	1	1	1	1	1	0	–	89	Excellent
Palomäki et al. [95]	1	1	1	1	1	1	1	1	1	–	100	Excellent
Anselma et al. [69]	1	1	1	1	1	0	0*	1	0	–	67	Good
Spengler et al. [65]	1	1	1	1	0	1	0	1	0	–	67	Good
Unger et al. [59]	1	1	1	1	1	1	0	1	0		78	Good
Kryst et al. [56]	1	1	1	1	1	0	0	1	1	–	78	Good
Nebiker et al. [66]	1	1	1	1	0	1	0	1	0	–	67	Good
Kutac et al. [99]	1	1	1	1	1	0	0	1	1	–	78	Good
Artymiak et al. [100]	1	1	1	1	1	0	0	1	1	–	78	Good

* Some of the samples were representative

^a Some of the tests were valid and reliable, but not all of them were

† Such studies do not consider the statistical power to detect a clinically important effect when making sample size calculations, but rather the representativeness of the population. This was only done in Sandercock et al. [90]. This item was excluded for the final score

1 = Clear hypothesis/aim/objective; 2 = Clear main outcomes to be measured; 3 = Clear participant characteristics for inclusion; 4 = Clear study description; 5 = Clear description of main findings; 6 = Actual probability values reported (except p < 0.001); 7 = Representative samples; 8 = Appropriate statistical test; 9 = Valid and reliable outcome measures (based on Castro-Piñero et al. [76] and Welsman and Armstrong [77]); 10 = Sufficient power to detect a clinically important effect where the probability value for a difference being due to chance is less than 5%; Descriptor = 0–20%, Bad; 21–40%, Poor 41–60%, Fair; 61–80%, Good; 81–100%, Excellent

Model selection based on a comparison of the AIC and the BIC is reported in Table 3. To numerically describe the health-related physical fitness trends among European children and adolescents, effect estimates, standard errors, and p values of each physical fitness component LMM, are reported in Table 4. Factors such as sex and age group are presented first, followed by linear, quadratic, and cubic trends over time, as well as interactions.

Table 3.

Model selection based on information criteria

		Cubic	Quadratic	Linear	Selected model
CR endurance	AIC	150.28	174.51	172.78	Cubic
	BIC	187.05	207.60	202.19
Body composition	AIC	− 303.24	− 300.72	− 300.46	Linear
	BIC	− 268.95	− 269.86	− 273.03
UB strength	AIC	181.74	249.90	258.74	Cubic
	BIC	220.21	284.52	289.51
LB strength	AIC	70.24	69.83	94.01	Quadratic
	BIC	107.80	103.64	124.05
Flexibility	AIC	− 45.57	− 34.16	− 4.96	Cubic
	BIC	− 19.52	− 10.71	15.88

For each physical fitness component, linear, quadratic and cubic mixed models were compared using Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC). Lower values indicate better model fit (bold)

CR cardiorespiratory, LB lower body, UB upper body

Table 4.

Effects estimates of five linear mixed models

LMM	Effect	Estimate	Standard error	p values
CR endurance	Intercept	− 0.1547	0.0712	0.044
	Sex	0.0050	0.0341	0.883
	Age group	0.0005	0.0684	0.994
	Time (linear)	− 0.5141	0.0454	< 0.001
	Time (quadratic)	0.0987	0.0236	< 0.001
	Time (cubic)	0.0738	0.0139	< 0.001
	Sex × time	− 0.0225	0.0364	0.537
	Age group × time	− 0.0665	0.0534	0.214
Body composition	Intercept	− 0.0287	0.0295	0.356
	Sex	0.0284	0.0230	0.218
	Age group	0.0062	0.0158	0.693
	Time (linear)	− 0.0944	0.0114	< 0.001
	Sex × time	0.0474	0.0191	0.014
	Age group × time	0.0029	0.0156	0.854
UB strength	Intercept	0.0516	0.0832	0.546
	Sex	0.0189	0.0324	0.559
	Age group	− 0.0085	0.0457	0.852
	Time (linear)	− 0.1443	0.0496	0.004
	Time (quadratic)	− 0.2681	0.0299	< 0.001
	Time (cubic)	− 0.1666	0.0189	< 0.001
	Sex × time (linear)	0.0277	0.0331	0.403
	Age group × time (linear)	0.0088	0.0392	0.822
LB strength	Intercept	0.0141	0.0369	0.707
	Sex	0.0134	0.0299	0.655
	Age group	− 0.0458	0.0423	0.279
	Time (linear)	− 0.2017	0.0285	< 0.001
	Time (quadratic)	− 0.0799	0.0146	< 0.001
	Sex × time	− 0.0552	0.0335	0.100
	Age group × time	− 0.0295	0.0352	0.403
Flexibility	Intercept	0.0329	0.0692	0.647
	Sex	0.0681	0.0337	0.046
	Age group	− 0.0769	0.0445	0.087
	Time (linear)	− 0.1511	0.0654	0.023
	Time (quadratic)	− 0.2115	0.0279	< 0.001
	Time (cubic)	− 0.0983	0.0259	< 0.001
	Sex × time	− 0.0283	0.0386	0.466
	Age group × time	− 0.0172	0.0398	0.667

According to Fühner et al. [38], lower-order trends and main effects are usually qualified by the highest-order significant trend or interactions. Thus, the interpretation of polynomial trends must start with the highest order trends. The LMMs analyses, developed using Jamovi, are available in Supplementary Material 5. Graphical representations of these trends can be found in Fig. 2. As a verification of the quality of the models, residual plots are illustrated in Supplementary Material 6—Fig. S1.

Fig. 2.

Temporal trends in cardiorespiratory endurance, body composition, upper body strength, lower body strength and flexibility. To facilitate readability, body composition z scores were multiplied by − 1. Therefore, positive scores indicate better health-related physical fitness. Panel a shows male observations, and panel b shows female observations. Solid lines and black circles represent children, whereas dashed lines and white triangles represent adolescents. Shaded area represents 95% confidence intervals for the model functions. CR cardiorespiratory, LB lower body, UB upper body

The coefficients define the functions of the secular trends shown in Fig. 2. Significant effects are in bold

CR cardiorespiratory, LB lower body, UB upper body

Cardiorespiratory Endurance

The LMM for CR endurance included 292 observations and 315,205 participants (47.61% females; 41.6% adolescents). The highest-order significant component was the positive cubic trend (B = 0.0738; SE = 0.0139; p < 0.001), which mainly defines the shape of CR endurance trend. Additionally, the positive quadratic (B = 0.0987; SE = 0.0236; p < 0.001) and negative linear (B = − 0.5141; SE = 0.0454; p < 0.001) effects were also significant. The CR endurance trend suggests an increase until the 1980s (no data available for female children), followed by a decline, which appears to be decelerating between 2015 and 2023 (Fig. 2). The LMM for CR endurance excluding 20 m shuttle run test included 204 observations and 278,357 participants (44.9% females; 42.5% adolescents). When the 20 m shuttle run test was excluded, the linear trend was found to be the best fit according to the information criteria (AIC = 45.87; BIC = 72.41). Therefore, the negative linear trend (B = -0.2340; SE = 0.0376; p < 0.001) determined its function. Figure 3 (left plot) shows a graphical comparison between the cubic trend of CR endurance, and this negative linear trend (excluding 20 m shuttle run test), from 1974 to 2023.

Fig. 3.

Temporal trends in cardiorespiratory endurance and body composition derived from alternative model specifications. The left plot compares cardiorespiratory endurance model including all available tests (solid line) and excluding the 20 m shuttle run test (dashed line). The right plot compares temporal trends in body composition (solid line) and BMI (dashed line). To facilitate readability, body composition and BMI z scores were multiplied by − 1. Therefore, positive z scores indicate better health-related physical fitness. Shaded area represents 95% confidence intervals for the model functions. BMI body mass index, CR cardiorespiratory, 20 m SRT 20 m shuttle run test

Body Composition

The LMM for body composition included 228 observations and 69,619 participants (22.6% females; 51.2% adolescents). Body composition best fitting trend is linear (Table 3). In this model, the negative linear trend (B = -0.0944; SE = 0.0114; p < 0.001) defines the shape. This LMM indicates trend differences between males and females (B = 0.0474; SE = 0.0191; p = 0.014). Figure 2 shows a continuous downward trend between 1972 and 2022, with no acceleration or deceleration periods. This trend appears to be more pronounced among males than females. In contrast to body composition, BMI followed a cubic trend, according to AIC and BIC (AIC = − 460.23; BIC = − 424.17). The LMM for BMI included 272 observations, 345,625 participants (44.9% females; 42.5% adolescents), and the positive cubic trend (B = 0.0443; SE = 0.0083; p < 0.001) is the critical source of variance. The quadratic (B = 0.0578; SE = 0.0087; p < 0.001) and linear trends (B = − 0.1235; SE = 0.0175; p < 0.001) were also significant. Figure 3 (right plot) shows a graphical comparison of the negative linear trend in body composition and the positive cubic trend in BMI. The BMI trend followed a slightly oscillating pattern: it improved before the 1980s, worsened until the early 2000s, and then improved again up to 2022.

Upper Body Strength

The LMM for UB strength included 346 observations and 318,330 participants (49.6% females; 43.2% adolescents). The primary source of variance affecting the UB strength function is the negative cubic trend (B = − 0.1666; SE = 0.0189; p < 0.001). There are also significant effects for the quadratic (B = − 0.2681; SE = 0.0299; p < 0.001), and the linear trend (B = − 0.1443; SE = 0.0496; p = 0.004). As shown in Fig. 2, there is a decline in UB strength from 1965 to 2023, with a period of stabilisation occurring between the 1980s and the 2000s. Furthermore, UB strength experienced the greatest decline of the five analysed physical fitness components.

Lower Body Strength

The LMM for LB strength included 316 observations and 317,246 participants (50.5% females; 54.5% adolescents). LB strength was the only physical fitness component whose best fitting trend is quadratic (Table 3). It is primarily characterized by a negative quadratic trend (B = − 0.0799; SE = 0.0146; p < 0.001). There is also a significant effect of the linear trend (B = − 0.2017; SE = 0.0285; p < 0.001). Figure 2 suggests an upward trend up to 1985–1990. After this period, the pattern gradually changed, showing a consistent decline until 2023.

Flexibility

The LMM for flexibility included 100 observations and 30,544 participants (40.3% females; 49.2% adolescents). In this model, the negative cubic trend (B = − 0.0983; SE = 0.0259; p < 0.001) defines the shape of flexibility trend. The quadratic (B = − 0.2115; SE = 0.0279; p < 0.001) and the linear trend (B = − 0.1511; SE = 0.0654; p = 0.023) are also significant. Additionally, this LMM indicates a systematic difference between males and females (B = 0.0681; SE = 0.0337; p = 0.046), while no evidence of sex-specific temporal trends was observed (Sex × Time, p > 0.05). Figure 2 illustrates a decline in flexibility which has accelerated over the last two decades.

Discussion

This systematic review aimed to analyse and update the temporal trends of the health-related physical fitness components in European children and adolescents aged 6–16 years. This review included 38 studies, five LMMs, and 1282 observations. Overall findings indicate a decline in pediatric health-related physical fitness from 1990 to the present. Eight studies included data up to 2020 [56, 59, 66, 67, 86, 89, 99, 100], and four studies [59, 66, 99, 100] had at least one time point after this date. The trend in all analysed physical fitness components during and after the COVID-19 pandemic, is a decline, except for CR endurance, which has stabilised since before the pandemic. However, this stabilisation does not appear when the 20 m shuttle run test is excluded. Further analysis is needed to fully understand the trends for each specific component of health-related physical fitness.

The European findings on CR endurance trends align with previous global studies, such as those conducted by Tomkinson et al. [34, 35] during their respective analysis periods and the LMM developed by Fühner et al. [38]. Despite using different LMM approaches (one being a nested LMM under physical fitness components and the other involving five separate LMMs for each component) both support a declining trend in CR endurance until the 2010s, at which point the negative trend appears to have stabilised or even tended to improve. The analysis which excluded the 20 m shuttle run test, showed a different trend, which decreased linearly throughout the entire analysed period. The new observation years provided by the present review, support the previously identified stabilisation period after the 2010s [35, 38]. However, when the 20 m shuttle run test is excluded from the analysis, the model suggests that this period may never have existed. This finding, raises the question of whether the observed stabilisation reflects a true change in the CR endurance trends, or is driven by the influence of 20 m shuttle run test on trend estimation. Additionally, any of both CR endurance LMMs did not show any interactions related to sex or age in European children and adolescents, unlike the global model, which included 12 European studies out of 22 and found sex interactions [38].

With regard to body composition in European children and adolescents, the trends have shown a linear decrease, which is more pronounced among males. This trend complements previous secular trends in body fat between 1951 and 2004 [44]. As both trends coincide from 1972 to 2004, the present review extends the monitoring period by 18 years. Despite the fact that this review includes more body composition outcomes and use a different statistical approach, the only difference is an acceleration breakpoint in 1985 [44]. Regardless, both studies are aligned in showing a consistent worsening of body composition (increase in adiposity). Another relevant finding of the present study is the clear difference between body composition and BMI trends. BMI showed an initial improvement before the 1980s, followed by a decline until the 2000s, and final improvement up to 2022. These changes in the direction of the trend, occurred within a small range, showing that the BMI trend was fairly stable. A global study, showed that European countries experienced only minor changes in BMI (< 0.5 kg/m²), and some of them have begun to slow this trend compared to others in the developed countries [41]. Considering that BMI has not been shown to reliably detect excess adiposity [43], and comparing its trend with the body composition model, it could be asserted that the BMI trend underestimates the magnitude and persistence of the public health crisis of the obesity epidemic [39, 40].

Regarding the trends in pediatric strength in Europe, the UB strength reveals a clear decrease from before the 1970s to the present, with a period of stabilisation between the 1980s and 2000s, while the LB strength increased until 1980s, and then gradually decreased until 2023. There were no differences in these trends between sexes or age groups. To understand this difference between UB strength and LB strength, it is first important to conceptualize the type of strength, and the type of test used to assess each limb. The tests used to measure UB strength in the included studies mostly evaluated strength endurance (or muscular endurance) and maximal strength, whereas all the tests used for LB strength were explosive strength or power manifestations. This understanding is necessary for the interpretation of the strength results, and their comparison with previous studies.

With respect to UB strength, Dooley et al. [47], focused on the absolute handgrip strength and found a clear and progressive global increase from 1967 to 2017, but they mentioned the limitations of the absolute handgrip trends they analysed, compared to relative handgrip values. The UB strength model included four studies which used absolute handgrip strength, but most studies used tests of strength relative to body mass. The findings in UB strength are in line with reviews by Masanovic et al. [36] and Eberhardt et al.[37], in which most included studies reported a decrease in this component. In a year-by-year analysis, the decline until the 1980s and the second decline after the 2000s are consistent with the LMM developed by Fühner et al. [38] for global relative strength, which describes a decline until 1982, and a second decline after 2006. However, it is not consistent with the period from the 1980s to the 2000s, when European UB strength shows a stabilisation of the trend, while the nested model of Fühner et al. [38] shows an increase between 1982 and 2006. Nevertheless, it is important to note that the present model focuses exclusively on UB Strength tests in European children and adolescents, whereas the aforementioned model covered the whole world and included only five UB strength tests out of eight [38]. Thus, the information provided by the UB strength model complements the previous findings on relative strength [38], and extend the monitoring beyond 2014 to 2023, indicating that UB strength has continued to decline.

Relative to LB strength, understood as lower body explosive strength or power due to the included jumping tests, is in line with the global trends which converge in finding an increase in this physical fitness component until the 1980s and a general decline after the 1990s [38, 45, 46]. Tomkinson et al. [46], in their analysis of the standing broad jump trend, found that the tendency slowed down during the 1990s, and Fühner et al. [38], found differences between sex in the year when the power started to decrease: 1982 for males and 1994 for females. This trend difference between sexes was not detected by the present LB strength model. Nevertheless, all the evidence, including the present systematic review, seems to indicate a trend that follows a quadratic function, with slight variation in the years when the trend changed direction: between the 1980s and the1990s. The present study extends this trend from 2015 to 2023, and confirms that the quadratic function has not changed, meaning that the LB strength continues to decrease.

Finally, flexibility in European children and adolescents shows a clear and progressive decline until the turn of the century, when the decrease accelerated to 2020. The findings are in line with those from earlier global systematic reviews, with the majority of the included studies reporting a decline in flexibility [36, 37]. The lack of a previous systematic review with statistical analysis that accurately describes flexibility trends, and the smaller number of observation points in this model compared to the others, confirms that this is probably the least studied component of health-related physical fitness. Masanovic et al. [36] affirmed that flexibility studies were less frequent and insufficient to draw meaningful conclusions. However, the present flexibility model had sufficient statistical power to detect significant trends, although there were no trend differences by age group or sex.

As final considerations, to combat the decline in health-related physical fitness among European children and adolescents, it is first necessary to optimise the monitoring trends. The first step could be to optimize the monitoring of these trends by reaching a European consensus on the tests used to assess the physical fitness, allowing easy and homogeneous comparisons between countries and time points, without the need for score transformations such as z scores. One of the first tries for the standardization was the EUROFIT test battery [25, 26] in 1988, followed by other batteries [27–32]. Despite these attempts and given the variety of tests used in the included articles in this review, even those published more recently, it seems that after more than 30 years, this European standardization has not yet been achieved. Considering the most commonly used tests, this review is broadly consistent with the recent consensus on the YFIT battery [33]. Although the YFIT consensus recommends using BMI to assess body composition, the authors acknowledge that it showed the lowest level of agreement, with some experts questioning its suitability due to its inability to differentiate between fat mass and fat-free mass. Nevertheless, they support its inclusion because BMI is the accepted measure for defining overweight and obesity, and because it is simple and feasible. Flexibility is not included in this battery, but the recommendation based on the most frequently used test in the articles of the present review would be to use the sit-and-reach test if the aim is to assess this component.

In addition to these methodological improvements, health and education institutions must also intensify their efforts to ensure compliance with the WHO’s physical activity guidelines [1]. The WHO’s most recent Global Status Report on Physical Activity revealed that only 25% of European youths aged 11 to 17 engage in at least 60 min of moderate-to-vigorous physical activity daily [103]. A further study conducted across 28 European countries, showed that only 19.4% of adolescents engaged in muscle-strengthening activities for at least three days per week [104], whereas the global prevalence of meeting this recommendation is 38.5% [105]. According to the present results, the components with the worst slopes in recent years were the strength trends. This is consistent with the prevalence of muscle-strengthening recommendations being lower among Europeans than globally (19.4% vs. 38.5%). Therefore, institutions must increase their emphasis on promoting the recommendation to engage in muscle and bone-strengthening activities three times per week. Nevertheless, the observed trends in CR endurance and body composition remain negative, highlighting the ongoing relevance of these issues as public health concerns and the need to continue promoting the WHO’s recommendations for moderate-to-vigorous physical activity.

Limitations

Despite its strengths such as the use of specific LMMs for each physical fitness component, the incorporation of age groups to detect potential differences between children and adolescents, and the inclusion of a study quality assessment, this review has some limitations. First, it is based on published studies only, which may exclude relevant national data that have not been published in scientific journals, limiting the scope of the findings. Second, not all the European countries are represented in this review. Third, the variability in the study methods, sample sizes, and especially the tests used to assess the physical fitness, introduces heterogeneity that may have affected the results. Although z score standardisation was attempted to mitigate this problem, it should be noted that it does not fully resolve comparability issues. Fourth, while all of the included tests were field-based, and so not as valid or reliable as laboratory tests, it should be considered that some fitness tests may not be appropriate for this population. For instance, although reviews have supported the validity of the 20 m shuttle run test using correlation analyses [106, 107] an age-specific validation study using agreement-based analyses, reported wide limits of agreement between field-estimated and laboratory-measured peak oxygen uptake in children [77]. Another different example is that the bent arm hang test has been shown to have low discriminatory power due to a high proportion of zero scores, which may limit its sensitivity [69]. Therefore, it should be acknowledged that field-based physical fitness test scores, as well as estimates derived from them, are indirect approximations of physical fitness, and that trends should be interpreted with caution. Fifth, although the five LMMs for each physical fitness component approach may provide more specific results, it has less statistical power than a nested LMM under the physical fitness components [38]. Sixth, the study quality assessment revealed that only nine studies collected data from representative samples of their respective countries, and a further nine studies collected data from at least one representative sample. Of these studies, only six analysed data from national samples at all time points, and five analysed data from at least one time point.

Conclusions

This systematic review provides an updated synthesis of the temporal trends in health-related physical fitness among European children and adolescents incorporating recent data from after 2014 and the period of COVID-19 pandemic. Overall, the findings suggest persistent declines in several health-related components of physical fitness, with variations in the shape and magnitude of trends. From a surveillance perspective, the heterogeneity between studies highlights the need for a standardised, valid and feasible single European test battery. Meanwhile, statistical approaches that mitigate the comparability issues will remain necessary. The deterioration of health-related physical fitness observed in recent decades, particularly in terms of strength, reinforces the need to intensify efforts to promote physical activity. In this context, public institutions should continue to promote the WHO physical activity recommendations, with particular emphasis on the muscle-strengthening recommendations.

Abbreviations

CR

Cardiorespiratory

BMI

Body mass index

UB

Upper-body

LB

Lower-body

SE

Standard error

LMM

Linear mixed model

Author Contributions

Conceptualization: DDV, AML, JAC; Methodology: DDV, AML, JAC; Investigation: DDV, AML; Data curation: DDV, GLB, AM, AGB; Formal analysis: DDV, GLB, AM, AGB, AML; Visualization: DDV; Writing – original draft: DDV; Writing – review and editing: DDV, GLB, AM, AGB, AML, JAC; Supervision: AML, JAC. All authors read and approved the final version of the article.

Funding

The article processing charges for this publication were founded by the EXER-GENUD research group (S72-23R) of the Government of Aragon. DDV, who wrote the original draft of the manuscript, received a PhD grant from the Government of Aragon (CUS/621/2023).

Availability of Data and Material

All data generated or analysed during this study are included in this published article and its supplementary information files.

Declarations

Ethics Approval and Consent to Participate

Not applicable.

Consent for Publication

Not applicable.

Competing Interests

The authors declare that they have no competing interests.